7 views

Uploaded by erik

- Uncertainties in climate change projections and regional downscaling 2010
- Shelter for Warm-humid Climates
- Jan 15 CS
- Application of Data Mining Techniques in Weather Prediction and Climate Change Studies.pdf
- Local Perceptions of Climate Variability and Change in Tropical Forests of Papua, Indonesia
- Typh0000
- One of the Biggest Thresholds Has Already Been Reached
- Adapting to Climate Change Thresholds, Values, Governance
- 5 pidato
- climateregion
- Solar Ete
- Climate Change
- 2017 Was the Year of the Billion-Dollar Disaster _ Climate Central
- Hurricanes
- Dispersion Modeling
- States Have No Responsibility to Other States
- Revised Geo 1200 Syllabus Fall 13
- A Nemo Metro
- Ao Xiii, 1-2, 2006 Papers
- try out un XII

You are on page 1of 74

Statistical Climatology

University of Lneburg

12 to 16 March 2001

1

STOCHASTIC CLIMATE MODELS 18

To the nature of red noise in the climatic series, G.V. Alekseev (Arctic and Antarctic Research

Institute, St.-Petersburg, Russia) 18

Frequency Modulation in the Instrumental Record, P. Carl (Institute for Freshwater Ecology and Inland

Fisheries, Research Association Berlin of the Leibniz Society, Berlin, Germany) 18

Statistical algorithms for solving inverse problems on predictability of regional climate, Anatoly I. Chavro, Yegor

V. Dmitriev (Institute for Numerical Mathematics RAS, Russia) 18

Validating Interannual Variability in Atmospheric General Circulation Models, Carsten S. Frederiksen1 and

Xiaogu Zheng2 (1Bureau of Meteorology Research Centre, Melbourne, Australia; 2National Institute of

Water and Atmospheric Research, Wellington New Zealand) 19

The character of prediction utility in dynamical systems relevant to climate and weather, Richard Kleeman

(Courant Institute for Mathematical Sciences) 19

Stochastic dynamics in a simple model of seasonal open ocean deep convection, Till Kuhlbrodt (Potsdam Institute

for Climate Impact Research, Climate System Dept., Potsdam, Germany) 19

Prediction of ENSO using an optimized mixture of linear models, Frank Kwasniok (Leibnitz-Institut fr

Atmosphrenphysik, Khlungsborn, Germany) 20

Simple Stochastic Models of the Thermohaline Circulation, Adam Monahan (Humboldt-Universitaet zu Berlin) 20

Improvement of prediction using SSA methods in addition to MEM: Case of NAO study, Sylvia Antunes1, H.

Oliveira Pires1,2 e Alfredo Rocha2 (1Instituto de Meteorologia, R. C ao Aeroporto, Lisboa, 2Universidade

de Aveiro) 20

On the interrelationships between multisite climatic temperature forecasts obtained from downscaling regional

sea-level-pressure fields, Ulrich Callies (Institute for Hydrophysics, GKSS Research Center; Geesthacht,

Germany) 20

Model Error and Uncertainty, Peter Challenor (Southampton Oceanography Centre, Southampton) 21

The Winter Temperature Downscaling Experiments Using Canonical Correlation Analysis in China, Youmin

Chen (Agrometeorology Institute, Chinese Academy of Agric Sciences, Beijing, China) 21

On-line Gaussian Processes and their Potential in Data Assimilation, Dan Cornford, Lehel Csato and Manfred

Opper (Computer Science, Aston University, Birmingham) 21

Statistical Downscaling of Largescale GCM Data to Local Ice Accumulation in Greenland, Traute Crueger, Hans

v. Storch (GKSS Research Center, Geesthacht, Germany) 21

Spatialisation of 24 hours precipitation with 100 years return period, Mojca Dolinar1, Damijana Kastelec2

(1Hydrometeorological institute of Slovenia, Ljubljana, Slovenia; 2University of Ljubljana, Biotehnical

Faculty, Ljubljana, Slovenia) 22

Bayesian estimate of halphen type a law's parameters, S. El Adlouni1, A. Bouziane, D. Ouazar, A. Zoglat1 (1

Faculte des Sciences-Rabat; Ecole Mohammadia d'Ingenieurs, Rabat, Morroco) 22

EMAD: An Empirical Model of the atmospheric Dynamics, Juan Pedro Montavez and Jin Song von Storch

(Meteorological Institute of Hamburg University) 23

2

On the Natural Regularity and Predictability of Terrestrial and Extra-Terrestrial Parameters, Nityanand Singh

(Indian Institute of Tropical Meteorology, Pashan, Pune, India) 23

Stochastic Tropical Atmospheric Boundary-Layer Dynamics: Convective Downdrafts and Recovery, Discharge-

Recharge Processes, and 1/f-Noise, Jun-Ichi Yano and Richard Blender (Meteorolo-gisches Institut,

Universitaet Hamburg, Hamburg, Germany) 23

Variational Assimilation of botanical proxy Data for the Reconstruction of Palaeoclimate, Christoph Gebhardt

(Meteorologisches Institut der Universitt Bonn, Germany); Norbert Khl (Institut fr Palontologie,

Universitt Bonn, Germany); Andreas Hense (Meteorologisches Institut der Universitt Bonn, Germany);

Thomas Litt (Institut fr Palontologie, Universitt Bonn, Germany) 24

How realistic is GCM-simulated precipitation? Nikolaus Groll, Martin Widmann (Institute of Hydrophysics,

GKSS Research Centre, Geesthacht, Germany) 24

Homogenisation and Interpolation of incomplete climatological Data with simple physically based Models,

Andreas Hense, Christoph Gebhardt (Meteorologisches Institut, Universitt Bonn,Bonn, Germany) 25

Trends of Maximum and Minimum Temperatures in Northern South America, Ramon A. Quintana-Gomez

(Department of Earth Sciences, Universidad de los Llanos Ezequiel Zamora, Barinas, Estado Barinas,

Venezuela) 25

Investigation of Spatial and Temporal Ozone Immission Structures under Stochastic-Deterministic Model

Assumptions, Heike Hoffmann (Graduate School of Spatial Statistics, Freiberg, University of Mining and

Technology, Germany) 25

Reduced-order schemes for data assimilation in oceanographiy based on the Kalman filter, I. Hoteit and D.T.

Pham (Laboratoire de Modelisation et Calcul, Grenoble, France) 25

Cloudiness and sea surface temperature: tackling the problem of missing values, Richenda Houseago-Stokes and

Peter Challenor (Southampton Oceanography, Centre Southampton, UK) 26

Removing errors in simulated distributions of climate elements using a non-parametric adjustment procedure,

Radan Huth1, Jan Kysel1,2 (1 Institute of Atmospheric Physics, Prague, Czech Republic; 2 Dept. of

Meteorology and Environment Protection, Charles University, Prague, Czech Republic) 26

Simplified EOFs - three alternatives to rotation, Ian Iolliffe (University of Aberdeen, UK); Mudassir Uddin

(University of Karachi, Pakistan); Karen Vines (Open University, UK) 26

Reconstruction of large-scale climate, J.M. Jones (GKSS Research Centre, Geesthacht, Germany) 27

On a probabilistic interpretation of variational data assimilation and model error estimation, Gennady A. Kivman

(Alfred-Wegener-Institut for Polar- and Marine Research, Bremerhaven, Germany) 27

Statistical presentation of wind wave climate, Leonid Lopatoukhin1,2 , Valentine Rozhkov1, Alexander

Boukhanovsky 1,3, (1 State Oceanographic Institute; 2 St. Petersburg State University (Dep. Oceanology); 3

Institute of High Performance Computing and Data Base, Russia) 27

Model-based spatial modelling of climatological data, J. Mateu 1, P. Juan 1, C. Antolin 2,3, C. Ano 3 28

(1 Department of Mathematics, Universitat Jaume I, E-12071 Castellon, Spain; 2 Department of Vegetal Biology,

Universitat de Valencia, Burjassot, Spain; 3 CIDE-Centre for Desertification, Cami de la Marjal s/n.,

Albal (Valencia), Spain) 28

Regional GEV-growth curves of yearly precipitatation depths in Belgium for durations ranging from 10 minutes

to 30 days, Daniel Gellens, Gaston Demare (Royal Meteorologuical Institute of Belgium, Brussels,

Belgium) 28

3

Extreme precipitation events in southeastern South America and the relationships with summertime large-scale

South American Stationary Systems (SASSY), Leila M. V. Carvalho1,2, Charles Jones1, Brant Liebmann3

(1Institute for Computational Earth System Sciences (ICESS), University of California, Santa Barbara,

California, USA; 2Dept. of Atmospheric Sciences, University of Sao Paulo, Sao Paulo, Brazil; 3NOAA-

CIRES Climate Diagnostics Center, Boulder, Colorado,USA) 28

Diagnosis of variations in the climate-related flood risk - Some results from the Interior Western United States,

Shaleen Jain, Martin Hoerling, Gary Bates, Randall Dole (NOAA/CIRES Climate Diagnostics Center,

Boulder, CO) 28

The Climatological Probability of Drought in Northern Province of South Africa, Tiba Kabanda (School of

Environmental Sciences, University of Venda, Thohoyadou, South Africa) 29

Statistical Analysis of Point Discharge Current at Different Time Scale Over the Pune Region, S.S.

Kandalgaonkar, M.I.R. Tinmaker, M.K. Kulkarni (Indian Institute of Tropical Meteorology, Pune, India)

29

A Dedacal Study of Premonsoon Seasons Thunderstorms Over Pune Region by Using Markov Chain Model, S.S.

Kandalgaonkar, M.K. Kulkarni, M.I.R. Tinmaker (Indian Institute of Tropical Meteorology, Pune, India)

29

Trends of Atlantic wave extremes as simulated in a 40-year wave hindcast using kinematically reanalyzed wind

fields, Xiaolan Wang* and Val R. Swail (Network Strategies Division, AMWSD Meteorological Service of

Canada, Downsview Ontario, Canada) 30

Hurricane Landfall Winds Along the United States Gulf and Atlantic Coasts: Monte Carlo Simulations, Gary D.

Skwira*, Douglas A. Smith+ and Richard E. Peterson* (Wind Science and Engineering Program, Texas

Tech University, Lubbock, TX, USA; *Department of Geosciences, +Department of Civil Engineering) 30

Simulation of ocean wave climate changes through wind variance changes on different timescales, E. Bauer1,R.

Weisse2 (1PIK, Potsdam; 2GKSS, Geesthacht, Germany) 30

Frequency and Intensity of Mediterranean Cyclones in a doubled CO2 scenario, P.Lionello (University of Lecce),

F.Dalan (University of Padua) 30

Extreme waves and surge events in a doubled CO2 scenario in the Adriatic Sea. P.Lionello (University of Lecce)

E.Elvini, A.Nizzero (University of Padua) 31

Extreme value analysis of absolute minimum temperature and minimum dew-point temperature at Sodankyl,

Finland, Jaakko Helminen, Niklas Sjblom (Finnish Meteorological Institute (FMI), Helsinki, Finland) 31

Observed Changes in Temperature Extremes of Korea, Won-Tae Kwon, Hee-Jeong Baek*, Byung-Choel Choi

(Meteorological Research Institute, Korea Meteorological Administration; * Seoul National University,

Korea) 32

The role of atmospheric large-scale circulation over Europe and North Atlantic in sea level variability on the

Polish coastal zone, Miroslaw Mietus (Institute of Meteorology and Water Management, Maritime

Branch, Gdynia, Poland) 32

Space-time variograms and a functional form for total air pollution measures, De Iaco, S.1, Myers, D. E.2, Posa,

D.3,4 (1 Universita di Chieti, Pescara Italy 2 University of Arizona, Tucson AZ - USA; 3 Universit di

Lecce, Lecce - Italy; 4 IRMA - CNR, Bari - Italy) 32

geostrophic model, Carlos Pires (CGUL, - Centro de Geofisica da Universidade de Lisboa, Lisboa,

Portugal) 33

A multisite self-consistent Weather Generator for Impact Studies in Portugal, Budong Qian, Joo Corte-Real,

Hong Xu (ICAT, Faculty of Sciences, University of Lisbon, Portugal) 33

Monthly mean forecasts of rainfall and temperature in Portugal, Joao Rio, Carols Pires (CGUL, - Centro de

Geofisica da Universidade de Lisboa, Lisboa, Portugal) 33

4

Sensitivity Studies on Wind Measurement Errors in Mesoscale Model Simulations, Radhika Ramachandran,

Prakash M Dolas (Space Physics Laborator, Vikram Sarabhai Space Centre, 34

Trivandrum, India) 34

Building the bridge between sequential data assimilation and geostatistics: the stationarity issue, Julien Snegas,

Hans Wackernagel (Centre de Geostatistique, Ecole des Mines de Paris, Fontainebleau, France) 34

Public Debate and Climate Model Products: Whereof must one be silent? Leonard A. Smith (Centre for the

Analysis of Time Series, London School of Economics, London, UK) 34

Multiple Analysis of Series for Homogenization (MASH). Seasonal Application of MASH. Automatic using of

Meta Data, Tamas Szentimrey (Hungarian Meteorological Service, Budapest, Hungary) 35

A data assimilation method with a coupled ocean-land-atmosphere model and its impact on the ocean state and on

medium-range pre-dictions, Clemente A. S. Tanajura, Konstantin P. Belyaev (Laboratorio Nacional de

Computao Cientifica (MCT/LNCC), Petrpolis, RJ, Brazil) 35

Ben P. Kirtman (Center for Ocean-Land-Atmosphere Studies (IGES/COLA), Calverton, MD, USA) 35

Assessing Model Skill in Climate Simulations, Karl E. Taylor (Program for Climate Model Diagnosis and

Intercomparison (PCMDI), Lawrence Livermore National Laboratory, Livermore, CA, USA) 36

XXth Century Behaviour of Temperature and Precipitation Extremes in Hungary, Jnos Mika, Tams

Szentimrey (Hungarian Meteorological Service, Budapest, Hungary); Judit Bartholy, Rita Pongrcz

(Etvs Lrnd University, Dept. Meteorology, Budapest, Hungary); Lszl Szeidl (Department of

Mathematics, University of Pcs, Hungary) 36

Homogenized Daily Temperatures for Trend Analyses in Extremes over Canada, Lucie Vincent, Xuebin Zhang,

Barrie Bonsal and Bill Hogg (Climate Research Branch, Environment Canada) 36

The analysis of extreme temperature distributions on the example of Poland, Joanna Wibig (Department of

Meteorology and Climatology University of Lodz) 37

Comparison of 20- and 50-year return values in GCM-simulated, downscaled and observed temperature series,

Jan Kysel1,2, Radan Huth1 (1Institute of Atmospheric Physics AS CR, Prague, Czech Republic; 2Dept. of

Meteorology and Environment Protection, Charles University, Prague, Czech Republic) 37

Test on the probability distribution of maximum air temperatures, Jana Jurekov1, Jan Picek2, Jaroslava

Kalvov1* (1Charles University in Prague, 2Technical University in Liberec, Czech Republic; *Dept. of

Meteorology and Environment Protection, Faculty of Mathematics and Physics, Charles University,

Praha, Czech Republic) 37

Three Approaches to determine extreme years of Global Atmospheric Temperature, Luis Gimeno1,*, Juan Ael 1,

Higinio Jorge1, Pedro Ribera1, David Gallego2, Ricardo Garcia2, Emiliano Hernandez2 (1Universidad de

Vigo. Facultad de Ciencias de Orense, Orense, Spain; 2Universidad Complutense de Madrid. Facultad de

Ciencias Fisicas, Madrid. Spain; *Corresponding author) 38

Bayesian Statistical Modelling of Nonlinear Climate Processes, E. P. Campbell, Y. Li (CSIRO Mathematical and

Information Sciences, , Wembley, Australia) 38

Modelling and characterizing extreme drought events, Jess Abaurrea, Ana Carmen Cebrian (Dpto. Metodos

Estadsticos, Universidad de Zaragoza, Ed. Matemticas. Pedro Cerbuna, 12., Zaragoza, Spain,) 38

Extremes in Instationary Time Series, Jrgen Grieser (Institute for Meteorology and Geophysics, University

Frankfurt, Germany) 39

Comparison of Several Methods for the Estimation of Extreme Value Distribution in Time Series with Trends,

Henrique Oliveira Pires, Joao Pestana Ferreira (Instituto de Meteorologia, Lisboa, Portugal) 39

The Possible Influence of Large Volcanic Eruptions on El Nino, Philippe Naveau (National Center for

Atmospheric Research, Boulder, USA) 39

5

SIMULATION MODELS 39

Simplified atmosphere models with realistic variability, Ulrich Achatz (Leibniz-Institut fuer Atmosphaerenphysik

an der Universitaet Rostock e.V., Kuehlungsborn, Germany); J.D. Opsteegh (KNMI, De Bilt, The

Netherlands) 39

Modelling of Temporal Rainfall by the Use of Stochastic Point Process Models in View of Scaling, Konrad Bogner

(Institute of Hydrology, University of Agricultural Research,Vienna, Austria) 40

Statistical Modelling of Tornado Occurrence, Harold E. Brooks (NOAA/National Severe Storms Laboratory

Norman, Oklahoma) 40

A method to extend the temporal validity of nested regional climate simulations, Udo Busch and Dietrich Heimann

(DLR Oberpfaffenhofen, Institut fr Physik der Atmosphre, Germany) 40

Stochastic generation of daily precipitation, Aristita Busuioc1, Hans von Storch2 and Rodica Tomozeiu1 (1

National Institute of Meteorology and Hydrology, Bucharest, Romania; 2 Institute of Hydrophysics,

GKSS Research Center, Geestacht, Germany) 41

Statistical Downscaling of Daily, Multi-Site Precipitation in Tropical and Sub-Tropical Climates, Stephen P.

1 1 2 3 1

Charles , Bryson C. Bates , Stephen J. Crimp and James P. Hughes ( CSIRO Land and Water,

2

Wembley, Australia; Queensland Department of Natural Resources, Indooroopilly, Australia;

3

Department of Biostatistics, University of Washington, Seattle, USA) 41

Stochastic Weather Generator Model for Simulation of Multiple Meteorological Time Series, Ildik Dobi-

Wantuch, Jnos Mika (Hungarian Meteorological Service, Budapest, Hungary); Lszl Szeidl

(Department of Mathematics, University of Pcs, Hungary) 42

Interdiurnal and interannual variability in stochastic daily weather generator: Modelling and the role in

agricultural and hydrologic studies, Martin Dubrovsky1, Josef Hejzlar2, Jan Kysely3, Josef Buchtele4,

Zdenek Zalud5, Daniel Marvan6 (1Institute of Atmospheric Physics AS CR, Hradec Kralove, Czech

Republic (CR), 2Hydrobiological Institute AS CR and Faculty of Biological Sciences USB, Ceske

Budejovice, CR; 3Institute of Atmospheric Physics AS CR, Prague, CR; 4Institute of Hydrodynamics AS

CR, Prague, CR; 5Mendel University of Agriculture and Forestry, Brno, CR; 6Faculty of Mathematics

and Physics, Charles University, Prague, CR) 42

Temperature predictability in the greater Mediterranean area, R. Garca, E. Hernndez, T. Muoz, P. Ribera*

and L. Gimeno* (Dto Fsica de la Tierra II, Universidad Complutense, Madrid Spain; Dto Fsica

Aplicada, Universidad de Vigo, Campus de Orense, Orense, Spain) 42

A new validation scheme for the evaluation of climate model outputs, F.-W. Gerstengarbe, M. Kcken, P.C.

Werner (Potsdam-Institute for Climate Impact Research, Potsdam, Germany) 43

The construction of multi-site and multi-variate scenarios for Mediterranean regions based on Monte Carlo

simulations with a conditional weather generator, C.M. Goodess and J.P. Palutikof (Climatic Research

Unit, University of East Anglia, Norwich, UK) 43

Reconstructing large-scale variability from paleoclimatic evidence by means of pattern nudging, Martin Widmann

1

, Hans von Storch 1, Ingo Kirchner 2 (1 GKSS Research Center, Geesthacht, Germany, 2 Max-Planck-

Institute for Meteorology, Hamburg, Germany) 44

Spatio-temporal patterns of Meteosat Satellite Thermal Infrared Derived Convective Clouds over Ethiopia as

Described by PCA & MVC, Abebe Yeshanew (Senior Meteorologist, National Meteorological Services,

Addis Ababa, Ethiopia) 44

Skill of seasonal hindcasts as a function of the ensemble size, F.W. Zwiers and V.V. Kharin (Canadian Centre for

Climate Modelling and Analysis, Victoria, BC, Canada) 44

6

Determining extreme sets in multivariate distributions, onika. Lakatos (Hungarian Meteorological Service); I.

Matyasovszky (Department of Meteorology, Etvs Lornd University, Budapest) 44

Climate extremes in Germany and its relation to the objective weather type classification of the German Weather

Service, Peter Bissolli, Udo Baum (Deutscher Wetterdienst, Offenbach, Germany) 45

Statistical methods and case studies investigating extreme climatology of the Carpathian Basin, Judit Bartholy1,

Laszlo Szeidl2 (1Dept. of Meteorology Eotvos Lorand University, Budapest Hungary; 2Dept. of

Mathematics University of Pecs, Hungary) 45

SIMULATION MODELS 46

Simulation of multivariate time series of meteorological data, Asa Forsman, Claes Bjrklund and Anders Grimvall

(Department of Mathematics, Linkping University, Linkping, Sweden) 46

Modelling the probability of precipitation using the Marcov chain MonteCarlo technique, Martin Kappler

(Fachbereich Statistik, Universitt Dortmund); Andreas Hense (Meteorologisches Institut Universitt

Bonn); Siegfried Schach (Fachbereich Statistik Universitt Dortmund) 46

Reproduction of variance in statistical downscaling: inflation vs. noise addition, Radan Huth1, Jan Kysel1,2

(1Institute of Atmospheric Physics, Prague, Czech Republic; 2Dept. of Meteorology and Environment

Protection, Charles University, Prague, Czech Republic) 46

Simulating meteorological data by block resampling, Anders Nordgaard, Asa Forsman, Anders Grimvall

(Department of Mathematics, Linkping University, Linkping, Sweden) 47

The Amazon climate: A multiscale and nonstationary processes, Guillermo O. Obregn, Carlos Nobre (Centro de

previso de Tempo e Estudos Climticos - CPTEC / Instituto Nacional de Pequisas Espaciais INPE,

Cachoeira Paulista, So Paulo Brazil) 47

First indication of anthropogenic climate change in local wave conditions in the North Sea, Arnt Pfizenmayer

(Institute for Coastal Research, GKSS Research Centre; Geesthacht, Germany) 47

Ensemble Predictions with a Limited-Area Mesoscale Atmospheric Model - Implications for the Interpretation of

its Direct Model Output, Susanne Theis, Andreas Hense (Meteorologisches Institut der Universitt Bonn,

Bonn, Germany); Ulrich Damrath, Volker Renner (Deutscher Wetterdienst, Offenbach, Germany) 47

Modeling stochastic structure of local climate as downscaling of GCM fields, 1,2Marina Timofeyeva and 1Rachael

Craig. (Department of Geology, Kent State University, Kent, Ohio; UCAR/NOAA/National Weather

Service, Silver Spring, Maryland) 48

Numerical experimentation with regional atmospheric models, Hans von Storch, Ralf Weisse (Institute of Coastal

Research, GKSS Research Centre, Germany) 48

Exploiting the emerging signal of anthropogenic climate change in probabilistic climate forecasting, Myles Allen

(Space Science & Technology Department, Rutherford Appleton Laboratory, Chilton, Didcot, UK) 49

Significance of Anthropogenic Radiative Forcings in Simulation of Temporal Pattern of Surface Air Temperature

in Monsoon Climate, Ajay Singh (School of Management, Indian Institute of Technology- Bombay, Powai,

India) 49

Decadal variations of Greenland precipitation over the past 1400 years, K. K. Andersen and P. D. Ditlevsen,

(Department of Geophysics, Copenhagen, Denmark) 49

Lessons taught from the Homogenization of 200 US First Order Stations, Auguste C. Boissonnade, Ph.D.1 and

Larry Heitkemper2 (1Risk Management Solutions, Inc. Menlo Park, California, USA; 2EarthSat

Corporation, Washington D.C., USA) 49

7

North Atlantic osccillation projection on regional climate predictability, Roxana Bojariu and Daniela Paliu

(National Institute of Meteorology and Hydrology, Bucharest, Romania) 50

What influences decision-makers in drought prone areas when allocating water resources? Bernard Cornlis

(Dpartement de Gomatique Universit de Lige) 50

Assessing uncertainty in anthropogenic signal estimates using space-time and trend-based optimal detection

approaches, Nathan Gillett, Gabi Hegerl, Myles Allen 50

Applications of Statistical Climatology in Finance, Stephen Jewson, Anders Brix (Risk Management Solutions) 51

Trends in northern hemisphere winter cyclone activity based on a comparison of cyclone indices, Christopher J.

Paciorek (Department of Statistics, Carnegie Mellon University); James S. Risbey (Department of

Engineering and Public Policy, Carnegie Mellon University); Richard D. Rosen (Atmospheric and

Environmental Research, Inc. Cambridge, Massachusetts); Valerie Ventura (Department of Statistics,

Carnegie Mellon University) 51

Signal analysis of the atmospheric mean temperature 500/1000 hPa north of 55 N between 1949 and 1994, Heiko

Paeth, Andreas Hense (Meteorologisches Institut, Universitt Bonn, Bonn, Germany) 51

Expert Assessment of Uncertainties in Detection and Attribution of Climate Change, James Risbey (University of

Utrecht, Department of Science, Technology and Society, Utrecht, The Netherlands) 52

A Maximum Entropy Approach for Extracting Forecast Probabilities from Finite Ensembles, M.S. Roulston1,

L.A. Smith1 and C. Ziehmann2 (1Mathematical Institute, Oxford University, UK; 2Institute of Physics,

Potsdam University, Germany) 52

Coupled sea ice and atmosphere variability in the Weddell Sea, Antarctica, Silvia A. Venegas (Danish Center for

Earth System Science (DCESS), Niels Bohr Institute for Astronomy, Physics and Geophysics, University

of Copenhagen, Denmark); Mark R. Drinkwater (Oceans/Sea-Ice Unit, ESTEC Earth Sciences Division

(APP-FSO), Noordwijk, The Netherlands) 52

A Multi-Index Bayesian Approach to Climate Change Detection and Attribution, Klaus Hasselmann, Reiner

Schnur (Max Planck Institute of Meteorology, Hamburg, Germany) 53

Changes in statistical properties of daily precipitation in transient climate change experiments with a coupled

atmosphere-ocean GCM, V.A. Semenov and L. Bengtsson (Max Planck Institute for Meteorology,

Hamburg, Germany) 53

Near Two Centuries Monsoon Rainfall Variations across India: Reconstructed Past and Predicted 10-year Future,

N.A. Sontakke (Indian Institute of Tropical Meteorology, Pashan, India) 53

Statistical detection of the anthropogenic greenhouse signal in observed global and regional climate data fields, T.

Staeger, J. Grieser, C.-D. Schnwiese (Institut fr Meteorologie und Geophysik, J.W. Goethe Universitt

Frankfurt a.M.) 54

Trends in the Diurnal Temperature Range in the CCCma Coupled Model, Daithi A. Stone, Andrew J. Weaver

(School of Earth and Ocean Sciences, University of Victoria, Canada) 54

Detection of the causes of recent trends in radiosonde temperatures, P.W.Thorne1, P.D.Jones1, S.F.B.Tett2,

M.R.Allen3, D.E.Parker2, T.D.Davies1, and T.J.Osborn1 (1Climatic Research Unit, UEA, Norwich, UK.;

2

Hadley Centre for Climate Prediction & Research, Meteorological Office, Bracknell, Berkshire, UK;

3

Rutherford Appleton Laboratory, Chilton, Didcot, UK.) 54

Signal Analysis and Detection Studies Using Neural Networks, Andreas Walter (Institute of Meteorology and

Geophysics, Johann-Wolfgang Goethe University, Frankfurt, Germany) 55

Potentially Predictable Patterns, Xiaogu Zheng1 and Carsten S. Frederiksen2 (1 National Institute of Water and

Atmospheric Research, Wellington, New Zealand; 2Bureau of Meteorology Research Centre, Melbourne,

Australia) 55

Extratropical cyclone variability in the Northern Hemisphere winter from the NCEP/NCAR Reanalysis data, Olga

Zolina, Sergey Gulev, Sergey Grigoriev (P.P.Shirshov Institut of Oceanology, RAS, Moscow) 55

8

POSTER CONTRIBUTIONS TO THE 8 INTERNATIONAL MEETING ON STATSITICAL

CLIMATOLOGY 56

An Explanation for the Multidecadal, Winter Rainfall Decline Over South-Western Australia, Bryson C. Bates1,

Stephen P. Charles1, James P. Hughes2, and Edward P. Campbell3 (1CSIRO Land and Water, Wembley,

WA 6913, Australia; 2Department of Biostatistics, University of Washington, Seattle WA 98195, USA;

3

CSIRO Mathematical and Information Sciences, Wembley, WA 6913, Australia) 56

Statistical Downscaling to Monthly Precipitation in the Northwestern of South America (Colombian Region),

Nestor R. Bernal S. (Nacional University of Colombia) 56

Better Quality of StatisTical Downscaling Models Without Preceding PC-Analyses For Predictands? Example:

Rainfall Assessments for Namibia U. Beyer (Institute of Geography, University of Wuerzburg,

Wuerzburg) 56

Multiple Non-linear Statistical Model for Estimating Mean Maximum Urban Heat Island in a Medium-Sized

Town Szeged, Hungary, Zsolt Bottyn, Jnos Unger (Department of Climatology and Landscape Ecology,

University of Szeged, Szeged, Hungary) 57

Application of factor analysis for quantification of climate forcing processes in the Baltic Sea region, Arunas

Bukantis Dept. Of Hydrology and Climatology of Vilnius University, Vilnius Lithuania) 57

A New Approach of Interpolation and Prediction of Climate Time Series, Hong-Xing Cao, Xiao-Jing Jia (Chinese

Academy of meteorological Sciences, Beijing ,China) 57

Wind Field Retrieval From Scatterometer Data, Dan Cornford, David J. Evans and Ian T. Nabney 58

Homogenization and Quality Control in State Space Models, Thomas K. Friedli, Evi Schuepbach, Tania Steiner,

Jrg R. Hsler (Department of Mathematical Statistics and Actuarial Sciences, University of Berne, Bern,

Switzerland) 58

Statistical Analysis of Modern Climate Changes observed in the West and East Antarctic, Jagovkina, S. V.1,

Lagun, V. E.2 (1Main Geophysical Observatory, St.-Petersburg, Russia; 2Arctic and Antarctic Research

Institute, St.-Petersburg, Russia) 58

Statistical Forecasts of Summertime South American Stationary Systems (SASSY), Charles Jones1 and Leila M. V.

de Carvalhol2 (1 Institute for Computational Earth System Sciences (ICESS), University of California,

Santa Barbara, CA; 2 Dept. of Atmospheric Sciences, University of So Paulo, Brazil) 59

Using of cluster analysis in processing of correlation coefficients, Stanislava Kliegrov, Ladislav Metelka (Czech

Hydrometeorological Institut, Hradec Krlov Svobodn Dvory, Czech Republic) 59

A new look at the temporal structure of a time series, Ken-Chung Ko (Department of Geography, National

Kaohsiung Normal University, Kaohsiung, Taiwan) 59

Extreme value analysis applied on precipitation time series in the Czech Republic, Kveton, V. (Czech

Hydrometeorological Institute, Prague, Czech Republic) 59

Statistical climatology and interannual variability of the extratropical cyclone parameters in the global

atmosphere, Victor E. Lagun (Air-Sea Interaction Department, Arctic and Antarctic Research Institute,

St. Petersburg, Russia) 59

Spatial Organization of Decadal and Bidecadal Rainfall Fluctuation on Southern North America and Southern

South America, Omar Abel Lucero, Norma C. Rodriguez (Instituto Nacional del Agua and National

University of Cordoba, Cordoba, Argentina) 60

Interrelationship between Interdecadal Fluctuations in the Southern Oscillation and the Sea Surface Temperature

of the Tropical Pacific, Omar Abel Lucero, Norma C. Rodriguez (Centro de la Region Semirida (INA),

and National University of Cordoba.Cordoba, Argentina) 60

Variability of annual pressure and precipitation on Iberian Peninsula, M Dolores Manso Orgaz* and Liliana

Caramelo** (*Depart. of Physics. University of Aveiro, Portugal; ** Depart. of Physics University of

Trs-os-Montes and Alto Douro . Vila Real) 60

9

Reproduction of dry and wet Periods by Stochastic Daily Precipitation Generators, D. Marvan1, M. Dubrovsky2

(1Faculty of Math. and Physics, Charles University, Prague, Czech Republic; 2Institute of Atmospheric

Physics, Hradec Kralove, Czech Republic) 61

Semi-parametric statistical approaches for the analysis of climatological data, J. Mateu1, C. Alvarez2, T. Sanfeliu2,

M.M. Jordan3 (1 Department of Mathematics; 2 Department of Experimental Sciences, Universitat Jaume

I, E-12071 Castellon, Spain; 3 Department of Experimental Sciences, Universitat Miguel Hernandez,

Elche, Spain) 61

Statistical Methods in climatology: an application to the longest precipitation series in Portugal, Solange

Mendona Leite (Geophysical Centre of the University of Lisbon, Portugal) 61

Spatial patterns of the five-day rainfall all over North-Western Iberian peninsula, M. Andrs (Atmospheric

Physics Department, University of Salamanca, Spain); Clemente Toms (Atmospheric Physics

Department, University of Salamanca, Spain); Fernando de Pablo (Atmospheric Physics Department,

University of Salamanca, Spain); Solange Mendona Leite (Geophysical Centre of the University of

Lisbon, Portugal) 62

The effects of large scale phenomena on local climate variability over Romanian territory, Daniela Mihaela Paliu,

Roxana Bojariu, (National Institute of Meteorology, Hydrology and Water Management, Bucharest,

Romania) 62

Engineering, University of Puerto Rico, Mayaguez); Jorge Gonzalez (Department of Mechanical

Engineering, University of Puerto Rico, Mayaguez); Felipe Chong Rengifo (Department of Industrial

Engineering, University of Puerto Rico, Mayaguez) 62

Mid-term prediction of the daily maximum/minimum temperatures in South Korea using Dynamic Linear

Models, Keon Tae Sohn (Pusan National University, Korea); Jeong Hyung Lee (Dong-A University,

Korea); Byungsoo Kim (Inje University, Korea) 62

Statistics in Agricultural Climatology adapting a workshop for African needs, S. M. Gathara, C. Oludhe, S. B.

Otengi, P. K. Runanu, R. D. Stern 63

Changes of large-scale atmosphere circulation during XX century and their influence on the climate in Ukraine,

V.F. Martazinova, T. Sverdlik (Ukrainian Hydrometeorological Institute, Kiev, Ukraine) 63

Validating a Downscaling Model of the Stochastic Structure of Daily Temperature, Marina Timofeyeva1, Rachael

Craig2 (1UCAR, NOAA/NWS, Silver Spring, MD; 2Dept. of Geology, Kent State University, Kent, OH) 63

Statistical Decomposition of European Temperature Time Series, Silke Trmel, Jrgen Grieser, C.-D. Schnwiese

(Institut fr Meteorologie und Geophysik, Universitt Frankfurt, Frankfurt /M., Germany) 64

Coupled sea ice and atmosphere variability in the Weddell Sea, Antarctica, Silvia A. Venegas (Danish Center for

Earth System Science (DCESS), Niels Bohr Institute for Astronomy, Physics and Geophysics, University

of Copenhagen, Denmark); Mark R. Drinkwater (Oceans/Sea-Ice Unit, ESTEC Earth Sciences Division

(APP-FSO), Noordwijk, The Netherlands) 64

10

MONDAY 12.3.2001

Plenary session: Stochastic Climate Models

9:00 Greg Holloway

9:45 Comment: Jin-Song von Storch

10:00 Comment: Andreas Hense

11:15 Norm McFarlane

11:45 Lunch

13:00 Poster introduction

14:30 G.V. Alekseev:

To the nature of red noise in the climatic series

14:50 P. Carl:

Frequency Modulation in the Instrumental Record

15:10 Anatoly I. Chavro:

Statistical algorithms for solving inverse problems on predictability of regional climate

15:30 Carsten S. Frederiksen:

Validating Interannual Variability in Atmospheric General Circulation Models

The character of prediction utility in dynamical systems relevant to climate and weather

16:40 Till Kuhlbrodt:

Stochastic dynamics in a simple model of seasonal open ocean deep convection

17:00 Frank Kwasniok:

Prediction of ENSO using an optimized mixture of linear models

17:20 Adam Monahan:

Simple Stochastic Models of the Thermohaline Circulation

14:30 Sylvia Antunes:

Improvement of prediction using SSA methods in addition to MEM: Case of NAO study

14:50 Ulrich Callies:

On the interrelationships between multisite climatic temperature forecasts obtained from

downscaling regional sea-level-pressure fields

15:10 Peter Challenor:

Model Error and Uncertainty

15:30 Youmin Chen:

The Winter Temperature Downscaling Experiments Using Canonical Correlation Analysis in

China

On-line Gaussian Processes and their Potential in Data Assimilation

16:40 Traute Crueger:

Statistical Downscaling of Largescale GCM Data to Local Ice Accumulation in Greenland

17:00 Mojca Dolinar:

Spatialisation of 24 hours precipitation with 100 years return period

17:20 S. El Adlouni:

Bayesian estimate of halphen type a law's parameters

TUESDAY 13.3.2001

11

Plenary session: Statistics, Models and Data Assimilation

9:00 Geir Evensen

9:45 Comment: Carlos Pires

10:00 Comment: Christian Lajaunie

11:15 Jean-Paul Chils

11:45 Lunch

13:00 Juan Pedro Montavez:

EMAD: An Empirical Model of the atmospheric Dynamics

13:20 Nityanand Singh:

On the Natural Regularity and Predictability of Terrestrial and Extra-Terrestrial Parameters

13:40 Jun-Ichi Yano:

Stochastic Tropical Atmospheric Boundary-Layer Dynamics: Convective Downdrafts and

Recovery, Discharge-Recharge Processes, and 1/f-Noise

13:00 Christoph Gebhardt:

Variational Assimilation of botanical proxy Data for the Reconstruction of Palaeoclimate

13:20 Nikolaus Groll:

How realistic is GCM-simulated precipitation?

13:40 Andreas Hense:

Homogenisation and Interpolation of incomplete climatological Data with simple physically

based Models

14:00 Ramon A.Quintana-Gomez

Trends of Maximum and Minimum Temperatures in Northern South America

14:20 Heike Hoffmann

Investigation of Spatial and Temporal Ozone Immission Structures under Stochastic-

Deterministic Model Assumptions

14:40 I. Hoteit

Reduced-order schemes for data assimilation in oceanographiy based on the Kalman filter

Cloudiness and sea surface temperature: tackling the problem of missing values

15:50 Radan Huth

Removing errors in simulated distributions of climate elements using a non-parametric

adjustment procedure

16:10 Ian Iolliffe

Simplified EOFs - three alternatives to rotation

16:30 J.M. Jones

Reconstruction of large-scale climate

16:50 Gennady A. Kivman

On a probabilistic interpretation of variational data assimilation and model error estimation

17:10 Leonid Lopatoukhin

Statistical presentation of wind wave climate

17:30 J. Mateu

Model-based spatial modelling of climatological data

13.00 Daniel Gellens

Regional GEV-growth curves of yearly precipitatation depths in Belgium for durations

ranging from 10 minutes to 30 days

12

13.20 Leila M. V. Carvalho

Extreme precipitation events in southeastern South America and the relationships with

summertime large-scale South American Stationary Systems (SASSY)

13.40 Shaleen Jain

Diagnosis of variations in the climate-related flood risk - Some results from the Interior

Western United States

14.00 Tiba Kabanda

The Climatological Probability of Drought in Northern Province of South Africa

14.20 S.S. Kandalgaonkar

Statistical Analysis of Point Discharge Current at Different Time Scale Over the Pune Region

14:40 S.S. Kandalgaonkar

A Dedacal Study of Premonsoon Seasons Thunderstorms Over Pune Region by Using Markov

Chain Model

15.00 Xiolan Wang

Trends of Atlantic wave extremes as simulated in a 40-year wave hindcast using kinematically

reanalyzed wind fields

Hurricane Landfall Winds Along the United States Gulf and Atlantic Coasts: Monte Carlo

Simulations

16.10 Eva Bauer

Simulation of ocean wave climate changes through wind variance changes on different

timescales

16.30 P.Lionello

Frequency and Intensity of Mediterranean Cyclones in a doubled CO2 scenario

16:50 P.Lionello

Extreme waves and surge events in a doubled CO2 scenario in the Adriatic Sea

17.10 Niklas Sjblom

Extreme value analysis of absolute minimum temperature and minimum dew-point

temperature at Sodankyl, Finland

17.30 Won-Tae Kwon

Observed Changes in Temperature Extremes of Korea

WEDNESDAY 14.3.2001

9:00 Adri Buishand

9:45 Comment: Hans Alexandersson

10:00 Comment: David Stephenson

11:15 Richard Chandler

11:45 Rick Katz

12:15 Lunch

13:30 Miroslaw Mietus

The role of atmospheric large-scale circulation over Europe and North Atlantic in sea level

variability on the Polish coastal zone

13:50 S. De Iaco

Space-time variograms and a functional form for total air pollution measures

14:10 Carlos Pires

13

Performance of four-dimensional variational assimilation and dynamical instability,

experiments on a quasi-geostrophic model

14:30 Budong Qian

A multisite self-consistent Weather Generator for Impact Studies in Portugal

14:50 Joao Rio

Monthly mean forecasts of rainfall and temperature in Portugal

15:10 Radhika Ramachandran

Sensitivity Studies on Wind Measurement Errors in Mesoscale Model Simulations

Building the bridge between sequential data assimilation and geostatistics: the stationarity

issue

16:20 Leonard A. Smith

Public Debate and Climate Model Products: Whereof must one be silent?

16:40 Tamas Szentimrey

Multiple Analysis of Series for Homogenization (MASH). Seasonal Application of MASH.

Automatic using of Meta Data

17:00 Clemente A. S. Tanajura

A data assimilation method with a coupled ocean-land-atmosphere model and its impact on the

ocean state and on medium-range predictions

17:20 Karl E. Taylor

Assessing Model Skill in Climate Simulations

13:30 Janos Mika

XXth Century Behaviour of Temperature and Precipitation Extremes in Hungary

13.50 Lucie Vincent

Homogenized Daily Temperatures for Trend Analyses in Extremes over Canada

14.10 Joanna Wibig

The analysis of extreme temperature distributions on the example of Poland

14.30 Jan Kysel

Comparison of 20- and 50-year return values in GCM-simulated, downscaled and observed

temperature series

14.50 Jaroslava Kalvov

Test on the probability distribution of maximum air temperatures

15.10 Luis Gimeno

Three Approaches to determine extreme years of Global Atmospheric Temperature

16.00 E. P. Campbell

Bayesian Statistical Modelling of Nonlinear Climate Processes

16.20 Jess Abaurrea

Modelling and characterizing extreme drought events

16.40 Jrgen Grieser

Extremes in Instationary Time Series

17.00 Henrique Oliveira Pires

Comparison of Several Methods for the Estimation of Extreme Value Distribution in Time

Series with Trends

17.20 Philippe Naveau

The Possible Influence of Large Volcanic Eruptions on El Nino

14

Session 4: Simulation Models

13:30 Ulrich Achatz

Simplified atmosphere models with realistic variability

13:50 Konrad Bogner

Modelling of Temporal Rainfall by the Use of Stochastic Point Process Models in View of

Scaling

14:10 Harold E. Brooks

Statistical Modelling of Tornado Occurrence

14:30 Udo Busch

A method to extend the temporal validity of nested regional climate simulations

14:50 Aristita Busuioc

Stochastic generation of daily precipitation

15:10 Stephen Charles

Statistical Downscaling of Daily, Multi-Site Precipitation in Tropical and Sub-Tropical

Climates

Stochastic Weather Generator Model for Simulation of Multiple Meteorological Time Series

16:20 Martin Dubrovsky

Interdiurnal and interannual variability in stochastic daily weather generator:

Modelling and the role in agricultural and hydrologic studies

16:40 R. Garcia

Temperature predictability in the greater Mediterranean area

17:00 Friedrich Wilhelm Gerstengarbe

A new validation scheme for the evaluation of climate model outputs

17:20 Clare Goodess

The construction of multi-site and multi-variate scenarios for Mediterranean regions based on

Monte Carlo simulations with a conditional weather generator

THURSDAY 15.3.2001

9:00 Jim Hughes

9:45 Comment: Francis Zwiers

10:00 Comment: Hans von Storch

11:15 Ed Waymire

11:45 Lunch

13:00 Martin Widmann

How realistic is GCM-simulated precipitation?

13:20 Abebe Yeshanew

Spatio-temporal patterns of Meteosat Satellite Thermal Infrared Derived Convective Clouds

over Ethiopia as Described by PCA & MVC

13:40 Francis Zwiers

Skill of seasonal hindcasts as a function of the ensemble size

13.00 Monika Lakatos

Determining extreme sets in multivariate distributions

15

13:20 Albert R. Boehm

The Probability Of An Extreme In A Correlated Time-Space Domain

13.40 Peter Bissolli

Climate extremes in Germany and its relation to the objective weather type classification of

the German Weather Service

14.00 Judit Bartholy

Statistical methods and case studies investigating extreme climatology of the Carpathian Basin

13:00 Anders Grimvall / Asa Forsman

Simulation of multivariate time series of meteorological data

13:20 Andreas Hense / Martin Kappler

Modelling the probability of precipitation using the Marcov chain MonteCarlo technique

13:40 Radan Huth

Reproduction of variance in statistical downscaling: inflation vs. noise addition

14:00 Anders Nordgaard

Simulating meteorological data by block resampling

14:20 Guillermo Obregon

The Amazon climate: A multiscale and nonstationary processes

14:40 Arnt Pfizenmayer

First indication of anthropogenic climate change in local wave conditions in the North Sea

Ensemble Predictions with a Limited-Area Mesoscale Atmospheric Model - Implications for

the Interpretation of its Direct Model Output

15:50 Marina Timofeyeva

Modeling stochastic structure of local climate as downscaling of GCM fields

16:10 Hans von Storch

Numerical experimentation with regional atmospheric models

13:00 Myles Allen

Exploiting the emerging signal of anthropogenic climate change in probabilistic climate

forecasting

13:20 Ajay Singh

Significance of Anthropogenic Radiative Forcings in Simulation of Temporal Pattern of

Surface Air Temperature in Monsoon Climate

13:40 Katrine Krogh Andersen

Decadal variations of Greenland precipitation over the past 1400 years

14:00 Auguste Boissonnade

Lessons taught from the Homogenization of 200 US First Order Stations

14:20 Roxana Bojariu

North Atlantic osccillation projection on regional climate predictability

14:40 Bernard Cornlis

What influences decision-makers in drought prone areas when allocating water resources?

Assessing uncertainty in anthropogenic signal estimates using space-time and trend-based

optimal detection approaches

15:50 Stephen Jewson / Anders Brix / Rodrigo Caballero

Applications of Statistical Climatology in Finance

16:10 Christopher J. Paciorek

Trends in northern hemisphere winter cyclone activity based on a comparison of cyclone

indices

16

16:30 Heiko Paeth

Signal analysis of the atmospheric mean temperature 500/1000 hPa north of 55 N between

1949 and 1994

16:50 James Risbey

Expert Assessment of Uncertainties in Detection and Attribution of Climate Change

17:10 Christine Ziehmann

A Maximum Entropy Approach for Extracting Forecast Probabilities from Finite Ensembles

FRIDAY 16.3.2001

9:00 Granger Morgan

9:45 Comment: Myles Allen

10:00 Comment

11:15 Mark Berliner

11:45 Chris Forest

12:15 Lunch

13.20 Silvia A. Venegas

Coupled Sea Ice and Atmosphere Variability in the Weddell Sea, Antarctica

13:40 Reiner Schnur

A Multi-Index Bayesian Approach to Climate Change Detection and Attribution

14:00 V.A. Semenov

Changes in statistical properties of daily precipitation in transient climate change experiments

with a coupled atmosphere-ocean GCM

14:20 N.A. Sontakke

Near Two Centuries Monsoon Rainfall Variations across India: Reconstructed Past and

Predicted 10-year Future

14:40 Tim Staeger

Statistical detection of the anthropogenic greenhouse signal in observed global and regional

climate data fields

Trends in the Diurnal Temperature Range in the CCCma Coupled Model

15:50 P.W. Thorne

Detection of the causes of recent trends in radiosonde temperatures

16:10 Andreas Walter

Signal Analysis and Detection Studies Using Neural Networks

16:40 Xiaogu Zheng

Potentially Predictable Patterns

17:00 Olga Zolina

Extratropical cyclone variability in the Northern Hemisphere winter from the NCEP/NCAR

Reanalysis data

17

inference on respective 'castings') calls for high-re-

Stochastic Climate Models solution, quantitative, joint analyses of multiple sig-

nal attributes among which the dimensions of mo-

dulation must not be neglected. Adaptive waveform

To the nature of red noise in the climatic series,

("Matching Pursuit"; MP) analysis using a highly

G.V. Alekseev (Arctic and Antarctic Research In-

redundant dictionary of frequency-modulated

stitute, St.-Petersburg, Russia)

Gabor atoms has been applied to this end to a

representative set of climatic time series of the

It is considered the forming of low-frequency com-

19/20th centuries. FM-MP "structure books" cove-

ponents in the climatic spectra of atmospheric cha-

ring all parameters of these (non-orthogonal) de-

racteristics that was named low-frequency weather

compositions - a sort of empirical time series mo-

noise. The base of approach is statistical model of

dels in a seven-dimensional signal space - are pro-

weather variability that include individual weather

vided and discussed. Comparison is made to the

disturbances (IWD). This IWD are re-presented

WT which yields a time-scale energy distribution

with help of self-model function including constant

'as is', and to the STFT which provides a spectro-

shape function and set of IWD parameters

gram that relies (as the FT heavily does) on mutual

(amplitude, duration, time of IWD start). It is

interference of Fourier modes. Like the STFT, the

known that spectrum of a process described of such

MP method interprets the energy density distri-

model depends on spectrum shape of individual

bution in terms of its dictionary, but overcom-

disturbance. On the base of experiments with

pleteness of the latter and intrinsic provision of

proposed model was shown that as a rule spectrum

nonstationarity relaxes massive interferences to

of many IWD with duration T has a maximum at

modest here. Unlike WT, the MP approach bears

periods much more than T. Averaging of time

"scale" as a truly independent parameter, resulting

series including IWD amplifies this effect. In this

in localizations that only depend on the scale of a

connection it should be noted that low-frequency

(partial) signal, not on its specific position in the

rise of climatic spectra could be more consequence

fre-quency domain; spectrograms provide high re-

of transfer of local structures energy to the low-

solution throughout.

frequency part of spectra in response to linear pro-

Deep FM is generally found to fit best into WT

cedures of spectral analysis and averaging than real

energy density distributions of the instrumental re-

low-frequency variability.

cord as far as considered. This is a surprise which

might underscore the potential topological aspects

mentioned. Though the signal space has to be fur-

Frequency Modulation in the Instrumental Re-

ther extended to include harmonic amplitude modu-

cord, P. Carl (Institute for Freshwater Ecology

lation (AM) and modulated drifting modes at least,

and Inland Fisheries, Research Association

and the high demand on resolution poses resource

Berlin of the Leibniz Society, Berlin, Germany)

problems which call for massively parallel

computation, another surprise at the present stage

Multiple cyclic forcing and the climate system`s

of analysis is a clear phase coincidence between

internal oscillatory capabilities (or nature), from

leading modes of both insolation and temperatures

intraseasonal to decadal and longer, may bear

since the 70th. The phase of leading 'slow' modes

complex, evolving modal structures as found in

points to decadal scale cooling ahead (but a con-

energy density spectrograms (resp. 'scalograms`)

servative projection is intrinsic to the method). The

throughout the instrumental record. Customary

problem of Global Warming is 'preserved', how-

methods used to visualize features in the time-fre-

ever, in second-rank modes of growing signal

quency plane, including "Short-Time" (windowed)

energy which should be taken seriously with a view

Fourier and Wavelet Transforms (STFT, WT), are

on their next warming phase.

suited to provide a qualitative view of these evo-

lutions. The question arises to what extent such

apparent structural nonstationarities reflect anhar-

monic (maybe anthropogenic) climate drifts and/or Statistical algorithms for solving inverse problems

just higher-order cyclostationarity comprising the on predictability of regional climate, Anatoly I.

frequency domain itself via systematic frequency Chavro, Yegor V. Dmitriev (Institute for

modulation (FM). As a basic conception of commu- Numerical Mathematics RAS, Russia)

nication, FM of a carrier mode might be a principle

New algorithms for solving inverse problems on the

at work in climatic signal transmission as well---

restoration of small-scale structure of geophysical

and might even hint at topological constraints (i.e.

fields in separate regions by large, space- and time-

geometrical structures in the back) of the system`s

averaged fields are proposed. The algorithms are

phase space trajectories. Recent paleoclimatic

multi-dimensional statistical models for forecasting

studies hint at harmonic FM of Milankovich cycles,

small-scale structure of most probable anomalies of

and the challenging departure since about the 40th

geophysical fields in separate regions using space-

from apparent astronomical control of the phase of

averaged values of global fields (or low-frequency

the seasonal cycle, as found in temperature records

component part of the spectrum), which can be

of the instrumental period, deserves further clari-

predicted by the general circulation models with

fication.

rough resolution.

Reconstruction from time series of the score behind

the 'sounds' of climate and climatic change (and

18

The problems are solved in the frameworks of line- A new parameter of dynamical system predicta-

ar models. It was supposed that anomalies of regio- bility is introduced which measures the potential

nal climate of field f are connected with global ave- utility of predictions. It is shown that this parameter

raged field x as x =Af+n (1) where A is a linear satisfies a generalized second law of thermody-

operator and n is the error of such representation. namics in that for Markov processes utility declines

An operator R is constructed on the base of a priory monotonically to zero at very long forecast times.

information such that acting on equation (1) one Expressions for the new parameter in the case of

can get a best (in a sense of mean-square error) esti- Gaussian prediction ensembles are derived and a

mation of the regional field f=Rx. useful decomposition of utility into dispersion

To construct R we used the methods of reduction (roughly equivalent to ensemble spread) and signal

[1] or statistical regularization [2].The proposed components (roughly equivalent to ensemble mean

methods also allow to control the reliability of sta- squared) is introduced. Earlier measures of predic-

tistical models for solving the inverse problem. The tability are shown to have only considered the dis-

results of numerical experiments of the restoration persion component of utility. A variety of simple

of the regional values of such geophysical fields as dynamical systems with relevance to climate and

mean surface temperature, high-frequency compo- weather prediction are introduced and the behavior

nent of H500 field and precipitation. The proposed of their potential utility is analyzed in detail. For

methods allow to restore over 70% of anomalies of the (stochastic) climate systems examined here, the

geophysical fields using the large-scale values pre- signal component is at least as important as the

dicted by the general circulation models. dispersion in determining the utility of a particular

set of initial conditions. The simple ``weather'' sys-

tem examined (the Lorenz system) exhibited diffe-

Validating Interannual Variability in Atmospheric rent behavior with the dispersion being more im-

General Circulation Models, Carsten S. portant than the signal at short prediction lags. For

Frederiksen1 and Xiaogu Zheng2 (1Bureau of longer lags there appeared no relation between

Meteorology Research Centre, Melbourne, utility and either signal or dispersion. On the other

Australia; 2National Institute of Water and hand there was a very strong relation at all lags

Atmospheric Research, Wellington New Zealand) between utility and the location of the initial con-

ditions on the attractor.

This paper is an effort to explore general statistical

procedures for validating the inter-annual variabili-

ty in ensembles of atmospheric general circulation Stochastic dynamics in a simple model of seasonal

model simulations forced by observed sea surface open ocean deep convection, Till Kuhlbrodt

temperatures (SSTs). The method relies on being (Potsdam Institute for Climate Impact Research,

able to decompose the interannual variance of a Climate System Dept., Potsdam, Germany)

seasonal mean of daily meteorological values into

three parts: a forced component, a low frequency Aspects of open ocean deep convection variability

internal source component and a weather-noise are explored with a two-box model of convective

component. How well models reproduce the inter- mixing in a water column. In order to locate the

annual variability of the components of seasonal model in a region of parameter space relevant to the

means is a topic of great interest. Ideally, in asses- real ocean, it is fitted to observational data of the

sing, or validating, a model's simulation of climate Labrador Sea. Seasonal cycles in the boundary

variability, the emphasis should be on the extent to conditions allow for a realistic model fit without

which the forced component of variation is correct- changing the basic dynamics of the model. The

ly simulated since this is potentially predictable. model suggests that Labrador Sea in the years

Here, we propose an estimation method for the 1964--1974 is close to a bistable regime where con-

correlation coefficient between the SST-forced vection is either `on' or `off'. The role of high-fre-

components of simulated and observed seasonal- quency atmospheric variability in triggering con-

means. The technique has been applied to DJF and vection events is simulated by adding a noise term

JJA seasonal means of 200hPa geopotential heights to the surface buoyancy forcing. Depending on the

using daily data from an ensemble of 4 runs of the underlying deterministic model state, the effect of

Bureau of Meteorology Research Centre (BMRC) the added noise is very different. First, when the de-

AGCM for the period 1958 - 1991. We have taken terministic model state is convective, the noise

as our "proxy" observed data the daily 200hPa leads to shorter convection events that are, how-

geopotential height field from the National Centre ever, more efficient in cooling the deep ocean.

for Environmental Prediction (NCEP) and National Second, when the deterministic state is non-

Centre for Atmospheric Research (NCAR) reana- convective, the noise induces convection events,

lyses over the same period. and for an optimal noise level a maximum vertical

density gradient is achieved. Third, for the bistable

deterministic state the noise leads to jumps between

The character of prediction utility in dynamical the two model states of annual convection and

systems relevant to climate and weather, Richard stable stratification. Here, the dependence of the

Kleeman (Courant Institute for Mathematical mean residence times in either state on model

Sciences) parameters is studied. This reveals a mechanism for

19

scale interaction between the synoptic and the (1Instituto de Meteorologia, R. C ao Aeroporto,

decadal timescale. Lisboa, 2Universidade de Aveiro)

The results highlight the crucial role of the high-

frequency surface forcing for the ocean's long-term Singular Spectral Analysis (SSA) is a method that

variability and its climatological mean state. allows the separation of series in isolated compo-

nents some identified as trend, other as components

with characteristics of periodicity. Using selected

Prediction of ENSO using an optimized mixture of significative components, grouped or isolated, this

linear models, Frank Kwasniok (Leibnitz-Institut method includes the reconstruction of series.

fr Atmosphrenphysik, Khlungsborn, Reconstructed components are noisily reduced and

Germany) present a regular behaviour and these are the cha-

racteristics that permit the improvement in pre-

A novel statistical prediction scheme for ENSO is diction.

presented. The predictor-predictand relationship is The method consists in fitting AR predictors esti-

modelled by a mixture of linear regression models. mated by Maximum Entropy Methods (MEM) to

Each model is associated with a probability distri- selected reconstructed components rather to the ori-

bution describing its domain of influence in phase ginal series. In order to guarantee the independence

space. The joint probability distribution of predic- of data the predictors are estimated using half of the

tors and predictands is approximated by interpo- series and tested in the remaining part. The pro-

lation among the models. The locations and the cedure is applied separately for each of the com-

prediction coefficients of the models are ponents and the final forecast is obtained by

determined by maximizing the likelihood of summing the forecasts corresponding to each com-

observing a given learning data set using an ponent.

expectation-maximization procedure. The North Atlantic Oscillation index, obtained u-

Unlike many other techniques, the method offers sing the Island pressure series and the recently

forecasts of prediction errors. Two applications of corrected pressure series of Lisbon is used to per-

the concept described above to prediction of sea form this study.

surface temperature indices are presented. Firstly, a

linear ENSO prediction model is constructed from

observational data which takes into account the On the interrelationships between multisite clima-

annual cycle by conditioning the regression tic temperature forecasts obtained from down-

coefficients on the phase of the annual cycle. Se- scaling regional sea-level-pressure fields, Ulrich

condly, a state dependent modelling approach is Callies (Institute for Hydrophysics, GKSS Re-

tried, i. e., the model coefficients depend on some search Center; Geesthacht, Germany)

ENSO-related variables. A cross-validated predic-

tion skill is estimated and compared with that of A case study of evaluating the information content

other prediction schemes. of climatic temperature forecasts at four European

stations, obtained from downscaling regional sea-

level-pressure information, is presenteded. Fore-

Simple Stochastic Models of the Thermohaline casts are informative in the sense of this study if a

Circulation, Adam Monahan (Humboldt- great proportion of observed temperature variability

Universitaet zu Berlin) can be attributed to these forecasts subsequent their

post-calibration based on a linear regression model.

Since the pioneering work of Stommel, simple As observations at different locations are generally

deterministic models have played a central role in correlated, each forecast delivered for one

shaping our understanding of the thermohaline cir- particular site will be informative about other sites

culation (THC). In particular, THC dynamics is as well. Even when the respective local forecasts

often analysed within the context of dynamical sys- are found to be best adapted to the conditions at a

tems theory. The presence of even weak stochastic particular station, it will generally be benecial to

perturbations can qualitatively alter the character of establish a multiple-calibration scheme that

variability in simple models of the THC. This talk involves additional information from forecasts

will address some of these effects of stochasticity, delivered for other locations. The technique of

including relative stabilisation of regimes and fitting conditional mo-dels is used to measure

stochastic induction of sensitivity to initial condi- incremental information obtained by consulting

tions. several forecasts. Its application also allows the

determination whether additional information

gained from forecasts delivered for other sites is

genuine or if it is merely mediated by correlated

Statistics, Models and Data observations.

Assimilation Results of conditional independence modelling are

con-trasted with the results of other techniques that

describe the inter-relationship between data by in-

Improvement of prediction using SSA methods in troducing artificial variables obtained from opti-

addition to MEM: Case of NAO study, Sylvia mum coordinate transformations (empirical ortho-

Antunes1, H. Oliveira Pires1,2 e Alfredo Rocha2 gonal function and canonical correlation analyses).

20

the impact of topography on downscaling skill. The

Model Error and Uncertainty, Peter Challenor downscaling model partially capture the tempera-

(Southampton Oceanography Centre, ture trend change even when the raw data are de-

Southampton) trended, but a better-simulated trend can be reached

if the raw data trends are retained for developing

Climate research is increasingly dependent upon models, which can lead to a better downscaling

com-plex numerical models. While such models skill.

give reasonable estimates of the state of the climate

system as yet we do not have any measure of un-

certainty, such as confidence limits, on such esti- On-line Gaussian Processes and their Potential in

mates. One approach is to run the models many Data Assimilation, Dan Cornford, Lehel Csato

times and use the resulting ensemble to evaluate the and Manfred Opper (Computer Science, Aston

uncertainty. However realistic models are expen- University, Birmingham)

sive to run and it is impossible to create ensembles

that are large enough to allow the uncertainty to be Recently, methods have been proposed that use on-

evaluated properly. In this paper an alternative is line algorithms for training Gaussian processes,

proposed. A Bayesian approach is taken and the based on a Bayesian probabilistic framework. The

output of the climate model is represented by a online method is an iterative algorithms for upda-

Gaussian process. A limited number of model runs ting the mean function and covariance kernel,

are needed to characterise the Gaussian process which removes the requirement for large matrix in-

which is then used to calculate the probability version. When combined with a sparsity heuristic,

density function (pdf) of a model output given the the algorithm has been shown to produce small ge-

pdf of some model inputs (which could be the for- neralisation errors on standard regression and

cing function or uncertain internal model para- classification problems. In this paper we illustrate

meters). The technique is illustrated with a simple the application of this method for assimilating scat-

climate model (where we can run a large enough terometer observations to retrieve wind fields using

ensemble to validate the method). The extension of an online algorithm. We show how the algorithm

such methods to large state-of-the-art models is may be adapted to cope with the multi-modal

discussed. nature of the distribution of the local wind vector

given the local scatterometer observations. The

local wind vector probability distribution is

The Winter Temperature Downscaling Experi- modelled using a mixture density network. Since

ments Using Canonical Correlation Analysis in the method used will track a local Gaussian mode

China, Youmin Chen (Agrometeorology Institute, in the posterior distribution of the wind field given

Chinese Academy of Agric Sciences, Beijing, the scatterometer observations, we show how the

China) algorithm could be used to sample from this

posterior distribution. We go on to show how

High quality monthly temperatures at 147 stations sparsity can be used to select a restricted set of

in China and the NCAR/NCEP reanalysis circula- spatial locations, which capture the variability in

tion data during winter (the average of Dec. Jan. the full wind field model but reduces redundant

Feb) of 1951/52-1997/98 are applied in this paper. computational and storage requirements. We

With the canonical correlation analysis (CCA) tech- suggest this may be a useful pre-processing tool for

nique being employed, the downscaling model with general data assimilation systems where the

East Asia circulation as predictor and the monthly observations have spatially variable redundancy,

temperature as predictand were established. Some which is often the case for remotely sensed

experiments with different PC truncations and observations. By reducing the number of data

different CCA modes retained were conducted in points required to represent the field in question in

order to find the optimum downscaling model. It is a pre-processing step more time is available for a

found that as many as possible PC truncation and 4- more complete data assimilation system.

6 CCA modes can make the better model skill. The

optimum model was found with 13 PC truncation in

predictors and 8 PC truncation in predictand and Statistical Downscaling of Largescale GCM Data

also 5 CCA modes retained. The validation results to Local Ice Accumulation in Greenland, Traute

with cross-validation method were compared be- Crueger, Hans v. Storch (GKSS Research Center,

tween the models with and without detrending to Geesthacht, Germany)

the raw data. There are markedly different model

skills for different sub-area of this country, chan- Object of the work is the creation and investigation

ging from 10% to 70% in Brier-based score. The of local 'Artificial Ice Core' accumulation estimated

distribution of model skill score is rather consistent from GCM data. Using GCM runs performed with

with topography characteristics, in that the northern different CO2 concentrations and solar forcing, the

China and eastern China have a relatively higher natural and anthropogenic influence on ice accu-

skill for its easy impact subject to circulation, while mulation shall be investigated. For estimation of the

in the area where the sharp orographic relief can artificial ice accumulation a statistical downscaling

impede airflow from Siberian high, the model skill model is developed. In this presentation the statisti-

is low. Therefore the model results indirectly reflect

21

cal technique and an application for one north * 50 stations with precipitation registration in-

Greenland ice core are described. struments (precipitation is measured every five

The statistical technique is based on a multiple line- minutes)

ar regression using yearly accumulations as pre-

dictant and principle components (PCs) of seasonal Gumbel method was used for calculating 24-hours

averaged largescale fields as predictors. In order to precipitation with 100 years return period for all

separate temperature from dynamic effects on pre- stations with registration instruments. As the first

cipitation the streamfunction (500hPa) and tem- step we analysed the correlation between 24-hours

perature (500hPa, 700hPa) fields are tested as pre- precipitation with 100 years return period and

dictors. extreme daily precipitation. The preliminary spatial

We vary the boundaries and the months of the analysis was made to find out the dependence of

fields, whose mean are used for the EOF expansion. extreme precipitation on some geographical varia-

This way the patterns with the highest correlations bles. In general, a positive trend was evident on

between their PCs and the ice accumulation are altitude. The subregions were defined regarding the

found. The PCs of these patterns are used as pre- residuals (difference to station measurements) of

dictors for the multiple linear regression. The para- regression with altitude.

meters for which the explained variance is highest Different methods of spatial interpolation were

(after cross-validation) are part of the final form of tested:

the model.

The data used to develop the model are the monthly * cokriging - taking the maximum daily pre-

NCEP-Reanalysis and yearly accumulations of cipitation as covariable

Greenland ice cores from 1948-1992. The cores * universal kriging - taking the maximum daily

were obtained and analyzed by the Alfred Wegener precipitation as external drift

Institute for Polar and Marine Research Bremer-

haven, Germany (AWI). For one core from the Cross-validation technique was used to determine

northern slope of the Greenland ice sheet the the appropriate variogram model. The kriging pre-

greatest part (65% explained variance after dictions were calculated in the grid with 1 km reso-

validation) of the yearly variability of the ice core lution. We used the altitude data from the digital re-

accumulation is described by the third and second lief model with the same grid. For the kriging pre-

EOF pattern of the 500hPa streamfunction field dictions we considered different influential sur-

averaged from May to August. The area of the roundings and the most efficient was chosen with

streamfunction field reaches from northeast Ameri- cross-validation technique.

ca to west Asia. The patterns correspond well to the

synoptic conditions leading to precipitation in

northern Greenland investigated otherwise. The Bayesian estimate of halphen type a law's para-

influence of the temperature on ice accumulation in meters, S. El Adlouni1, A. Bouziane, D. Ouazar,

northern Greenland seems to be small. A. Zoglat1 (1 Faculte des Sciences-Rabat; Ecole

An 'Artificial Ice Core' is estimated by applying the Mohammadia d'Ingenieurs, Rabat, Morroco)

regression model to a 1000 year coupled atmos-

phere/ocean GCM control run. The Halphen distributions family has some sta-

tistical properties that make them appropriate for

modeling some hydrological series. In particular,

Spatialisation of 24 hours precipitation with 100 the Halphen type A law is a useful tool in extreme

years return period, Mojca Dolinar1, Damijana value analysis.

Kastelec2 (1Hydrometeorological institute of Despite of the large set of distributions used in

Slovenia, Ljubljana, Slovenia; 2University of hydrology, none is judged globally adequate for

Ljubljana, Biotehnical Faculty, Ljubljana, adjusting hydrological series. The 3-parameter dis-

Slovenia) tributions (Halphen type A, B, and B-1) developed

by Halphen (1941) meet some specific require-

The information about heavy precipitation amount ments for fitting these series. However, the litera-

and the probability of its occurrence is valuable ture about these distributions is very restrictive and

especially in hydrology and civil engineering. Since it seems that, except the papers by Halphen (1941),

the network of precipitation registration stations is Halphen (1955) and Morlat (1956), there is no work

not dense, it is very difficult to provide re- that has rigorously explored in detail their mathe-

presentative data to the customer. In order to be matical and statistical properties.

able to respond on customer requests we wanted to In their paper Perreault et al. (1997) present a com-

make a digital map of 24-hour precipitation with plete study and pointed out some interesting mathe-

100 years return period with resolution 1 km. matical and statistical properties of the Halphen

For analysis we used the data from type A distribution. This is a member of the expo-

nential distributions family, so that the maximum

* 360 stations with daily precipitation measure- likelihood estimation of its parameters is appro-

ment (precipitation is measured once per day at 7 priate and easy to proceed. As a measure of esti-

a.m.) mators accuracy we consider the bias and stan-

dardized error. These two statistics involve the

parameters and their estimators, we reviewuate

22

them using simulations. In this work we give a modelling-extrapolation-prediction (or TS-MEP)

method for generating Halphen type A distributed process. Perhaps equally important component of

samples. Using these samples we study some this method is the development of effective filtering

parameters estimators properties. In the second part module. The filtering mechanism should be such

we use a Halphen type A law with a Bayesian that it effectively suppresses the high frequency, or

framework, to improve the estimations of the maxi- unpredictable, variations and carves out the low

mum likelihood, especially for samples with small frequency mode, or predictable, variation of the

size. To compute bayesian estimators we uses Mon- given series. By incorporating a possible solution to

te Carlo Markov Chains algorithms (MCMC). To those propositions a new TS-MEP method has been

conclude we point out some pertinent remarks. developed in this paper. A Variable Harmonic

Analysis (VHA) has been developed to decompose

the time series into sine and cosine waveforms for

any desired wave resolution within the data length

Stochastic Climate Models (or fundamental period). In the Classical Harmonic

Analysis (CHA) the wavelength is strictly an

integer multiple of the fundamental period. For

EMAD: An Empirical Model of the atmospheric

smoothing the singular spectrum analysis (SSA)

Dynamics, Juan Pedro Montavez and Jin Song

has been applied. The SSA provides the mechanism

von Storch (Meteorological Institute of Hamburg

to decompose the series into certain number of

University)

principal components (PCs) then recombine the

first few PCs, representing the dominant modes of

In this work we present a hybrid model consisting

variation, to get the smoothed version of the actual

of an ocean GCM (HOPE-G) and an empirical mo-

series.

del of the atmospheric dynamics (EMAD). The

Twenty-four time series of terrestrial and extra-

goal is to extract, in a most complete way, the air-

terrestrial parameters, which visibly show strong

sea interaction processes represented by a fully

regularity, are considered in the study. They can be

coupled AO-GCM and express the statistical

broadly grouped into five categories: (i) inter-

properties of these processes in terms of a simple

annual series of number of storms/depressions over

model (EMAD).

the Indian region, seasonal and annual mean

EMAD is formulated for the global ocean, i.e. in-

northern hemisphere land-area surface air tempera-

cluding the regions covered by sea ice. The model

ture and the annual mean sunspot number (chosen

is able to generate the heat, fresh water and

cases of long term/short trends or oscillation); (ii)

momentum fluxes to drive the ocean model, and to

monthly sequence of zonal wind at 50-mb, 30-mb

response to the surface conditions given by the

and 10-mb levels over Balboa (representative of

ocean. The formulation of the model consist of two

quasi-biennial oscillation); (iii) monthly sequence

parts, a deterministic part formulated in a truncated

of surface air temperature (SAT) over the Indian

EOF space and a stochastic part formulated in the

region (strongly dominated by seasonality); (iv)

physical space. The model is fitted to 200 years of

monthly sequence of sea surface temperature (SST)

daily data from an integration with the coupled

of tropical Indian and Pacific Oceans (aperiodic

ECHO-G GCM. The skill of the model is better

oscillation related to El Ni0/La Nia); (v) se-

than a persistence model.

quence of monthly sea level pressure (SLP) of se-

The model is able to produce realistic responses of

lected places over ENSO region (seasonality and

the atmosphere to oceanic anomalies related to

aperiodic oscillation). Best predictions are obtained

ENSO and polynyas. Finally a coupled HOPE/-

for the SLP followed by SAT and SST due to

EMAD run has been performed. Using cross-corre-

strong seasonality and/or aperiodic oscillations. The

lation functions between SST and surface fluxes, it

predictions are found satisfactory for the lower stra-

is shown that the hybrid model is able to reproduce

tospheric zonal wind over Balboa, which displays

the most features of the air-sea interaction pro-

quasi-periodic oscillations. Because of a steep

cesses obtained in the fully coupled ECHO-G

declining trend a reliable prediction of number of

GCM.

storms/depressions over India is possible by the use

of the method. Prediction of northern hemisphere

surface air temperature anomaly is not found satis-

On the Natural Regularity and Predictability of

factory.

Terrestrial and Extra-Terrestrial Parameters,

Nityanand Singh (Indian Institute of Tropical Me-

teorology, Pashan, Pune, India)

Stochastic Tropical Atmospheric Boundary-Layer

Extrapolation of dominant modes of fluctuations Dynamics: Convective Downdrafts and Recovery,

after fitting suitable mathematical function to the Discharge-Recharge Processes, and 1/f-Noise,

observed long period time series is one of the Jun-Ichi Yano and Richard Blender (Meteoro-

approaches to long-term weather or short-term cli- lo-gisches Institut, Universitaet Hamburg, Ham-

mate prediction. Experiences suggest that reliable burg, Germany)

predictions can be made from such approaches

It was recently shown by the authors (manuscript

provided the time series being modelled possesses

submitted to J. C.) by analyzing TOGA-COARE

considerable regularity. Choice of the suitable

sonde data that the tropical convective variability

function is also an important task of the time series

23

measured as CAPE (convective available potential which is minimized in dependence of the climatic

energy) varies as 1/f-noise for the time scale of 1- state by a variational approach. The implementation

100 days. Variability of the surface moisture and, to of the conditional probabilities leads to a non-qua-

a lesser extent, temperature are responsible for this dratic cost function interpretable as a maximum-li-

1/f-noise behavior of CAPE. It is speculated that kelihood function. A simplified dynamical model

stochasticity of convective downdrafts, which sud- and a first-guess field of the palaeoclimate are in-

denly dry and cool the surface boundary-layer, and plemented as weak constraints. The reconstructed

the following sudden recovery is the main physical climate is found at the minimum of the cost func-

mechanism. This mechanism may be further idea- tion.

lized as a discharge-recharge process, which can be, The method is tested on European palaeobotanical

furthermore, regarded as a type of self-criticality. In data to reconstruct temperatures of two time slices

order to generate a 1/f-noise from a random series of the Eemian interglacial ( 125-115 kyears BP)

of such pulse-like events, these time intervals must which are characterized by relatively stable climatic

be close to Gaussian distribution (rather than Pois- and vegetational conditions similar to the present

son) with the mean interval corresponding to the time.

minimum period for the 1/f-noise. The present

paper reports on the statistical analysis of the

TOGA-COARE research vessel surface obser- How realistic is GCM-simulated precipitation?

vations, which are frequent enough (1-15 min) to Nikolaus Groll, Martin Widmann (Institute of Hy-

resolve the individual downdraft events. drophysics, GKSS Research Centre, Geesthacht,

Germany)

Statistics, Models and Data estimating regional precipitation variability from

Assimilation changes in the synoptic-scale flow.

The first step is to assess how well the NCEP re-

analysis data captures precipitation variability. The

Variational Assimilation of botanical proxy Data rationale behind this is that the NCEP reanalysis

for the Reconstruction of Palaeoclimate, can be thought of as a proxy for an ideal general

Christoph Gebhardt (Meteorologisches Institut circulation model (GCM) which nearly perfectly

der Universitt Bonn, Germany); Norbert Khl represents the synoptic-scale flow, but in which

(Institut fr Palontologie, Universitt Bonn, areas of complex topography are not sufficiently

Germany); Andreas Hense (Meteorologisches resolved. Reanalysis precipitation is entirely cal-

Institut der Universitt Bonn, Germany); Thomas culated according to the model physics and para-

Litt (Institut fr Palontologie, Universitt Bonn, meterisations, and is not adjusted to observations.

Germany) Precipitation fields from the NCEP reanalysis have

therefore been evaluated using high-resolution grid-

The terrestrial palaeorecords of botanical fossils box datasets for precipitation in Oregon and Wa-

(e.g. pollen, macro fossils) serve as proxy data shington, as well as over the Alps. The reanalysis

which provide important information that can be captures precipitation varability on time scales of

used for the reconstruction of near-surface palae- months to years accurately. Nevertheless there are

oclimate. However, a spatially coherent recon- over- and underestimations of the seasonal means,

struction is often hampered by a relatively sparse which affect not only spatial scales up to a few

number of fossil sites and by factors influencing model gridcells over Oregon and Washington, but

each fossil record on a regional scale. In this case, a also the total area mean over the Alpine region.

reconstruction of climatological fields requires ad- These systematic biases can be mainly attributed to

ditional information like a background field or dy- the poor representation of the topography in the

namical constraints. We present a simple two-di- NCEP model.

mensional variational assimilation of the proxy data To determine how well precipitation variability is

resulting in a reconstructed climate which is con- represented in GCMs that are used for climate

sistent with the observed palaeovegetation as well change studies, the ECHAM4 model has been

as with large-scale features of the atmospheric dy- investigated. In contrast to the NCEP Reanalysis

namics. the sequence of atmospheric states in the ECHAM4

In a first step, transfer functions quantifying the re- control run is not directly related to the observed

lation between climate and the botanical proxy data states. In order to allow for a comparison of

are found by combining a gridded climatology of simulated and observed precipitation, we developed

30-year (1961--1990) monthly means with distri- an analog-based resampling method (ABR). This

bution maps of the recent vegetation. This results in method selects for each observed day the most

Bayesian transfer functions describing conditional similar day from the control run (with respect to the

probabilities for the existence of the important spe- SLP field over a specified region). The observed

cies given certain climatic states. Thus a most pro- precipitation can then be compared with the simu-

bable geographical distribution of each species can lated precipitation of the analog day. ABR is cur-

be related to a given climate. The deviation of this rently applied to SLP fields over the North Atlantic

conditional vegetation from the actually observed Ocean and Europe in order to compare ECHAM4

palaeovegetation is described in a cost function

24

simulated precipitation with observed precipitation Spatial and temporal properties of ozone immission

over the Alps. structures are studied under stochastic-deterministic

Our results show that simulated precipitation might model assumptions. The given data are the half-

be a good predictor for regional precipitation down- hour averages of pollutants and meteorological pa-

scaling. rameters at 30 monitoring stations from 1995 to

1999. The sites are irregularly distributed over Sa-

xony, concentrated in the strongly loaded areas

Homogenisation and Interpolation of incomplete (with respect to primary pollutants), in the cities

climatological Data with simple physically based and along the moutain ranges along the Czech-

Models, Andreas Hense, Christoph Gebhardt German boundary (Ore Mountains). The target of

(Meteorologisches Institut, Universitt the work is the production of maps of the ozone

Bonn,Bonn, Germany) load for half-hour intervals.

In the given case spatial interpolation is rather diffi-

A characteristic of climatological data sets like the cult because of the heterogenity of the monitoring

Comprehensive Ocean Atmosphere Data set network. This concerns the spatial distribution of

COADS is the incompleteness in space and time. the stations as well as their representativeness,

Furthermore, even data sets like the Climate which differs for city stations, outskirts stations and

Research Unit (CRU) temperature data set may land stations. Originally the monitoring network

suffer from insufficient sampling on the sub-month- was constructed to observe the maximum concen-

ly time scale leading to unrepresentative data with trations of the primary pollutants like nitrogen

respect to neighbouring grid points. oxides or sulphur dioxide; therefor stations are mis-

We will present a simple but physically based inter- sing at positions where extreme point concentra-

polation method ("data model") to interpolated and tions of ozone are expected (in the lee of the big

homogenize existing climatological data sets up to cities). Big problems also result from the fact that

an a-priori selected spatial scale. For the COADS some stations close together (e.g. three stations in

1950-1994 in the Atlantic we analyse wind (vector the city of Leipzig) are located in regions of dif-

and velocity) and sea level pressure using an opti- ferent air-chemical regimes, what leads to a special

mal averaging procedure and an Ekman wind mo- kind of trend in the data.

del as a dynamic constraint. The data model is Despite these difficult circumstances, the Saxonian

formulated through a variational approach. The first ozone data could be nevertheless spatio-temporally

result is that the artificial trends in the wind data set trend-corrected by methods such as meanpolishing

can be reduced taking into account the pressure following Cressie so that finally geostatistical pro-

information but unfortunately as a second result it cedures for stationary processes could be used. In

turns out that the simple Ekman model is insuf- particular, a reasonable variogram could be esti-

ficient even on the monthly time scale due to the mated showing sufficient correlations between the

omission of curvature effects. The interpolation and ozone values at neighbouring stations. The long

homogenization of the CRU temperature data is range of the variogram of more than one hundred

again done in a variational framework. A two di- kilometers is in accordance with what is known

mensional advection diffusion model defines modes about the spatial process of ozone distribution.

which are used as a projection basis. The ampli- These spatial investigations were combined with

tudes are estimated from the available data under temporal analyses. In particular, the temporal inves-

the constraint that the resulting tendencies of the tigations were carried out for the whole data set as

advection-diffusion model are minimal in the least well as for single stations respectively seasons,

square sense. This data model is such efficient that leading to interesting relationships which are useful

Monte Carlo methods like cross validation can for the prediction of the ozone distribution.

provide a quality control by detecting unrepre-

sentative data and estimating an a-posteriori error

measure of the derived complete fields. Reduced-order schemes for data assimilation in

Results will be presented for the period 1881-1998 oceanographiy based on the Kalman filter, I.

and the area 2.5S to 77.5N. Hoteit and D.T. Pham (Laboratoire de Modeli-

sation et Calcul, Grenoble, France)

Trends of Maximum and Minimum Temperatures Brute-force implementation of the extended Kal-

in Northern South America, Ramon A. Quintana- man (EK) filter in realistic ocean models is not

Gomez (Department of Earth Sciences, possible because of its prohibitive cost. To remedy

Universidad de los Llanos Ezequiel Zamora, to this problem, different degraded forms of the EK

Barinas, Estado Barinas, Venezuela) filter, which basically reduce the dimension of the

system through some kind of projection onto a low

dimensional subspace, have been proposed. The

Investigation of Spatial and Temporal Ozone main principle of our contribution is to study the

Immission Structures under Stochastic-Deter- usefulness of the evolution in time of the reduced-

ministic Model Assumptions, Heike Hoffmann order state space. We will therefore compare the

(Graduate School of Spatial Statistics, Freiberg, theoretical and practical aspects of the singular

University of Mining and Technology, Germany) evolutive extended Kalman (SEEK) filter intro-

duced by Pham et al. and the reduced-order

25

extended Kalman (ROEK) filter introduced by nterpart. The method is non-parametric and consists

Cane et al.. Indeed, the reduced state space of the in fitting equally-spaced percentiles in the observed

SEEK filter is let to evolve in time according to the and simulating distributions and linearly inter-

model dynamic while the evolution of the reduced polating between them. The only assumption it

space of the ROEK filter is replaced, and thus re- makes is that percentiles of a distribution (not

mains fixed in time, by the evolution of the error in values themselves) are simulated correctly.

the reduced dynamic defined by the reduced space. The method is applied to daily temperature and

On the other hand, the SEEK and the ROEK filters precipitation at several sites in central Europe, si-

remain expensive in real operational oceanography. mulated by two GCMs and a downscaling proce-

Hoteit et al. have proposed different degraded dure. The statistical moments (up to the fourth) of

forms of the SEEK filter to reduce its cost. Their the simulated temperature distributions after the

approach consists in simplifying the evolution of adjustment are very close to those observed, even if

the reduced space, which is the most expensive part the simulated distributions themselves were seve-

of the SEEK filter. To reduce the cost of the ROEK rely distorted relative to observations. On the other

filter, we introduce the use of an autoregressive hand, the temporal and spatial structure of the

statistical model to let the error evolve in the simulated temperatures (e.g., lag-1 autocorrelations,

reduced space, which is the most expensive part of day-to-day variability, spatial autocorrelations, fre-

this filter. Finally, using a twin experiments ap- quency and duration of extreme periods) are not

proach, these filters have been implemented and affected by the adjustment. For precipitation, the

tested by assimilating altimetric data in a realistic numbers of dry days are corrected after the ad-

setting of the OPA model in the tropical Pacific justment and the probabilities of a wet-to-wet and

ocean. dry-to-wet transitions are considerably improved.

A modification of the method is applied to 2xCO2

GCM simulations. It assumes that GCMs correctly

Cloudiness and sea surface temperature: tackling simulate the temperature difference between the

the problem of missing values, Richenda future and present climate for the equally spaced

Houseago-Stokes and Peter Challenor percentiles. We show that the method successfully

(Southampton Oceanography, Centre removes GCMs errors from the temperature series.

Southampton, UK) In validation studies of GCM-simulated or down-

scaled climate variables for the purposes of impact

Sea surface temperature from infra red radiometers assessments, the main emphasis is put on the first

on satellites is an important climate variable. Un- two statistical moments of their distributions. We

fortunately, because of cloud, the data set contains argue that a detailed validation of them is of se-

many missing values. This makes it difficult to ap- condary importance as one can easily adjust their

ply Principal Component Analysis to look at the distributions to fit the observations. Instead, the va-

dominant modes of variability. We investigate two lidation studies should put much more emphasis on

solutions to this problem. The first is to interpolate temporal and spatial structure of the variables, for

the missing data using spatial optimal interpolation which such an adjusting method does not exist.

on the non-missing data and apply standard me-

thods to evaluate the empirical orthogonal functions

(EOF's). The other method is to use the recently de- Simplified EOFs - three alternatives to rotation,

veloped method of Probabilistic Principal Compo- Ian Iolliffe (University of Aberdeen, UK);

nent Analysis and apply the EM algorithm. This Mudassir Uddin (University of Karachi,

algorithm enables us to estimate the EOF's and the Pakistan); Karen Vines (Open University, UK)

missing values simultaneously. To test the two me-

thods we compare them on data where we have si- Principal component analysis (PCA) is widely used

mulated missing values by deleting points at in atmospheric science, and the resulting empirical

random and thus know the true results. We then ap- orthogonal functions (EOFs) are often rotated to aid

ply them to fifteen years of sea surface temperature interpretation. We describe three methods that

data in the North Atlantic from the Advanced Very provide alternatives to the standard two-stage pro-

High Resolution Radiometer (AVHRR) on US sa- cedure of PCA followed by rotation. The techni-

tellites. ques are illustrated on a simple example involving

Mediterranean sea surface temperature.

The first method compresses the usual two stages

Removing errors in simulated distributions of cli- into one, by successively maximizing a criterion

mate elements using a non-parametric adjustment that combines variance and simplicity. The second

procedure, Radan Huth1, Jan Kysel1,2 (1 Institute technique is based on an idea from regression, the

of Atmospheric Physics, Prague, Czech Republic; LASSO (Least Absolute Shrinkage and Selection

2

Dept. of Meteorology and Environment Operator). Our technique adds a constraint on the

Protection, Charles University, Prague, Czech sum of absolute values of elements in an EOF to

Republic) the usual PCA optimisation problem, which has the

effect of driving some elements to zero. Our third

We propose a method that allows one to adjust a approach seeks high variance EOFs such that the

simulated (GCM-simulated or downscaled) distri- elements of each EOF are proportional to (hope-

bution of any climate element to its observed cou- fully low-valued) integers. The approach achieves

26

this by applying a series of variance maximising mation provided by the data onto the model state

but "simplicity-preserving" transformations to an space as another degree of believe. Then, having

initial set of particularly simple, but not necessarily two probability distributions over the same space,

high variance, initial EOFs. we may apply the probability theory to the pro-

Each technique is shown to provide different sim- babilities themselves. This leads to a natural cri-

plified interpretations for the major sources of terion of compatibility of the data with the model

variation in a data set. Comparisons will be made solution for given data and model uncertainties.

between the three methods and with standard PCA Contrary to the chi2 criterion, it is applicable to any

and its rotated version. We conclude that all three probability distributions.

techniques have advantages compared to the usual As the probabilities on the model state space are

two-stage rotation. viewed as degrees of belief, the principle of maxi-

mum entropy should be employed to estimate

statistical parameters of the model errors since it

Reconstruction of large-scale climate, J.M. Jones yields the least unbiased solution for assigning pro-

(GKSS Research Centre, Geesthacht, Germany) babilities. It turns out that if the compatibility cri-

terion is fulfilled, the prediction made with the

It is aimed to reconstruct those large-scale patterns VAR may be expressed as the mean with respect to

of extratropical circulation which are strongly lin- an operator valued probability measure of what is

ked to variability in proxy data. Current work is told by the data and by the model. Maximization of

centered on the Antarctic circumpolar circulation. a generalization of Shannon's entropy for such

Using canonical correlation analysis we found pro- probabilities is proposed to use for estimating the

mising relationships between tree ring chronologies model error statistics. The performance of this

from Argentina, Chile and New Zealand, and large- approach is compared with other techniques widely

scale climate patterns representing the Antarctic used in the VAR such as the maximum likelihood

circumpolar circulation for sea level pressure and and the generalized cross-validation.

850hPa geopotential height, taken from the NCEP

reanalysis. This work shall also be extended to

Northern Hemisphere circulation modes. The suita- Statistical presentation of wind wave climate,

bility of other proxy records, such as ice core and Leonid Lopatoukhin1,2 , Valentine Rozhkov1,

1,3

coral data, for incorporation into these reconstruc- Alexander Boukhanovsky , (1 State

2

tions is also being investigated. Preliminary ana- Oceanographic Institute; St. Petersburg State

lysis of Greenland ice core accumulation shows University (Dep. Oceanology); 3 Institute of High

links with large-scale Northern Hemisphere atmos- Performance Computing and Data Base, Russia)

pheric circulation. These reconstructions are the

first stage of DATUN (Data Assimilation through 1. Existence of 40 years (1957-1996) of reanalysis

Upscaling and Nudging). The aim of the DATUN of meteorological data and verified spectral wave

method is to obtain a physically based, best guess models allows to calculate the ensemble of spatial-

for the large-scale states of the atmosphere during temporal wind and wave fields in a grid points.

the Late Holocene with annual temporal resolution. 2. Vector wind field V(r,t) is specified by mean

The second stage of this methodology is nudging mV(r,t), variance DV(r,t), covariance KV(r,?,t,?),

the large-scale variability of a GCM towards the where r is vector of spatial coordinates, ? vector

results from the upscaling. of displacement between the grids, t- time, ? -lag.

Invariant of tensors DV(?), KV(?) are used for

cartographic presentation. The annual variations of

On a probabilistic interpretation of variational wind speed are represented as the expansions of

data assimilation and model error estimation, mV (t), DV(t), KV(t,?) by the trigonometric basis in

Gennady A. Kivman (Alfred-Wegener-Institut for accordance with the theory of periodically corre-

Polar- and Marine Research, Bremerhaven, Ger- lated random processes.

many) 3. A set of 9 values (h,?,?)i , is used for parame-

terization of complicated fields of wind wave and

Estimating the model error statistics is a funda- swell spectra S(?,?,r,t). Here i=1, 2, 3 indexes for

mental problem of variational data assimilation wave height h, periods ? and directions ? for wind

(VAR). It cannot be solved without understanding waves (i=1) and swells (i=2, 3), calculated by mean

of the VAR in probabilistic terms. A serious de- of the moments mjk(j,k = 0, 1, 2) of truncated

terrent to the use of the probability theory is that spectra S(?,?) and peak frequency. Such a change

probabilities for the data and for the model va- of S(?,?) to a set of 9 random values ? simplify the

riables are conceptually diverse quantities. In envi- treatment the results of calculations. Besides,

ronmental studies, the former may be treated as applying of both the theory of non-random function

relative frequencies while the latter generally ex- S(?,?) = S(?,?,?) of random arguments ? and met-

press degrees of believe. Thus, it is difficult to give hod of statistical linearization propose the approach

an interpretation to probabilities on the joint data- for classification of climatic wave spectra.

model state space and to view the VAR as an 4. Annual maxima of wave heights and wind speed

application of Bayes' theorem. are recorded in different months and from different

Restoration of the probability interpretation of the directions. The ensemble of values (h, ?) ? (|V|,?) is

VAR may be achieved by transferring the infor-

27

a mixture of distribution and Monte-Carlo simula- A regional L-moments approach has been used to

tion of random values and function is used. study the yearly extreme k-days amounts in Bel-

5. Synoptic variability of waves is connected with a gium. Two different datasets originating from two

sequence of storms and weather windows and it pa- networks are available for this purpose: the dense

rametric representation (in accordance with a climatological network yielding yearly extremes of

theory of random outlier) may be given in a system k-days precipitation amounts (with k= 1, ..., 30

? of four random values {h+ ,h- ,?,?}. Where h+ ,h- days); and the sparser hydrometeorological network

the value of outlier (up and down), ?,? - duration equipped with recording raingauges with a time

of outlier. Method of estimations of wind and wave step of 10 minutes. The latter dataset provided the

extreme with different return periods is proposed yearly extremes for durations ranging from 10 mi-

and based on the results of analysis of distributions nutes through 7 days.

?. The regional homogeneity of the distribution being

verified, the growth curves of the regional GEV-

distributions have been established. This allows for

Model-based spatial modelling of climatological the comparison of the extreme value distributions

data, J. Mateu 1, P. Juan 1, C. Antolin 2,3, C. Ano 3 based upon yearly extremes from the two datasets

(1 Department of Mathematics, Universitat Jaume (including an "experimental" assessment of the

I, E-12071 Castellon, Spain; 2 Department of Herschfield factor H) and the sharp determination

Vegetal Biology, Universitat de Valencia, of the curve parameter of the GEV-distribution for

3

Burjassot, Spain; CIDE-Centre for a large range of durations.

Desertification, Cami de la Marjal s/n., Albal

(Valencia), Spain)

Extreme precipitation events in southeastern

The problem of estimation and prediction of a spa- South America and the relationships with sum-

tial stochastic process, observed at irregular loca- mertime large-scale South American Stationary

tions in space, is considered. The statistical model Systems (SASSY), Leila M. V. Carvalho1,2,

proposed is a mixture of two components, one in- Charles Jones1, Brant Liebmann3 (1Institute for

volving a non-parametric term, accounting for a de- Computational Earth System Sciences (ICESS),

terministic trend given by exogenous variables, and University of California, Santa Barbara,

a parametric component defining the purely spatial California, USA; 2Dept. of Atmospheric Sciences,

random variation. Several well known statistical University of Sao Paulo, Sao Paulo, Brazil;

3

models such as the full interactive model, ridge-re- NOAA-CIRES Climate Diagnostics Center,

gression or neural-network regression are particular Boulder, Colorado,USA)

cases of the proposed modelization. Field variables

usually show spatial dependence among observa- The South Atlantic Convergence Zone (SACZ) is

tions which is an important drawback to traditional characterized by a wide range of temporal varia-

statistical methods. However, geostatistical tech- bility, from synoptic to intraseasonal and inter-

niques analyze and describe the spatial dependence annual scales. Our recent study has objectively

and quantify the scale and intensity of the spatial identified the occurrence of South American Sta-

variation providing the essential spatial information tionary Systems (SASSY) during austral summer.

for local estimation. Geostatistics is principally the SASSY events are characterized by persistence of

application of regionalized variable theory. two or more days and large geographic extension

The estimation process is developed in a two-step over southeast Brazil. This study investigates the

procedure, first addressing the estimation of the occurrence of extreme precipitation events in south-

non-parametric component, and then the estimation east South America and the relationships with the

of the parametric term from the residual stochastic SASSY. Daily Outgoing Longwave Radiation

process. The proposed methodology is applied to (OLR) and precipitation data derived from surface

model the spatial process of rain intensity (rainfall stations from the states of the South and Southeast

erosivity) in a mediterranean region. Some other re- Brazil (1979 to 1999) are used to determine the

lated processes like elevation or distance to the sea frequency of occurrence of extreme precipitation

are also considered. These variables are considered events. Whenever the daily precipitation for a cer-

to be highly important in the mechanisms and pro- tain period exceeds a given percent of the

cesses of erosion by water in Mediterranean en- climatological mean for that period, it is considered

vironments. as an event. This presentation will focus on the

spatial and temporal variability of the SASSY in

association with extreme precipitation events.

Diagnosis of variations in the climate-related

flood risk - Some results from the Interior Wes-

Regional GEV-growth curves of yearly preci-

tern United States, Shaleen Jain, Martin

pitatation depths in Belgium for durations ran-

Hoerling, Gary Bates, Randall Dole

ging from 10 minutes to 30 days, Daniel Gellens,

(NOAA/CIRES Climate Diagnostics Center,

Gaston Demare (Royal Meteorologuical Institute

Boulder, CO)

of Belgium, Brussels, Belgium)

28

In the interior western United States, spring floods pitation formation, occurrence of lightning and

(i.e., peak streamflow) reflect an interesting integra- point discharge current are taking place. The non-

tion of the seasonal-to-interannual climate varia- linear conduction current flowing from the atmos-

tions (through winter snowpack development) and phere to the surface of the earth when the electric

Springtime weather patterns (responsible for the field is sufficiently strong is called Point Discharge

snowmelt). Consequently, trends and year-to-year Current (PDC). In the present paper minute interval

variations in the flood magnitude and timing may data of Point Discharge Current (PDC) collected for

be cast in a climate-related risk assessment frame- a total of 65 thunderstorm days occurred during 6

work. We present a detailed empirical-statistical years period (1972-1977) at Pune (18 32'N, 73

assessment of this climate-flood problem, with 51'E, 559 m asl) has been subjected to statistical

special emphasis on the sensitivity of flood peaks analysis namely frequency distribution at different

to: (a) large-scale climate signal during winter and time scale (distribution during the total diurnal

Spring, and their potential predictability, and (b) period). The total diurnal period (0000 - 2359 IST)

regional hydrologic features (e.g., elevation and is divided into 3 distinct time intervals i.e. after-

drainage area of river basins). noon (1400-2100 IST); night time (2100-0759 IST)

The observed flood probability distribution is exa- and day time (0800-1400 IST). These three time

mined for preferred tail behavior, consequent to the intervals are practically consistent with the require-

known modes of seasonal-to-interannual climate ment of separating the diurnal period into convec-

variability. Some potential application areas for tive and non-convective atmospheric regimes.

such studies (including water resources manage- Gaussian Model (computer base), is applied and the

ment, stream ecology, hydraulic design, and emer- results suggested that the distribution is normal and

gency response) are discussed. the chi-square value obtained (5.2008) is less than

the standard table value (9.542) for 22 degrees of

freedom. From this result it is seen that the hypo-

The Climatological Probability of Drought in thesis applied to the present data set holds good.

Northern Province of South Africa, Tiba Kabanda Since it is a normal distribution the data has been

(School of Environmental Sciences, University of subjected to Fischer's test for the skewness and

Venda, Thohoyadou, South Africa) kurtosis. The results thus obtained revealed that the

distribution is found to be normal with negative

This paper presents a study of the climatology of skewness and exhibits the property of leptokurtic

drought in the Northern Region of South Africa. distribution. The sustained occurrence of negative

Drought is described as a rare and severe event, and positive PDC at the end of the storm day in

therefore the frequency of occurrence and intensity second and third category of time interval may be

need to be explained by using a criterion that will due to the presence of space charges of either po-

distinguish normal dry spells from real drought. larity.

The paper is intending to describe such methods.

The study examines the geographical variability of

dry spells in the area of study and analyses drought A Dedacal Study of Premonsoon Seasons Thun-

(seasonal) in detail to reveal its frequency, cover- derstorms Over Pune Region by Using Markov

age, and persistence. Rainfall stations in the region Chain Model, S.S. Kandalgaonkar, M.K.

are the point-source data samples that represent the Kulkarni, M.I.R. Tinmaker (Indian Institute of

study region. All stations were intercorrelated to Tropical Meteorology, Pune, India)

produce blocks of highly correlating sites. The site

characteristics depict different physical proportions Probabilistic description of the thunderstorm phe-

such as geographical location and elevation. Rain- nomena in the premomonsoon season over Pune

fall data for 40 years is used in this study. (18 32'N, 73 51'E, 559 m asl) region has been

The study employs standardised precipitation Index studied with the help of Markov Chain Models. The

(SPI) in the analysis because it allows the deter- daily thunderstorm data of pre-monsoon season for

mination of the rarity of an episode in terms of pro- a period of 11 years i.e. from (1970-1980) have

bability. The results obtained from the study reveal been subjected to the Markov Chain models of first,

spatial-temporal distribution of drought within the second and third order. Use of Akaike's Information

study area. Criterion (AIC) clearly indicated that the first order

Markov Chain Model is the best which described

the atmospheric process of thunderstorm pheno-

Statistical Analysis of Point Discharge Current at mena over the Pune region. This is based on the

Different Time Scale Over the Pune Region, S.S. decision procedure depending upon the principle of

Kandalgaonkar, M.I.R. Tinmaker, M.K. Kulkarni maximum likelihood. The steady state probabilities

(Indian Institute of Tropical Meteorology, Pune, and mean recurrence time of dry and wet days

India) have also been calculated for first and second order

models for each separately and also for the total of

Thunderstorm is an important weather phenomenon 11-years. These computations revealed that the ob-

for the understanding of the many issues related to served and theoretical values of steady state

the atmospheric electricity. As the storm forms and probabilities realistically consistent.

dissipates several complex cloud physical and elec-

trical processes such as charge separation, preci-

29

Trends of Atlantic wave extremes as simulated in phenomenon. At a particular location the record of

a 40-year wave hindcast using kinematically past occurrence is too sparse to provide meaningful

reanalyzed wind fields, Xiaolan Wang* and Val R. statistics to use in planning for future landfalls.

Swail (Network Strategies Division, AMWSD In order to exploit the historical database more

Meteorological Service of Canada, Downsview extensively, Monte Carlo simulations have been

Ontario, Canada) carried out. Using tropical cyclone characteristics

compiled by Ho et al. (1987), probability density

In this study, both seasonal and monthly extremes functions were assembled for storm speed and di-

of wave height in the North Atlantic are analyzed. rection, central pressure and radius of maximum

The analysis is based on a 40-year (1948-1997) nu- wind. After random selection of parameters from

merical hindcast using an intensive kinematic the probability density functions, a tropical cyclone

reanalysis of wind fields. Changes in the ocean wind field model (Holland et al. 1991) was ini-

wave extremes are identified by performing the tialized to obtain the horizontal winds.

Mann-Kendall test, and are further related to Simulated propagation of the model cyclone then

changes in the atmospheric circulation (sea level allowed the time history of wind speed and direc-

pressure) by means of Redundancy Analysis. The tion at landfall to be tabulated at specific locations.

month-to-month variability of trends is also re- Repeated random sampling allowed a 1000-storm

vealed. climatology to be generated.

The results show that the trend patterns of wave ex- The simulation results are portrayed for 20 sites

tremes vary greatly from month to month. Gene- along the United States Atlantic and Gulf coasts.

rally, changes are most extensive in January and The contrasting simulated climatologies are dis-

March for the northern North Atlantic, and in July cussed in terms of the meteorological and physical

and October for the central and/or subtropical North settings as well as the impacts for planning and

Atlantic. In January and March, significantincreases design.

of wave extremes in the northeast North Atlantic

are closely related to changes in the North Atlantic

Oscillation (i.e., the deepening of the Icelandic low Simulation of ocean wave climate changes

and the strengthening of the Azores high) during through wind variance changes on different

the last four decades. In June, July and November, timescales, E. Bauer1,R. Weisse2 (1PIK, Potsdam;

2

the trend pattern features significant increases of GKSS, Geesthacht, Germany)

wave extremes in the ocean off the North America

coast; and it features a large area of significant in- The observed mean ocean wave height in the North

creases in the subtropical North Atlantic in Atlantic has grown significantly over the past 30-40

October. Changes in the other months are of little years. Since no corresponding growth of the mean

field significance. Generally, increases of ocean wind forcing was observed, the effect of wind

wave height are associated with an anomaly low variance changes on two different timescales was

pressure over the region. tested. Test cases were conducted with the wave

The results also indicate that the ocean off the model WAM using ECMWF-Reanalysis wind

North America coast has roughened in August, No- fields. The varying balance of the wave generation

vember and December during the recent decades, and wave dissipation processes suggested to esti-

experiencing higher wave extremes. mate the effect of storm frequency changes on daily

Since the kinematic reanalysis of winds correctly timescales and of high-frequency wind variance

re-assimilates in situ surface wind data, intensifys changes on hourly timescales. First, the storm fre-

wind fields in extratropical storms as necessary, quency changes were simulated by modifying the

and includes tropical cyclone boundary layer winds, time intervals between succeeding storm events.

the resulting wave hindcast shows greater rates of Second, the high-frequency wind variance was

changes than those identified from a wave hindcast simulated by a quasi-realistic wind generator deve-

using the NCEP reanalyzed winds for the same pe- loped from measurements with 20-minutes time

riod. resolution. The two simulation studies revealed a

Especially, more significant increases are detected mean wave height growth with decreasing storm

for the ocean off the Canadian coast in summer, frequency on the one hand, and with increasing

and for the central North Atlantic in fall. high-frequency wind forcing on the other hand. To

explain the observed wave height growth in the

North Atlantic support is provided by independent

Hurricane Landfall Winds Along the United observations of growing duration of westerly

States Gulf and Atlantic Coasts: Monte Carlo weather patterns in connection with the intensifying

Simulations, Gary D. Skwira*, Douglas A. Smith+ North Atlantic Oscillation (NAO) during that

and Richard E. Peterson* (Wind Science and period. So, the first simulation helps to identify

Engineering Program, Texas Tech University, mechanisms inducing wave climate changes where-

Lubbock, TX, USA; *Department of Geosciences, as the second simulation yields an estimate of

+Department of Civil Engineering) modelling uncertainty.

Gulf and Atlantic Coasts have had a greater impact Frequency and Intensity of Mediterranean Cyclo-

in recent years, they remain a relatively infrequent nes in a doubled CO2 scenario, P.Lionello

30

(University of Lecce), F.Dalan (University of This analysis is based on 30 year long simulations

Padua) of the wave field and of the storm surge in the

Adriatic Sea using the information provided by the

The analysis is based on two 30-year long global global (T106) time slice simulations of the CTR

simulations carried out with the ECHAM 4 model and CO2 scenario carried out by DMI.The wave

at the DMI (Danish Meteorological Institute): a field simulations have been carried out using the

CTR (control) experiment, which reproduces the WAM model at 1/6 degs. resolution. The surge

present climate, and a CO2 experiment which simulations have been carried out with a barotropic

simulates the effect of a doubled atmospheric CO2 shallow water model at 1/12 degs resolution.

content. Both experiment were carried out at T106 A crucial element for the accurate reproduction of

resolution. surge events and of the wave field is the quality of

An assessment of the variations in the cyclonic ac- the surface wind. Unfortunately, the T106 resolu-

tivity is, obviously, important. The atmospheric cy- tion is absolutely inadequate for the reproduction of

clones are a potential source of major damages in the surface wind field in the Adriatic Sea. There-

the Mediterranean area, because of their association fore, it is not possible to use directly the T106 wind

with storm surges, high wind waves, intense preci- fields for the evaluation of surges and waves

pitation and floods. scenarios.

The method is based on the search of the pressure The wind fields used for the regional simulations

minima in the T106 slp (sea level pressure) fields, have been obtained by statistical downscaling of the

available at time intervals of 6 hours. The search is T106 slp fields. The downscaling was based on the

based on the partitioning of the slp fields in depres- CCA analysis of the ERA-15 T106 slp fields and of

sions by the identification of sets of steepest the regional surface wind fields produced by the

descent paths leading to the same slp minimum. BOLAM (BOlogna Limited Area Model) at a re-

To each depression has been associated a trajectory, solution of 25km.

obtained by following the location of its slp mini- The 30 year long simulation carried out with WAM

mum in time. The procedure results in a list of and with the surge model have been used for an

cyclones, their trajectories, initial and final points, analysis of the distribution of the extreme values.

and the sequences of their pressure minima, of the The results show a significant diminished intensity

areas covered, and of the maximum geostrophic of the extreme events. It appears that the dimi-

wind at the sea surface. nished cyclonic activity results also in a diminished

The analysis shows that: 1) the total number of cy- danger of coastal floods (ignoring the effect of the

clones in the CTR experiment is in acceptable sea level rise) and a diminished intensity of the

agreement with the ERA-15 data and it can be wave regime. The evaluation of the surge scenarios

considered an acceptable reproduction of the pre- should be considered cautiously, because of the

sent climate. 2) There is a statistically significant systematic underevaluation (not yet quantitatively

decrease in the total number of cyclones in the CO2 assessed) of surge events at the Venetian littoral

experiment. This suggests that the CO2 doubling present in the CTR scenario.

would result in a milder climate in the Mediter-

ranean Area.

The interannual variability of the cyclonic activity Extreme value analysis of absolute minimum tem-

remains comparable to the difference between the perature and minimum dew-point temperature at

two scenarios, so that a mild year of the present Sodankyl, Finland, Jaakko Helminen, Niklas

scenario and a stormy year of the doubled CO2 Sjblom (Finnish Meteorological Institute (FMI),

scenario have an equivalent number of cyclones. Helsinki, Finland)

No significant change in the geographic distribution

of the cyclones has been found. The simulation On January 29, 1999 new record low absolute mini-

with CO2 doubling presents a few intense events mum temperatures were reached at several sites in

that suggest a possible intensification of extreme Finnish Lapland with severe impacts on local ener-

cyclones. The number of events is, anyway, too gy supply. This led to the need to re-estimate the

small to have a reliable statistical significance. corresponding recurrence values. The results based

on the peak-over-threshold (POT) approach of the

absolute minimum temperature at Sodankyl for

Extreme waves and surge events in a doubled CO2 periods 1908-1998 and 1908-2000 will be

scenario in the Adriatic Sea. P.Lionello presented and compared. The other interesting

(University of Lecce) E.Elvini, A.Nizzero question in this context to be asked is how low the

(University of Padua) absolute minimum temperature can reach all in all.

As such limit the minimum dew-point temperature

The Venetian littoral, the flat northern coast of the will be considered as it represents the temperature

Adriatic Sea, is particularly vulnerable to extreme value at which the formation of shallow ice crystal

marine events. An intensification or a change of the fog will stop further cooling due to long-wave

regimes of storm surges and waves, due to the in- radiation at the ground. To clarify the question of

creased CO2 content in the atmosphere, would the lowest possible absolute minimum temperature

require the reorganization of the coastal defenses values the POT-analysis results of the minimum

for the protection of an economically important dew-point temperature will be presented and

area and a unique historical heritage. discussed as well.

31

Variability of the first principal components of

annu-al time series contains statistically significant

Observed Changes in Temperature Extremes of trends.

Korea, Won-Tae Kwon, Hee-Jeong Baek*, The first pair of canonical patterns for annual mean

Byung-Choel Choi (Meteorological Research explains 23 per cent of variance of regional sea le-

Institute, Korea Meteorological Administration; * vel pressure and 95 per cent of the sea level va-

Seoul National University, Korea) riance at Polish coast. The correlation between pat-

terns equals 0.74. The first canonical map of sea

Changes in mean, maximum and minimum tem- level pressure presents the Northern and Central

perature with daily resolution are calculated with 14 Europe under the influence of pressure system with

stations data (with relatively long records) in center over Finland. The rest of continent and At-

Korea. In the twentieth century, temperature in lantic Ocean are under the influence of a weaker

Korea is increasing and especially in Seoul, the and more extend pressure system with center over

minimum temperature has shown a large increase. Atlantic Ocean, west of Brittany. Such pattern is

Since 1980's the trends in minimum, maximum, and responsible for the changes of water amount in the

as well as mean temperatures become steeper than Baltic Sea. The corresponding canonical map of sea

the earlier decades. Ullung-do, an Island station, level is similar to its well-known first empirical

shows no trend until 1980's but indicates an orthogonal function.

increasing trend since then. The temperature Monthly and daily variability of sea level were re-

increases during the twentieth century are between constructed by means of canonical correlations

1C for rural and island stations, and 2C for patterns or redundancy patterns. They remain in

large metropolitan areas. Another prominent good and very good agreement with in situ obser-

change occurs in daily temperature ranges, showing vations (from the years 1947-95) in all seasons of

decreasing trend due to a large increase in the year expect of summer. Correlations coeffi-

minimum temperature. We also analyzed the cients between reconstructed time series and in situ

changes in extremely warm and cold days, and measurements reach to 0.88. Reconstructed time se-

freezing days (T < 0C) and hot summer days ries reveal positive trends in the nearly 100 years

(Tmin > 25C). The numbers of extremely warm 1899-1995. Reconstructed positive trends are

day and hot summer day are increased and this caused by an intensification of westerly flow over

trend is more prominent during 1990's. The the Baltic Sea Basin and due to increasing stability

numbers of extremely cold day and freezing day of westerly flow. An increase in the range of sea

have decreasing trends and the trends show sharp level variability is also confirmed by tendency in

drops in 1950's and 1980's. And total numbers of increase of higher level percentiles.

extremely warm and cold days are increasing with In case of increasing CO2 concentration a contin-

time, which suggests that the climate of Korea has uation of nowadays-existing mean sea level rise is

become more extreme in recent decades. We will expected. ECHAM1/LSG pre-industrial experiment

present the detailed results from station data and shows that there is expected slight increase of sea

also mid-tropospheric data. level rise in the respect to control run. In the end of

6th decade of 21st century a jump of sea level in

range of 4-6 cm may have place. ECHAM3 time-

slice experiment shows that the mentioned rise

Statistics, Models and Data might be additionally accelerated by 0.7-0.9 cm

Assimilation decade-1 in the respect to 1xCO2 concentration.

The same tendencies are expected for minimum and

maximum sea level values. The expected increase

The role of atmospheric large-scale circulation of percentiles should be lower the increase of mean

over Europe and North Atlantic in sea level varia- sea level, however range of their variability should

bility on the Polish coastal zone, Miroslaw Mietus be greater.

(Institute of Meteorology and Water Management,

Maritime Branch, Gdynia, Poland)

Space-time variograms and a functional form for

An extensive study concerning the relationship be- total air pollution measures, De Iaco, S.1, Myers,

tween large-scale pressure field over Europe and D. E.2, Posa, D.3,4 (1 Universita di Chieti, Pescara

North Atlantic and climatic and sea level variability Italy 2 University of Arizona, Tucson AZ - USA;

3

on the Polish coast has been carried out for the time Universit di Lecce, Lecce - Italy; 4 IRMA - CNR,

span of the last century and in case of expected Bari - Italy)

increase of the CO2 concentration.

EOF analysis of mean sea level on the Polish coast There are different ways to quantify Total Air

shows that there is generally one dominating mode Pollution (TAP) for a region. Spatial averages, over

of spatial variability explaining more than 90 per a region of interest, or temporal averages, for a

cent of variance in each season (only in summer 87 specified time period, could be considered for each

per cent). The leading mode presents increasing contaminant or for a weighted average of several

deviation from the long-term mean when moving contaminants. Such averages might be interpreted

from the western part of the coast towards east. The as the total mass of contamination for a time period

second mode explains 7-8 per cent of variance. and a geographic region. Knowing the possible

32

sources and their individual emissions would allow leading finite singular values of the linear

a comparison to determine if pollution is flowing propagator. The effect of observation density is also

into or out of a region. The comparison might also accessed, especially near the tracking of large flow

be used to determine if there are un-identified sour- instabilities. In a real setting, this study can help to

ces. put limits of data assimilability and predictability of

As an application of quantifying total air pollution, 4DVAR assimilated solutions.

data for NO, NO2 and CO collected in the Milan

District, Italy are used to generate a combined in-

dicator of pollution, TAP1. The weightings are ob- A multisite self-consistent Weather Generator for

tained by multiple PCA analyses of the daily con- Impact Studies in Portugal, Budong Qian, Joo

centration levels. The first components are treated Corte-Real, Hong Xu (ICAT, Faculty of Sciences,

as data in space-time and a space-time product-sum University of Lisbon, Portugal)

variogram is fitted to the data. Using this model and

the dual form of kriging, i.e., Radial Basis Func- Through previous studies done in past years,

tions, a functional form for the TAP1 concen- downscaling of precipitation and temperature for

trations inspace-time could be obtained. Using this Portugal, from large-scale output of GCMs inclu-

functional form, spatial and temporal integrals pro- ding both control and transient (scenario) runs, has

duce measures of total air pollution over a region in essentially been achieved (Corte-Real et al., 1999).

space for a specified time period. Using the speci- However, previous downscaling studies were de-

fied legal limits on the three contaminants, a limit is veloped site by site, not taking into account in a

generated for TAP1. In turn this can be used to consistent way, the interrelationships between pre-

identify the region in space-time where the maxi- cipitation and temperature, while some models

mum for TAP1 is exceeded. designed for impact studies, such as hydrological

models, require spatially distributed precipitation

scenarios consistent with temperature scenarios as

Performance of four-dimensional variational assi- well. Stochastic weather generators have been de-

milation and dynamical instability, experiments veloped either to model climate elements (such as

on a quasi-geostrophic model, Carlos Pires precipitation, temperature and radiation in many

(CGUL, - Centro de Geofisica da Universidade de studies) at a given site or to simulate multisite

Lisboa, Lisboa, Portugal) distributions of a specific element such as preci-

pitation (Wilks, 1999). In this contribution, a sto-

We perform four-dimensional variational assimi- chastic downscaling model, developed for conti-

lation (4DVAR) of artificially generated noisy ob- nental Portugal, for obtaining spatially distributed

servations on a multi-level quasi-geostrophic model regional precipitation scenarios consistent with

in the perfect model setting. The quality of analyses temperature scenarios and conditioned on daily

and corresponding forecasts are specially improved circulation patterns is presented and validated. Im-

with the extension of the assimilation period Ta up plications of the model for impact studies are also

to 8-10 days (Pires et. al. 1996; Swanson et. al., discussed.

1998). One obtains an explicit formula of the

covariance matrix C of the assimilation error,

showing its dependence on the assimilation period Monthly mean forecasts of rainfall and tempe-

length Ta, the flow instability and the temporal rature in Portugal, Joao Rio, Carols Pires (CGUL,

observation density. An approximation of the - Centro de Geofisica da Universidade de Lisboa,

leading eigen structure of C is obtained using the Lisboa, Portugal)

linear and adjoint models. This approximation is

compared to that obtained by statistical estimation We assess the quality of monthly mean forecasts of

of C trough the generation of a large set of different rainfall and temperature in Portugal using a simple

realizations of observation errors. The test is and low-cost statistical model. The statistical pre-

performed for a blocking case in the Atlantic sector dictors are extracted from the NCEP Reanalysis of

where the present is put a few days before the the daily 1000hPa geopotential height field (z1000).

blocking onset. The 4DVAR assimilated solutions EOF analysis is performed on the North Atlantic-

are able to track the blocking transition if Ta is Europe area selecting the leading 10 PCs at the

sufficiently extended to the past. The eigen- forecast time as predictors. The monthly cumulated

structure of C tends to the space spanned by the rainfall and the monthly mean temperature are

asymptotic Lyapunov vectors at the present time. computed in three complementary regions: north,

However the assymptotic limit of C is obtained in center and south, from which we compute the

practice for a finite useful assimilation period Tau distribution terciles. That quantities are then classi-

depending much on the recent dynamical instability fied in one of the three classes: above, normal, be-

of the flow during the assimilation period. Weak low. The forecast is based on the analogue method

instabilities are favorable to small 4DVAR analysis using the angle between the predictors vector as

errors and large useful assimilation periods, similarity function. The search for analogues is

especially if that occurs recently in the assimilation performed in cross validation on a daily basis from

period. We access the analysis errors, data data within a period centered at the same month of

assimilability and predictability for different the forecast time. If the angle is below a given

patterns of dynamical instability, measured by the threshold then that day is considered an analogue.

33

The predicted class is the most populated from the Building the bridge between sequential data assi-

pool of chosen analogues. The forecast method is milation and geostatistics: the stationarity issue,

the one-step PC method presented by Vautard et. al. Julien Snegas, Hans Wackernagel (Centre de

(1996). Contingency tables with forecast/observed Geostatistique, Ecole des Mines de Paris, Fon-

classes are built for each region and variable from tainebleau, France)

which we compute the LEPS (Liner Error In Proba-

bility Space) to access the forecasts quality. This A central issue in spatio-temporal prediction is the

score vanishes for constant or random forecasts and modeling of the spatial covariance. This problem is

it is one for correct forecasts. The results show solved differently in sequential data assimilation

higher scores for the cumulated precipitation, es- and in geostatistics. In the first approach, regardless

pecially in winter months, which is associated to of the algorithm used, the spatial structure is propa-

the occurrence of large-scale patterns favourable to gated forward in time by means of the physical

frontal rainfall. In contrast, precipitation in summer model (Bennett, 1992, Evensen, 1997), so that the

which is mostly associated to local conditions spatial covariance at a given time step is the result

which are not captured by the chosen oredictors. of an initial structure and the dynamics of the

Further work is ongoing as we intend to increase system. In the second approach, the inference of the

the number of data stations and the predictor spatial structure relies on an assumption of spatial

information (e.g. using sea surface temperature in stationarity (or homogeneity) which makes it pos-

the Atlantic). sible to compute an experimental covariance on the

basis of a single realization. However, the assump-

tion of stationarity is often too strong, and several

Sensitivity Studies on Wind Measurement Errors extensions have been proposed, like the Intrinsic

in Mesoscale Model Simulations, Radhika Random Functions (Matheron, 1973) or more

Ramachandran, Prakash M Dolas (Space Physics recently a spatial deformation model (Guttorp and

Laborator, Vikram Sarabhai Space Centre, Sampson, 1994).

Trivandrum, India) The present paper points out the respecting ad-

vantages of these two types of methodologies and

Large errors in radiosonde sounding data find a indicates how they can benefit from each other.

mention in literature from time to time. Pressure From a geostatistical point of view, the update algo-

levels have to be determined accurately for obtai- rithm used in sequential data assimilation can in-

ning heights through the hydrostatic relation, for deed be seen as a very powerful tool to generate a

example. Temperature and humidity along with the nonstationary spatial covariance, liable to take into

winds are the other independent fields measured by account local variations like those induced by

a radiosonde. topography, meteorology, time dependent variables

We have initiated wind forecast studies for the or boundary conditions. Conversely, stationary

south west monsoon regime, comprising the months spatial covariances are of ten used in data assi-

of June, July and August in India .The domain milation to model the system noise, for which at

centre of the study is the coastal station Chennai least the stationarity assumption should be tested.

(13.1 N, 80.3 E), located off the Bay of Bengal. We show how a reasonable use of the physical

The wind regime at Chennai during this period is properties of the variables can serve, in both fields,

characterised by low level westerly jet centred the modeling of the spatial structure, especially

around the height of 2.5 km and the tropospheric when the available data are scarce in space.

easterly jet (TEJ) with core centred around 16 km. Several examples from hydrodynamics will be

The average maximum in the westerly regime is given to illustrate these features.

about 10 m/s while in the TEJ, the average easterly

maximum is about 35 m/s. The upper tropospheric

westerlies extend at least up to 35 km in the Public Debate and Climate Model Products:

stratosphere. TEJ has an azimuth very close to 90 Whereof must one be silent? Leonard A. Smith

with a steadiness factor of 98 % while the low - (Centre for the Analysis of Time Series, London

level westerly jet has an azimuth close to 285 with School of Economics, London, UK)

a steadiness factor of 94 %.

Presently we are conducting 2D and 3D simulations Climate modelling holds distinction in the physical

using the meso- scale Advanced Regional Predic- sciences in that it applies quantitative methods to

tion Systems (ARPS) model. This is a non-hydro- produce detailed predictions which can only be re-

static, fully compressible primitive equation model fined and evaluated in-sample. This raises the justi-

in a generalised terrain following co-ordinate sys- fied scepticism of statisticians, although climate

tem developed by the Centre for Analysis and scientists can hardly be held responsible for the

Prediction of Storms (CAPS), Oklahoma, USA. In time constants set by the problem at hand. They

this study we will discuss the effect of the errors in can, however, be held accountable for which model

the initialising wind field on the wind forecast. products (in particular the level of detail in any

forecast) are supplied to policy makers and to the

public. This paper argues for a robust attempt to

establish which model products might prove re-

alistic (i.e. what are the smallest space and time

averages can be discussed?).

34

A statistician may well ask why it has taken so difference series can be attributed to the candidate

many years to fit the Earth's global mean surface series.

temperature (~128 points) given models with 2^20 A new multiple break points detection procedure

degrees-of-freedom and an uncounted number of has been developed which takes the problem of sig-

parameters? The response, of course, is to appeal to nificance and efficiency into account. The sig-

the Laws of Physics: the number of free degrees-of- nificance and the efficiency are formulated accor-

freedom are many fewer than the number of varia- ding to the conventional statistics related to type

bles; both variables, parameterisations, and para- one and type two errors, respectively. This test

meter values are constrained by "the physics" and obtains not only estimated break points and shift

the resulting simulation yields a realistic reproduc- values, but the corresponding confidence intervals

tion of the entire planet's climate system to within as well. The series can be adjusted by using the

reasonable bounds. But what bounds? exactly? point and interval estimates. Since a MASH pro-

This paper sketches a framework with which to gram system has been developed for the PC, the

quantify those bounds. We define, for example, a application of this method is relatively easy. The

temporal credibility ratio: the smallest time-step in paper focuses on new developments of MASH

the model divided by the the shortest duration over which are connected with two special problems of

which some model variable must be averaged if it is the homogenization of climatic time series. One of

to compare favourably with observations. For a sin- themis the relation of monthly, seas onal and

gle model run under transient forcing scenario, annual series.

there are good reasons that this ratio must be small The problem arises from the fact, that the signal to

on statistical grounds; most of these reasons vanish noise ratio is probably less in case of monthly series

if an ensemble of runs are made. The statistics of than in case of derived seasonal or annual ones.

bounding boxes in moderate dimensional spaces is Consequently the inhomogeneity can be detected

discussed, and the likely number of ensemble mem- easier at the derived series although we intend to

bers required to capture reality is noted. This pro- adjust the monthly series.

vides a direct measure, for example, of the The second problem is connected with the usage of

(in)ability of a model to provide insight on regional meta data in the course of homogenization proce-

changes: if a model cannot capture regional va- dure. The developed version of MASH system

riations in the data on which the model was con- makes possible to use the meta data information - in

structed (that is, in-sample) claims that out-of- particular the probable dates of break points -

sample predictions of such regional averages automatically. The decision rule is based on the

should be used in policy making are vacuous. confidence intervals given for the break points.

Related points will also be touched upon, including

questions such as resource allocation: is it better to

have the best guess available (with no meaningful A data assimilation method with a coupled ocean-

quantitative estimate of its reliability whatsoever), land-atmosphere model and its impact on the

or to estimate the likely range of outcomes, further ocean state and on medium-range pre-dictions,

qualified by evidence of how well this particular Clemente A. S. Tanajura, Konstantin P. Belyaev

event (climate variable) can be simulated in this (Laboratorio Nacional de Computao Cientifica

particular model? (MCT/LNCC), Petrpolis, RJ, Brazil)

Ben P. Kirtman (Center for Ocean-Land-Atmos-

phere Studies (IGES/COLA), Calverton, MD,

Multiple Analysis of Series for Homogenization USA)

(MASH). Seasonal Application of MASH. Auto-

matic using of Meta Data, Tamas Szentimrey A data assimilation method based on the well-

(Hungarian Meteorological Service, Budapest, known Kalman filter scheme is considered with the

Hungary) COLA coupled ocean-land-atmosphere model. This

method is applied to assimilate in situ observations

The MASH (Multiple Analysis of Series for Homo- of surface and subsurface temperatures from the

genization) method is a relative homogeneity test PIRATA project. Results of the assimilation pro-

procedure that does not assume the reference series duce a new corrected temperaturesalinity ocean

are homogeneous. Possible break points and shifts state, which can be used to estimate different

can be detected and adjusted through mutual com- oceanographic quantities. In particular, the present

parisons of series within the same climatic area. paper contains new reconstructed meridional heat

The candidate series is chosen from the available fluxes, and focus on the tropical Atlantic Ocean.

time series and the remaining series are considered One-month forecast experiments during 1999 have

as reference series. The role of series changes step been performed. The impact of the assimilation on

by step in the course of the procedure. Several the forecasted temperature profiles of the upper 500

difference series are constructed from the candidate meters is investigated.

and weighted reference series. The optimal weigh-

ting is determined by minimizing the variance of

the difference series, in order to increase the

efficiency of the statistical tests. Providing that the

candidate series is the only common series of all the

difference series, break points detected in all the

35

Assessing Model Skill in Climate Simulations, or 30 years period; and frequency of extremes

Karl E. Taylor (Program for Climate Model Diag- defined as 10 % of the empirical distribution.

nosis and Intercomparison (PCMDI), Lawrence Four considerations are quantified, namely the

Livermore National Laboratory, Livermore, CA, long-term changes, characterised by variations of

USA) the 10 years statistics; the representativity of pre-

vious 30 years periods; the effect of homo-

Although skill scores in the weather forecast genisation; and the spatial synchronity of extremes

community are routinely used to assess model per- at the 10 stations in Hungary.

formance, this practice is rather uncommon in the

climate modeling community. Skill scores have Some typical questions from the 4 x 4 ones:

been proposed, but there is little agreement or even * Which decades exhibited much more (less)

understanding of the relative merits of various record settings than the theoretical 1/N ?

scores. Most skill scores are related in some way to * Did the frequency of extremes change

the common second-order statistics that charac- monotonously or in correlation to the global trends

terize the degree of pattern correspondence between ?

simulated and observed fields (e.g., correlation, * Did the homogenisation change the distribution

RMS difference, and ratio of variance). A family of of new records or the 10 % extremes ?

new measures of model skill is proposed, which * Can we estimate the monthly ranges from

allows one to specify the relative importance placed previous (i.e. 10 years old) 30 years periods ?

on succcessfully simulating the correct amplitude * Do the new monthly record setting occur para-

of variations compared to obtaining a high llelly in such a relatively small area ?

correlation. The scores account for the limits to

potential agreement between models and data that For preliminary conclusion, one can establish that,

result from unforced (internal) variabilty and obser- although even the 30 years periods exhibited con-

vational errors. The use of these skill scores is siderable variability in Hungary, neither the fre-

illustrated by applying them to output from the quency, nor the departure from the mean increased

Atmospheric Model Intercomparison Project with time or with the global temperature in the past

(AMIP), in which models were forced by observed century.

sea surface temperatures and sea ice cover. An

assessment of the relative merits of about 20

models participating in the project will be given, Homogenized Daily Temperatures for Trend Ana-

both in terms of their ability to simulate the lyses in Extremes over Canada, Lucie Vincent,

climatological mean state and the year to year Xuebin Zhang, Barrie Bonsal and Bill Hogg

deviations from that state. An ensemble of inde- (Climate Research Branch, Environment Canada)

pendent realizations of the simulated climate ob-

tained from a single model will then be used to A database of long-term and homogenized monthly

assess the statistical signficance of apparent diffe- temperatures has been created for climate change

rences in model skill and to estimate fundamental analyses over Canada. A technique based on re-

limits to agreement that must result from unforced gression models was developed to identify "in-

variability. The limitations of using a single mea- homogeneities" in individual annual maximum and

sure to quantify a model's skill in simulating an ob- minimum temperature time series using sur-

served field will also be discussed, and the use of a rounding stations. The inhomogeneities were often

new diagram to simultaneously show several mea- step changes caused by factors such as changes in

sures of skill will be described. the station exposure, location, instrumentation, and

observing program. Using this technique, it was

possible to divide the tested series into homo-

geneous segments at the identified steps. Monthly

Extreme Value Analysis adjustment factors, derived from the regression

models, were applied to bring each segment into

agreement with the most recent homogeneous part

XXth Century Behaviour of Temperature and

of the series.

Precipitation Extremes in Hungary, Jnos Mika,

Recent work has involved the adjustments to daily

Tams Szentimrey (Hungarian Meteorological

maximum and minimum temperatures. Reliable

Service, Budapest, Hungary); Judit Bartholy, Rita

daily datasets are essential for the accurate des-

Pongrcz (Etvs Lrnd University, Dept.

cription of spatial and temporal characteristics of

Meteorology, Budapest, Hungary); Lszl Szeidl

extreme temperature related variables. Adjustment

(Department of Mathematics, University of Pcs,

of daily temperatures has raised several concerns

Hungary)

mainly due to the large variability in day to day

temperature, particularly at high latitude stations.

Monthly extremes are analysed at ten stations

An approach was investigated using linear inter-

representing 93,000 sq. km of the country. Ten and

polation between "target values" derived from the

30 years statistics of four characteristics are

monthly adjustment factors. This approach has two

parallelly computed and evaluated. These are: the

main advantages. First, it doesnt require the use of

mean; the year of new record setting; the range

a daily reference series (which is often difficult to

between the two extremes of opposite sign in the 10

produce), and second, the average of the daily

36

adjustments over a particular month is equal to the Environment Protection, Charles University,

monthly adjustment factor. Therefore, Prague, Czech Republic)

homogenized daily temperatures compatible with

the homogenized monthly datasets can be obtained. Extreme weather and climate events severely in-

To assess the impact of these daily adjustments on fluence ecosystems and human society. High and

extreme temperatures, trends in the percentage of low temperature extremes are among the most fre-

days above (below) the 95th (5th) percentile have quently investigated extreme events. Nevertheless,

been computed for maximum and minimum daily relatively little work has been done as regards ex-

temperatures both before and after adjustment. tremes of surface temperature in GCM and down-

Seasonal trend maps were then examined for scaling studies even though it appears to be clear

various periods of time. Preliminary results show that impacts of climate change would result mainly

that the inhomogeneity having the greatest impact from changes in climate variability and extreme

on the trends during the past five decades is events. This study concentrates on the comparison

associated with the 1961 change to observing of 20- and 50-year return values of daily maximum

window at principal stations. An overall assessment and minimum temperatures in (i) observation, (ii)

of Canadian temperature extremes indicates no GCM simulated control climates (ECHAM3 and

consistent trends in the number of extreme hot CCCM2 GCMs are employed), (iii) statistical

summer days, but significant trends toward fewer downscaling from observation, and (iv) statistical

days with extreme low minimum temperatures du- downscaling from GCMs. As for the area involved,

ring the last century. several sites in central Europe and the nearest

gridpoints corresponding to the stations are

analyzed. Observations cover the period 1961-

The analysis of extreme temperature distributions 1990.

on the example of Poland, Joanna Wibig The downscaling method used is a stepwise mul-

(Department of Meteorology and Climatology tiple regression of 500 hPa height and 1000/500

University of Lodz) hPa thickness gridpoint values over most of Europe

and adjacent Atlantic Ocean. Two methods of

The surface air temperature of the world has enhancing the downscaled variance to become

warmed by about 0.3-0.6oC since the 19th century. equal to that observed are compared, namely the

This warming trend should be associated with the inflation of variance and white noise addition.

variation of daily extreme temperatures. There are Since downscaled temperature series reproduce the

indications of the rise of the daily minimum and observed means and variances, for a fair

maximum temperature in some areas. Brzdil at al. comparison between GCMs and downscaling the

(1996) have shown that in Poland the linear in- distributions of GCM-produced temperatures were

crease in mean seasonal daily maximum tempe- re-sized to have the observed mean and standard

rature is slightly higher than that of daily minimum deviation.

temperature in all seasons but autumn during the Extreme value analysis is performed by fitting the

period 1951-1990. Wibig (2000) has shown the Gumbel distribution and/or the generalized extreme

increase of minimum temperature at some stations value (GEV) distribution to the sample of annual

in Poland in a longer period. Therefore the changes extremes of maximum and minimum temperature.

are not uniform even on the area of Poland. The Various methods of the estimation of parameters of

question arise: has the frequency of extreme events extreme value distribution are intercompared,

changed during the present century? If so, the including the method of L-moments and maximum

extremes in temperature can have significant social likelihood method.

and economic impacts. Nicholls et al. (1996) states The comparison among GCM-simulated, down-

that although there is no evidence for global scaled and observed 20- and 50-yr return values of

changes they probably exist on the regional scale. temperature does not appear to be sensitive to what

These changes can manifest not only as a shift of statistical method of the estimation of parameters of

the mean value but also as variation in other extreme value distribution is used, although in-

parameters of the extreme values distribution. The dividual return values are influenced by the choice

aim of this paper is to fit the theoretical distribution of the method. The skill of both GCMs in repro-

to the records of daily extreme temperature and the ducing extreme high and low temperatures is

of extreme temperature for the longer period limited. The statistical downscaling from GCMs

(season, year) and analyze the long-term course of tends to improve the results although it generally

parameters of such distributions. Several distri- yields extremes that are too moderate compared to

butions and fitting methods were tested and the observed values (this holds for downscaling both

results were compared. from observation and GCMs). White noise addition

in downscaling from observation leads to more

realistic return values of maximum and minimum

Comparison of 20- and 50-year return values in temperature than variance inflation.

GCM-simulated, downscaled and observed tem-

perature series, Jan Kysel1,2, Radan Huth1

(1Institute of Atmospheric Physics AS CR, Prague, Test on the probability distribution of maximum

Czech Republic; 2Dept. of Meteorology and air temperatures, Jana Jurekov1, Jan Picek2,

Jaroslava Kalvov1* (1Charles University in

37

Prague, 2Technical University in Liberec, Czech with negative temperature anomalies from 1961-90,

Republic; *Dept. of Meteorology and 2) The global number of hot days (number of days

Environment Protection, Faculty of Mathematics with temperatures exceeding the 90 percentile from

and Physics, Charles University, Praha, Czech 1961-1990) and the number of cold days (number

Republic) of days with temperatures being lower than the 10

percentile from 1961-1990 and 3) The percentage

If we are interested in the extreme values of tem- of world area with annual temperature exceeding

peratures or precipitation, the tails of the underlying the 90 percentile from 1961-1990 and the per-

probability distribution are more interesting than its centage of world surface area with annual tem-

central part. Hence, a specific test on the shape of perature being lower than the 10 percentile from

tails is desirable. After a collection of observations, 1961-1990.

we would first like to decide whether the

distribution is light- or heavy-tailed, and what can

be the value of its Pareto index. More specifically, Bayesian Statistical Modelling of Nonlinear

our problem is to decide whether the summer daily Climate Processes, E. P. Campbell, Y. Li (CSIRO

maximum temperatures at the Czech Republic Mathematical and Information Sciences, ,

stations belong to the domain of attraction of the Wembley, Australia)

Frchet rather than of the Gumbel extreme value

distributions. In accordance with the previous Nonlinear statistical methods have enjoyed rapid

investigation of the authors, we suppose that the development during the 1990s, and offer great

vector of the summer daily maximum temperatures potential as a tool for climatologists seeking rela-

X(n) follows the autoregression model of order 1, tionships of predictive value. A key issue still to be

[Image] (1) resolved however is predictor selection. In the time

t=0, 1, 2, , where { et } is a sequence of series context this requires selection both of optimal

independent random variables, identically dis- predictors and their lags. A powerful new metho-

tributed with distribution function F and density dology for model selection within a Bayesian

(innovation density) f, generally unknown. The framework, known as reversible jump Markov

hypothesis states that - log(1 - F(x)) ~ m log x as chain Monte Carlo, has emerged in the statistical

x , where 0< m m0, m0 prescribed. A finite literature in the late 1990s. We show how this

value of m means that the distribution function F methodology can be used to identify and fit non-

has a heavy right tail and belongs to the domain of linear time series models, and illustrate the

attraction of the Frchet distribution. The proposed approach using threshold models for monthly rain-

test is based on the tail behavior of the extreme fall. We show how our approach can be generalised

autoregressive quantiles of the model (1). The to select nonparametric nonlinear time series mo-

asymptotic null distribution of the test criterion is dels.

normal and the test is consistent against the

alternative of lighter tails as well as of the

exponential tails. The test is applied to the series of Modelling and characterizing extreme drought

the daily maximum temperatures at the Czech events, Jess Abaurrea, Ana Carmen Cebrian

Republic station in the period 1961 - 1998. (Dpto. Metodos Estadsticos, Universidad de

Zaragoza, Ed. Matemticas. Pedro Cerbuna, 12.,

Zaragoza, Spain,)

Three Approaches to determine extreme years of

Global Atmospheric Temperature, Luis Gimeno1,*, The aim of this work is to develop a stochastic mo-

Juan Ael 1, Higinio Jorge1, Pedro Ribera1, David del to analyze the occurrence and severity of

Gallego2, Ricardo Garcia2, Emiliano Hernandez2 meteorological drought events. The description of

(1Universidad de Vigo. Facultad de Ciencias de the characteristics of droughts by probability distri-

Orense, Orense, Spain; 2Universidad Complutense butions provides measures such as the inter-drought

de Madrid. Facultad de Ciencias Fisicas, Madrid. recurrence time, the expected duration or deficit

Spain; *Corresponding author) and the return values, which are helpful in water

resource management.

Global atmospheric temperature is usually com- In order to define a drought event we use an opera-

puted from the average of a combination of land air tional definition based on the comparison of a sto-

temperature anomalies and sea surface temperature chastic process, s(t), describing the hydrological

anomalies . Both components of the dataset are ex- state of the system and a threshold, u(t), which re-

pressed as anomalies from 1961-90, as this makes presents a critical level for the process. For the

computation much easier. In this work a daily 2.5 analysis of long-term meteorological droughts

latitude x 2.5 longitude grid box temperatures presented in this work, we use as signal the

from NCAR-NCEP reanalysis at three pressure monthly moving annual rainfall. The proposed

levels (850, 500 and 200 hPa) and for the period model, based on results from Extreme Value

1958-1998 were used to determine extreme years of Theory and Excess over Threshold methods, is a

global atmospheric temperatures by means of three compound Poisson cluster process, CPClP, which

different approaches: 1) The difference between the combines a Poisson Cluster process, to represent

world surface area with positive temperature ano- the drought occurrence, with a vector series

malies from 1961-90 and the world surface area composed by three random variables -duration,

38

deficit and maximum intensity- to describe the and the generalised von Mises-Jenkinson distri-

drought magnitude. Six Spanish monthly rainfall bu-tion, but allowing for a linear variation of the lo-

series of about one hundred years long are studied cation parameter with time. The maximum like-

to check this model. lihood equations are solved by numerical methods.

The fitted CPClP model is also used as the basis to The maximums of the likelihood achieved can then

characterize the largest drought event to occur in a be compared. As an example we use a 100 years

given period of time. This characterization requires time series of the air temperature in Lisbon where

the knowledge of the distribution of the maximum there is a very significant increasing trend.

in samples of random size. Results about this issue

in samples with Poisson size are presented for some

of the most frequent distributions in environmental The Possible Influence of Large Volcanic Erup-

applications. tions on El Nino, Philippe Naveau (National

Center for Atmospheric Research, Boulder, USA)

Extremes in Instationary Time Series, Jrgen The El-Nino Southern Oscillation (ENSO) is the

Grieser (Institute for Meteorology and largest source of inter-annual climatic variability on

Geophysics, University Frankfurt, Germany) a global scale. To determine if large tropical erup-

tions influence the ENSO by introducing an ex-

Extreme values of meteorological variables threaten ternal forcing associated with a cooling period, or if

human life and property. Thus, knowledge of the deviations from the El-Nino cycle are only due to

statistical features of extremes is important. The random noise is of primary interest to the paleo-

concerned variables are the waiting time distribu- climate and climate change communities. Hence,

tion, return period and risk. These variables are the aim of this talk is to investigate the statistical

easily obtained if the underlying meteorological association between the ENSO and volcanic erup-

variable is independent and identically distributed. tions.

Otherwise the probability to exceed a threshold is To study the historical behavior of El-Nino events

not constant. It is demonstrated that the widely used over a long period of time 1525-1987, we use the

return periods are not a useful variable in this case. data set constructed by Quinn and Neal (1995). To

However, if the meteorological variable itself has a assess volcanic activities from 1525 to 1991, we

certain temporal structure also the probability to work with the VEI (Volcanic Explosivety Index)

exceed a threshold changes in time in a determined that describes the occurrences and the intensity of

way. volcanic eruptions (Newhall and Self 1982, updated

It is shown that the statistical features of extremes by Simkin and Siebert 1995).

can be drawn from the statistical features of the To investigate the existence of a potential link be-

meteorological variable itself. The latter one may tween these two signals, we develop a variety of

be taken from observations or from model runs. statistical models that can take into account the

Two real world examples demonstrate that the special characteristics of our data (e.g., volcanic

statistical features of extreme values may change eruptions seem to occur in clusters). The most

considerably even if the mean features of the under- complex model is a Hidden Markov Model (HMM)

lying meteorological variable have only slightly that is based on two main components: the ob-

changed. servations and some hidden atmospheric states. The

advantages of such a process is its flexibility and its

capabilities to model a variety of dependence struc-

Comparison of Several Methods for the Esti- tures between the ENSO and VEI. Finally, to

mation of Extreme Value Distribution in Time remove the risk of over-parameterization, we

Series with Trends, Henrique Oliveira Pires, Joao compare different HMM's to simpler models

Pestana Ferreira (Instituto de Meteorologia, (independent Bernouilli processes and Markov

Lisboa, Portugal) Chains) by using model selection techniques.

extreme value statistics, where the number of peaks

is modelled by a Poisson distribution and the value Simulation Models

over the threshold by an exponential distribution is

compared with methods based on extreme value

Simplified atmosphere models with realistic vari-

asymptotic distribution theory.

ability, Ulrich Achatz (Leibniz-Institut fuer

Estimation is made using Gumbel distribution (ex-

Atmosphaerenphysik an der Universitaet Rostock

treme value Type I) with just a location and a dis-

e.V., Kuehlungsborn, Germany); J.D. Opsteegh

persion parameters and, also, using the generalised

(KNMI, De Bilt, The Netherlands)

von Mises distribution where there is also a shape

parameter; von Mises-Jenkinson distribution re-

The realism of climate simulations by general cir-

duces to Gu mbel when this shape parameter is

culation models is partly due to their use of an

null.

enormous number of degrees of freedom. Unfor-

The influence of trends, which is essential in a

tunately this leads to an (ever increasing) loss of

changing climate, is evaluated by fitting, using

transparency of specific model results and low

maximum likelihood estimation, the Gumbel model

integration speeds, hampering both progress on

39

basic understanding of the climate system and the fractal pattern of rainfall time series. The

applicability of GCMs for climate impact assess- combination of these analysis of stochastic point

ment studies. There is therefore a need for com- process models and the scale of fluctuation will

plementary climate models of low or intermediate facilitate to improve the disaggregation possibilities

complexity which can reproduce the atmosphere's and to identify the limits of applicability of point

behavior based on simplified descriptions. In the process models.

light of its important role it seems desirable that

internal climate variability is also captured. Up to

now there is no low- or intermediate-complexity Statistical Modelling of Tornado Occurrence,

climate model which satisfies this demand. Here Harold E. Brooks (NOAA/National Severe Storms

progress might be brought by so-called reduced Laboratory Norman, Oklahoma)

models utilizing realistic dynamics in combination

with an optimal set of basis patterns and a good Tornadoes are rare events at any location. Possible

parameterization of the impact of scales and changes in the distribution over time could have

processes not explicitly resolved. The patterns are significant effects on public safety. Unfortunately,

chosen such that they can give a rather complete the quality of data collection has changed drama-

description of the climate attractor but simul- tically over the years in many countries (e.g., an-

taneously neglect highly improbable realizations of nual tornado reports in Germany have increased by

unrealistic states. A good candidate for this are a factor of 5-10 in the last 50 years). As a result,

empirical orthogonal functions (EOF). Achatz and changes in the "true" distribution of tornadoes can

Branstator (1999) have recently shown that the be hard to distinguish from changes in the repor-

empirical etermination from a GCM data set of ting. It appears, however, that the conditional distri-

linear parameterizations of the effect of ageos- bution of tornadoes by intensity is relatively

trophy, unresolved vertical and horizontal scales, consistent in many countries for tornadoes rated F2

and unresolved physics can improve a reduced or higher on the Fujita damage scale, decreasing

quasigeostrophic two-layer model projected onto logarithmically with increase F-scale value. This

EOFs such that it is able to reproduce the internal consistency allows us to make some estimate of the

variabiity of the GCM not only in midlatitudes but degree of underreporting in different locations. For

also in the tropics. In oder to also include the instance, it appears likely that only about 15% of

nonlinear tropical dynamics we have built on this French tornadoes are currently in their database.

result and developed a corresponding model class A 75-year record of significant tornadoes (those

based on the primitive equations. Comparison with rated F2 or higher) in the United States has

the climatology of a conventional GCM provided input to a variety of statistical models of

(ECHAM3) shows good agreement. Models based tornado occurrence that allow us to estimate the

on 500 and as few as 30 EOFs will be discussed range of variability. The first model is a Markov

with respect to the simulation of the seasonal chain model of occurrence of at least one

dependence of mean state and fluxes, and to the significant tornado anywhere in the United States

reproduction of recurrent anomalies. on a daily basis. It provides estimates of the

probability of long runs of consecutive days with or

without tornadoes. The second model uses an

Modelling of Temporal Rainfall by the Use of Sto- estimate of the probability of a significant tornado

chastic Point Process Models in View of Scaling, on a daily basis on a grid in the US, based on a

Konrad Bogner (Institute of Hydrology, University simple kernel density estimation technique. Those

of Agricultural Research,Vienna, Austria) daily probabilities are then modelled using a beta

distribution, and the beta distributions used to

Clustered rectangular point process models and va- generate a long model run that provides estimates

rious modifications are investigated in detail regar- of tornado occurrence and, of particular interest, the

ding the assessment of fit. Besides testing the interannual variability.

agreement between observed properties and those

used in fitting the model, but examined at different

levels of aggregation, the generalized extreme value A method to extend the temporal validity of nested

analysis is applied to compare simulated and regional climate simulations, Udo Busch and

observed extreme rainfall. In former publications a Dietrich Heimann (DLR Oberpfaffenhofen,

lack of fit for extreme values of hourly aggregation Institut fr Physik der Atmosphre, Germany)

levels is reported sometimes. However the in-

vestigated modifications, like the incorporation of This study introduces a statistical-dynamical extra-

third order moments in the fitting procedure, polation scheme (SDE) for nested regional climate

indicate an improvement and better agreement for model simulations. The method is based on a

small time intervals also. Furthermore assessment cluster analysis weather-type classification scheme.

of fit is made for simulated series of sub-hourly Predictor is the 500~hPa geopotential height and

time resolution considering the extreme values. predictands are the regional surface temperature

There are considerable differences between obser- and precipitation.

ved and simulated properties at these levels of For the validation of the scheme a 30-year GCM

aggregation. Therefore the method of wavelet trans- simulation of the Hadley Centre with a contin-

form is applied to reveal the self similarity and uously nested regional climate model (RCM) is

40

used. The basic features of the regional develop- model parameters (mean precipitation amount for

ment in Central Europe are quite well reproduced wet days and the probability of a wet day following

by the SDE. A comparision of the SDE and the a wet day) are not very well estimated over 1901-

direct RCM output shows only a deviation of 1930 interval. This result could be explained by the

0.13~K for the decadel mean winter temperature fact that before 1931 these parameters describe

and 8~\% for the decadel mean winter precipitation another precipitation regime (as urbanization effect)

of the greater Alpine Region. The SDE estimation or these parameters are no so strongly dependent on

of the interannual variability of the mean winter the large-scale circulation. Testing of model for

precipitation shows an excellent reproduction for other months (seasonally stratified) and other

the North Sea area and a slight underestimation for station placed in same geographical conditions are

the greater Alpine area. in progress.

The results of the SDE are two dimensional

available with the resolution of the nested regional

climate model. With the application of the SDE the Statistical Downscaling of Daily, Multi-Site Pre-

regional climate model has to be nested only for a cipitation in Tropical and Sub-Tropical Climates,

1 1

few years to get 30-year climatological statistics. Stephen P. Charles , Bryson C. Bates , Stephen J.

This render the possibility to obtain regional scale 2 3 1

Crimp and James P. Hughes ( CSIRO Land and

information for the length of GCM experiments. 2

30-year GCM time-slice experiments with different Water, Wembley, Australia; Queensland Depart-

CO2 concentrations could be compared within the ment of Natural Resources, Indooroopilly, Austra-

3

regional scale. The comparison and validation of lia; Department of Biostatistics, University of

regional models with different model physics and Washington, Seattle, USA)

parametrizations is possible for a higher number of

model runs including the possibility to cover time Regional climate simulation is becoming increa-

spreads of climatological periods (i.e. 30-years). singly important as the scientific and policy

community require realistic projections of the

potential regional impacts of climate change. Alt-

Stochastic generation of daily precipitation, hough general circulation models (GCM) and limi-

Aristita Busuioc1, Hans von Storch2 and Rodica ted area models (LAM) simulate large-scale atmos-

Tomozeiu1 (1 National Institute of Meteorology pheric processes reasonably well, they tend to over-

and Hydrology, Bucharest, Romania; 2 Institute estimate the frequency and under-estimate the

of Hydrophysics, GKSS Research Center, intensity of daily precipitation, and thus fail to

Geestacht, Germany) reproduce the observed precipitation statistics at the

spatial and temporal scale required for regional

The paper presents a stochastic model conditioned impacts assessment. Statistical downscaling can

by the large-scale climatic distributions to generate overcome these limitations by relating regional-

daily precipitation amount. With such a model, or point-scale precipitation to large-scale atmos-

numerous statistically equivalent but in its details pheric fields. The nonhomogeneous hidden Markov

different realization of long sequences of daily model (NHMM) is a statistical downscaling model

rainfall can be generated. The model uses a first previously shown to reproduce daily precipitation

order Markov chain combined with a downscaling probabilities, spatial patterns in precipitation

model based on the canonical correlation analysis occurrence, wet and dry spell length statistics, and

(CCA) to link the local precipitation with large- the probability distributions of precipitation

circulation. Precipitation occurrence is described by amounts at multiple sites in temperate midlatitude

a two-state, first-order Markov chain. The pre- regions (Pacific north-west USA and south-western

cipitation either occurs or it does not (the two Australia).

states) and the conditional probability of pre- Here, we apply the NHMM to tropical and sub-tro-

cipitation occurrence depends only whether preci- pical regions in Australia where precipitation is

pitation occurred in the previous day. There are two generated by a much wider variety of mechanisms

parameters describing the precipitation occurrence including monsoon, coastal and upper troughs,

process: the transition probabilities of a wet day tropical cyclones, fronts, and localised convective

following a dry day and the probability of a wet day systems. The literature contains very few examples

following a wet day. Variation of precipitation of statistical downscaling in tropical regions. A

amount on wet days is characterized using gamma suite of observed atmospheric predictors (derived

distribution, which has two parameters: the shape from NCEP/NCAR Reanalysis data), encompassing

parameter and the scale parameter. Therefore the measures of atmospheric flow, moisture content,

stochastic precipitation generation model depends and stability, provide NHMMs that reproduce the

on four parameters. To link the four parameters by observed summer and winter precipitation statistics

the large-scale circulation, represented by the sea for tropical, sub-tropical continental, and sub-

level pressure on the European scale, a regression tropical maritime regions in Queensland, Australia.

model based on the canonical correlation analysis Future work will compare downscaled simulations

have been used. This model has been tested for the generated from the observed atmospheric predictors

winter (January) daily precipitation amount at the with those generated from the corresponding LAM

Bucharest station over 1901-1994. It was found atmospheric predictors.

that, for the independent data set (1901-1950), two

41

Stochastic Weather Generator Model for Simu- Met&Roll is a WGEN-like four-variate (precipi-

lation of Multiple Meteorological Time Series, tation, solar radiation, temperature maximum, tem-

Ildik Dobi-Wantuch, Jnos Mika (Hungarian perature minimum) stochastic daily weather

Meteorological Service, Budapest, Hungary); generator used to provide synthetic weather series

Lszl Szeidl (Department of Mathematics, for models simulating growth of crops and forests

University of Pcs, Hungary) and hydrological regime in present and changed

climate conditions. The previous version of the

This work presents the development of a daily, generator was found not to satisfactorily reproduce

stochastic weather generator point-model, which is interdiurnal and interannual variability of the

based on precipitation occurrences, as a driving weather series. Two questions are raised in this

parameter. The model, simulating diurnal series of contribution: (1) What is an effect of these im-

maximum, mean and minimum temperature, preci- perfections on results obtained by the models

pitation, sunshine duration, cloudiness, relative (growth models, hydrological models) feeded by

humidity, wind speed and air pressure, is re- the synthetic weather series? (2) How to improve

commended for use in climate change impact the generators performance?

studies in Hungary. In the first part, improvements of Met&Roll aiming

The first problem, to solve, is the suitable modeling to better reproduction of statistical structure of the

of wet and dry sequences, since, as we are going to weather series are introduced: (i) To better repro-

discuss, some often used simulation approaches do duce interdiurnal variability of the weather series,

not appropriately work in our climate. Considering correlations among solar radiation and daily ex-

the frequently occurring long dry periods in Hun- treme temperatures are allowed to vary during a

gary, duration of sequences without measurable year and the Markov chain of the third order (in-

precipitation is modeled by mixed Poisson and stead of the first order) is used to model preci-

geometric distributions. pitation occurrence. (ii) As the generator was found

Another important aspect is the preservation of the to underestimate interannual variability of monthly

first four moments and the correlation structure of means, the underlying mathematical model was

the empirical time series. In order to retain these split in two steps: In the first step, time series of

statistical features for all variables, we applied monthly means of the four variables is generated

various uni- and multi-variate methods. using first-order four-variate autoregressive model.

To prepare the data vectors for further modeling, In the second step, daily weather series is generated

first we standardized the series considering the using the monthly means produced in the first step.

annual cycle of the means and standard deviations Impact of the new features of the generator on the

by applying Fourier transformation. Next, a centra- statistical structure of synthetic daily weather series

lization for conditional averages, depending on the will be demonstrated.

serial number of the given day within the dry or wet In the second part, effect of the interdiurnal and

series, is performed, since this dependence is also interannual variability of daily weather series on

significant in several respects. crop yields simulated by growth models (CERES

Factor analysis and test of normality suggest that and WOFOST) and on stream hydrology and

three of the variables, namely maximum tempera- stream water quality simulated by SAC-SMA

ture, relative humidity and air pressure, can be hydrological model will be shown.

parallelly modeled as a weakly stationary, 3-dimen-

sional process which can be fairly well described

by an AR(1) model. However, the other six climate Temperature predictability in the greater Mediter-

elements have non-normal distributions. This ranean area, R. Garca, E. Hernndez, T. Muoz,

recognition drives us to accept a multivariate exten- P. Ribera* and L. Gimeno* (Dto Fsica de la

sion of the Johnson distribution. This technique Tierra II, Universidad Complutense, Madrid

already matches the necessary statistics for the Spain; Dto Fsica Aplicada, Universidad de Vigo,

given elements and climatic region. Campus de Orense, Orense, Spain)

Interdiurnal and interannual variability in sto- seasonal average temperatures in an area centered

chastic daily weather generator: Modelling and over the Mediterranean Sea has been considered

the role in agricultural and hydrologic studies, from two different approaches intended to forecast

Martin Dubrovsky1, Josef Hejzlar2, Jan Kysely3, annual or seasonal average temperature values se-

Josef Buchtele4, Zdenek Zalud5, Daniel Marvan6 veral months in advance. The methods are

(1Institute of Atmospheric Physics AS CR, Hradec Optimate Climate Normals (OCN) and an extension

Kralove, Czech Republic (CR), 2Hydrobiological of it called Climate Normal Functions (CNF). Both

Institute AS CR and Faculty of Biological of them attempt to forecast temperatures from the

Sciences USB, Ceske Budejovice, CR; 3Institute of past behavior of its own time series. They have

Atmospheric Physics AS CR, Prague, CR; been applied to a dataset of monthly averaged tem-

4

Institute of Hydrodynamics AS CR, Prague, CR; peratures from 151 observatories from Europe and

5

Mendel University of Agriculture and Forestry, Africa distributed around the Mediterranean area,

Brno, CR; 6Faculty of Mathematics and Physics, covering the period 1955-1990.

Charles University, Prague, CR) This study makes use of annually and seasonally

averaged values. For every observatory, OCN and

42

FCN have been computed for each averaging ther generator, C.M. Goodess and J.P. Palutikof

period. Next, cluster analysis has been applied to (Climatic Research Unit, University of East

classify similitudes in the predictability charac- Anglia, Norwich, UK)

teristics. Three predictability patterns have been

identified: in the Northern area and 8 yr period has Empirical methods for downscaling GCM output

been detected; the North Atlantic Oscillation seems offer relatively simple and computationally effi-

to be the most important influence. In the Eastern cient means of constructing scenarios with a high

area the predictability coefficients reach their spatial and temporal resolution. However, rela-

maximum at the longest average periods. Finally, tively few downscaling studies have addressed the

the aera closest to the Mediterranean shows a one problem of constructing scenarios which are consis-

year persistence as the most characteristic feature. tent between sites and between variables on a day-

by-day basis. Here, a three-stage method for the

construction of self-consistent daily temperature

A new validation scheme for the evaluation of cli- and rainfall scenarios which has been developed

mate model outputs, F.-W. Gerstengarbe, M. and tested for two study areas in southeast Spain

Kcken, P.C. Werner (Potsdam-Institute for and southern Italy is described.

Climate Impact Research, Potsdam, Germany) First, daily rainfall scenarios were constructed for a

key site in each study area using a conditional

Model validation plays an important role in the weather generator in which rainfall occurrence de-

development of climate models. The validation en- pends on both the circulation type and on whether

compasses a simple optic comparison, the use of the previous day was wet or dry. Each simulation

generalising measurement figures (e.g. root mean set consists of 1000 runs, i.e. a Monte Carlo

square error - RMSE) and the application of fully approach is taken. This has a number of advan-

developed statistical methods. The latter often re- tages:

quire relatively high calculation time so that they

are rarely used for evaluation purposes. Never- a range of scenarios can be provided and the

theless, the evaluation of complex correlations can level of uncertainty can be indicated, for

only be done with such complex methods. In order example by calculating quantile values;

to make progress here, a method is presented in the ranked pairs can be used as climate scenarios,

following which is based on a multivariate pattern thus taking some account of systematic errors

recognition and provides measurement figures for in the downscaling models and underlying

quality evaluation that can be easily interpreted and GCMs; and,

whose spatial distribution and temporal develop- non-ranked pairs can be used in climate sen-

ment serves to derive information on possible sitivity studies, for example to maximise the

sources of error. changes in a particular season.

The question that should be answered by the new

method is: How well can a model represent defined Second, daily rainfall scenarios for a network of

complex structures of a climate regime? sites in each study area were constructed using an

In a first step the validation conditions are defined analogue approach, based on a reference scenario

and the selected parameters are grouped in patterns for each key site, i.e. by sampling from the ob-

for a reference field at each grid point of a model served data conditional on the circulation type and

area via n time steps by means of a cluster analysis whether the key site was wet or dry.

algorithm. It has then to be investigated how far the Third, multi-site daily temperature scenarios were

same parameter combination of the simulated field, constructed using a transfer function method, with

i.e. the climate model output, reflects these patterns. free atmosphere variables as predictors in multiple

This is done by a distance measurement which regression models. The principle of consistency is

provides information about the assignment of the maintained by constructing transfer functions sepa-

grid points to the clusters of the reference field. rately for wet days and dry days. The latter infor-

Thus, one receives a yes/no answer: the simulated mation was available from the multi-site rainfall

value either corresponds with the one of the scenarios.

reference field (the grid point belongs to the same The final scenarios are considered more plausible

cluster as in the analysis ) or the same grid points of than the raw GCM changes. This approach is, how-

the reference field and the simulated field belong to ever, affected by a number of problems which are

different cluster. In order to specify this yes/no common to all empirical downscaling approaches

answer, each degree of deviation, standardised to (i.e. stationarity, overdispersion, data availability

0;1 (0 no deviation, 1 largest deviation with a and the reliability of GCMs). A number of ways in

statistical security of 95%) will be determined in which the method could be further refined and

addition. developed to reduce these problems are proposed.

The effectiveness of the method is presented by a

comparison of the quality of different regional cli-

mate models.

scenarios for Mediterranean regions based on

Monte Carlo simulations with a conditional wea-

43

The most dominant spatial mode reveals a dipole

Statistics, Models and Data structure whereas the second depicts a tripole

Assimilation spatial configuration. The two spatial modes of

principal component analysis are used in a

minimum variance cluster analysis (MVCA) to

Reconstructing large-scale variability from paleo- delineate areas where CCD is homogeneous in

climatic evidence by means of pattern nudging, space. Here, the cubic clustering criterion (CCC) is

Martin Widmann 1, Hans von Storch 1, Ingo used to determine the number homogeneous region.

Kirchner 2 (1 GKSS Research Center, Geesthacht, Based on CCC and MVC, four regions of convec-

2

Germany, Max-Planck-Institute for tive activity regions are defined. The defined con-

Meteorology, Hamburg, Germany) vective regions show different temporal charac-

teristics of the duration of convective activities over

Pattern nudging is the second step of the newly the country. The convective activity regime with

developed Data Assimilation Through Upscaling maximum CCD is located over western and central

and Nudging (DATUN) technique. The aim of parts of the country. Whereas, the region charac-

DATUN is to obtain a physically-based, best guess terised by minimum occurrence of convective

for the large-scale states of the atmosphere during clouds entails northern, northeastern, southern,

the Late Holocene with annual temporal resolution. southeastern parts of Ethiopia.

DATUN uses statistical upscaling models to esti-

mate continental-scale and hemispheric climate va-

riability from climate proxy records. In a second Skill of seasonal hindcasts as a function of the

step, which is the subject of this presentation, the ensemble size, F.W. Zwiers and V.V. Kharin

large-scale variability of a GCM is nudged towards (Canadian Centre for Climate Modelling and

these estimates. Analysis, Victoria, BC, Canada)

Commonly used nudging procedures relax model

states towards a target field that is prescribed at Forecast skill as a function of the ensemble size is

every gridpoint. However, this approach is proble- examined in a 24-member ensemble of northern

matic in the context of paleoclimatic reconstruc- winter (DJF) hindcasts produced with the second

tions, since the historic atmospheric states are only generation general circulation model of the Cana-

incompletely known. To overcome this difficulty, dian Centre for Climate Modelling and Analysis.

we developed a nudging method that is purely These integrations are initialized from the NCEP

defined in a pattern space. Only large-scale patterns reanalyses at 6 hour intervals prior to the forecast

are nudged and the synoptic-scale variability is season. The sea surface temperatures that are ap-

therefore expected to be undamped but consistent plied as lower boundary conditions are predicted by

with the large-scale forcing. Pattern nudging has persisting the monthly mean anomaly observed

been implemented in the ECHAM4 model and prior to the forecast period. The potential pre-

tested using prescribed states of the Arctic Os- dictability that is attributed to lower boundary

cillation. An evaluation of these test runs will be forced variability is estimated.

presented. This work is undertaken as part of the The forecast skill in the first 2 weeks, which ori-

German project KIHZ (Klima in Historischen ginates predominately from the initial conditions, is

Zeiten, Climate in Historical Times). greatest for relatively small ensemble sizes. The

forecast skill increases monotonically with the

ensemble size in the rest of the season. The skill of

Spatio-temporal patterns of Meteosat Satellite DJF 500 hPa geopotential height hindcasts in the

Thermal Infrared Derived Convective Clouds over Northern Hemisphere and in the Pacific/North

Ethiopia as Described by PCA & MVC, Abebe America sector improves substantially when the

Yeshanew (Senior Meteorologist, National ensemble size increases from 6 to 24. A statistical

Meteorological Services, Addis Ababa, Ethiopia) skill improvement technique based on the Singular

Value Decomposition method is also more suc-

Convective clouds with top temperature at -40 C cessful for larger ensembles.

thresholds are derived from Meteosat satellite ther-

mal infrared channels over Ethiopia from July to

September fro the year 1988, 1989 and 1990. The

duration of convective activity (cold cloud duration, Extreme Value Analysis

CCD) maintaining its top cloud temperature at -40

C threshold are summed up to every dekad (8, 9, 10

or 11 days depending on the month considered). Determining extreme sets in multivariate distri-

The standardised values of the duration of convec- butions, onika. Lakatos (Hungarian

tive activity is subjected to principal component Meteorological Service); I. Matyasovszky

analysis to cull the most dominant and important (Department of Meteorology, Etvs Lornd

signals from the dataset. The first two principal University, Budapest)

components are selected based on the break point

of the eigenvalues verse the eigenvector The term extremes is simply defined as events

components plot. The first two principal having small probabilities at the tail(s) of an un-

components explained 83% of the total variance. derlying density function. Consider a multivariate

44

random variable with probability density function Climate extremes in Germany and its relation to

f(x). For any 0<<1 an extreme set T is defined as the objective weather type classification of the

German Weather Service, Peter Bissolli, Udo

T :{ x ; f ( x ) < c } such that f ( x )dx = , Baum (Deutscher Wetterdienst, Offenbach,

Germany)

T

where c=c(). When the analytical form of f is

Climate extremes can be attributed to special fea-

kownn the task to obtain T is straightforward.

tures of certain weather types. For the area of

Without knowing the form, which is a typical case,

Germany, such weather types can be described

an estimate of the density is needed. Therefore, a

using the objective weather type classification of

procedure is developed to determine T using a

the German Weather Service. This classification

kernel type estimator for f. The methodology is

provides 40 standard weather types, but, by using

applied to air temperature and wetness data in order

special indices, it is possible to define other special

to define critical dry and warm conditions.

extreme weather types too.

The Probability Of An Extreme In A Correlated

1. how to define climate extremes in Germany,

Time-Space Domain, Albert R. Boehm

2. how to define extreme weather types,

3. how to find relations between extreme

Consider an area with weather observations at a

weather types and climate extremes.

limited number of points and over a limited period

of record. Questions: What is the probability during

The definition of these extreme events is mainly

a specified target period in the future that there will

based on frequency distributions of different cli-

be:

mate elements (daily temperature maximum and

A: An extreme anywhere in the area?

minimum, daily wind speed maximum, daily

B: A specified fraction or more will experience an

precipitation amount and daily snow depth) and

extreme?

weather type indices (indices of advection, cyclo-

C: An extreme will persist during some fraction of

nality and humidity). In addition, the spatial varia-

the target period?

bility of climate elements within Germany is taken

D: Extremes will exist beyond some penalty

into account. Advantages and disadvantages of

function which could be complicated and is fitted to

these defined extremes are discussed, from a statis-

specific requirements such as crop growth or

tical point of view as well as from a climatological-

heating load.

practical one.

The relation between weather types and climate

Case 1: The area is homogeneous and correlation is

extremes can be defined by probability distri-

isotropic in space and stationary in time. This case

butions. Considering that the weather type classi-

can be handled by fitting the probability of an

fication also provides forecasts of weather types,

extreme at point and modeling correlation structure.

the forecast skill of these probability distributions

Questions A, B, C and some forms of D can be

can be estimated.

answered with the CUB algorithm defined at the

5IMSC 1992. CUB depends on the eigenvalues of

the domain correlation matrix. More complicated

Statistical methods and case studies investigating

forms of D can be simulated.

extreme climatology of the Carpathian Basin,

Case 2: The area is heterogeneous and the

Judit Bartholy1, Laszlo Szeidl2 (1Dept. of Meteo-

probability may have a seasonal variation, but cor-

rology Eotvos Lorand University, Budapest Hun-

relation is isotropic in space and stationary in time.

gary; 2Dept. of Mathematics University of Pecs,

In effect, the spatial and time domain consists of

Hungary)

sub regions that differ in their extreme probability

but are correlated. Since the heterogeneous pro-

One of the most challenging topics of the global

bability will be less than if all the domain were at

change research is predicting future changes in

the highest probability but greater than if all the

extreme climatological values. An overview of the

domain were at the lowest probability, a weighted

statistical methodology will be presented that can

mean can be used as an approximation. Since the

provide scientific and reliable answer to the follo-

correlation is assumed known, a conditional pro-

wing questions: how the extreme values and their

bability of a lesser sob-domain being extreme given

occurrences can be modelled, what kind of pro-

that a greater sub-domain is not extreme can be

blems may occur investigating climatic fields of a

calculated.

given meteorological parameter, etc. The presen-

Case 3. The domain is undergoing climate change

tation will focus on the extremes of climate of the

or fluctuation. This complicated case has an elegant

Carpathian Basin. Monthly temperature data obser-

solution: Take into account all possible changes

ved at 16 stations will be examined, furthermore,

(including no change) each weighed by the pro-

precipitation measurements from 162 and 8 stations

bability it will occur.

located in Hungary will be used in monthly and

daily scale evaluation, respectively. In this region

the precipitation trends evaluated during last

decades seems to be inconsistent. Beyond extreme

45

value analysis of precipitation, variability and fluc- in the statistical characteristics of the two types of

tuation characteristics will be determined. Spatial data.

distributions will be mapped of various statistical

characteristics for all seasons. Time series will be

also analyzed separately as robust and extreme. Modelling the probability of precipitation using

Distributions of extreme, robust and full time series the Marcov chain MonteCarlo technique, Martin

will be compared. Asymptotic approach will be Kappler (Fachbereich Statistik, Universitt Dort-

applied to precipitation and temperature time series mund); Andreas Hense (Meteorologisches Institut

using Gumbel distribution model. Model para- Universitt Bonn); Siegfried Schach (Fachbereich

meters will be estimated by method of moments. Statistik Universitt Dortmund)

Return values will be determined for pre-set

periods, e.g., 10, 20, 50 years, which thresholds will Although precipitation simulated in atmospheric

be exceeded once in average during that period. models is calculated deterministically, the inter-

More detailed investigations will be carried out for actions between the rain generating processes in

the two sensitive regions of the Carpathian Basin, space and time require a probabilistic interpre-

namely, the drainage basin of the shallow Lake tations of precipitation forecasts. This is especially

Balaton and Sio, and the Hungarian Great Plain. In true for high resolution weather forecast models

the present climatic conditions the water balance of like the Lokal-Modell LM of the German Weather

these regions tends to be negative, and therefore it Service DWD which has presently a horizontal

is likely that the global change may affect them the resolution of 7 km. We will show results for

most. forecasts of hourly accumulated rain fall from the

LM as input to a postprocessing model which

describes rainfall events as outcomes from Ber-

noulli trials with an unknown probability (PoP

Simulation Models probability of precipitation) to be estimated from

the data.

The postprocessing model is set up as a hierarchical

Simulation of multivariate time series of meteoro-

Bayesian network with three levels. The logit-trans-

logical data, Asa Forsman, Claes Bjrklund and

formed Bernoulli probability (level 1) is para-

Anders Grimvall (Department of Mathematics,

metrized as a normally distributed variable where

Linkping University, Linkping, Sweden)

the expectation is parametrized (level 2) linearily in

terms of a-priori chosen variables like orographic

A great variety of models used in the environmental

height, orographic gradient in wind direction and

sciences are driven by meteorological data. This

the precipitation field in the neighborhood of the 1

implies that simulations involving synthetic mete-

hour backward trajectory. Lastly the regression

orological inputs can play an important role in

coefficients (level 3) are characterized by the

several different contexts. First, such simulations

uninformative normal density function with expec-

can be used to determine the joint distribution of an

tation zero and a large variance. The Bayesian

arbitrary set of response variables. Secondly,

network and the specification of the above given

simulations followed by statistical analyses of the

conditional and priori densities allows the calcul-

response to different inputs can be used to elucidate

ation of the a-posterior density for the unknown

the dynamic behaviour of complex models. In the

parameters conditioned on the data of simulated

present paper, we discuss how the statistical charac-

precipitation. The parameter estimation is done by a

teristics of a given multivariate time series of me-

Gibbs sampling of the a-posteriori density with the

teorological data can be transferred to arbitrary

help of a Monte Carlo simulation (Markov chain

amounts of synthetic data. In particular, we de-

Monte Carlo). Data are takenonly from a quarter of

scribe a procedure involving the following steps: (i)

the model domain such that the remaining parts can

transformation of the marginal distribution of each

be used for validation. The estimated values for the

of the meteorological variables to a standardised

regression coefficients show that the gradient in

normal distribution; (ii) fitting of a vector auto-

wind direction and the backward trajectories are the

regressive model to the transformed data; (iii)

most important external variables. Taking the Brier

simulation of new observations according to the

skill score for the validation shows that the post-

fitted auto-regressive model; (iv) transformation of

processing model is clearly better than persistency

the data thus generated to synthetic meteorological

or a PoP estimated from a simple spatial average.

data. A nonparametric resampling method that

preserves the rank correlations of the different

meteorological variables is also described. In both

cases, seasonality is handled by dividing the ori- Reproduction of variance in statistical

ginal data sets into subsets representing the downscaling: inflation vs. noise addition, Radan

different months of the year. An empirical data set Huth1, Jan Kysel1,2 (1Institute of Atmospheric

comprising thirty years of daily data regarding Physics, Prague, Czech Republic; 2Dept. of

temperature, precipitation, vapour pressure, wind Meteorology and Environment Protection,

speed, and cloudiness was used to illustrate the Charles University, Prague, Czech Republic)

methods proposed. Examination of observed and

Statistical downscaling models can explain only a

simulated data revealed only marginal differences

part of the variance of the original (observed) time

46

series. Therefore, the variance of the downscaled The Amazon climate: A multiscale and nonsta-

series is underestimated. Two methods of enhan- tionary processes, Guillermo O. Obregn, Carlos

cing the downscaled variance to become equal to Nobre (Centro de previso de Tempo e Estudos

that observed are in common use: (i) inflation of Climticos - CPTEC / Instituto Nacional de

variance, consisting in multiplying downscaled Pequisas Espaciais INPE, Cachoeira Paulista,

anomalies by the ratio of the observed and So Paulo Brazil)

downscaled standard deviations, and (ii) noise

addition, consisting in adding a random process The goal of this work is to examine the climatic

(noise) to the downscaled series. The variance complexity of the western part of Amazon basin

inflation appears to be a physically doubtful related to both the ENOS and NAO phenomenon

concept as it assumes that all the local variability based on monthly level of the Negro river at

originates from large-scale variability, which Manacapur-Manaus (Q) from 1903 to 1999. The

apparently is not correct. Morlet wavelet transform (WT) and the probability

The comparison of the two approaches is conducted density function (PDF) are used as a tools. The WT

for daily maximum temperatures in May to Sep- shows that on Amazon basin exist four regimes of

tember and daily minimum temperatures in Novem- quasi-periodic oscillation. The 2-3 yr cycles and the

ber to March at six central European stations for a 5 yr cycles are in agreement with the ENOS signal.

30-year period. The downscaling method used is a The quasi-decadal cycles appears linked to the

stepwise multiple regression of 500 hPa height and NAO, principally before 1940, and the quasi-bi-

1000/500 hPa thickness gridpoint values over decadal cycles, remaining any abrupt climatic

Europe and adjacent parts of the Atlantic Ocean. change in time, seems to be associated with both

The random process added to the time series is a the ENOS and NAO. The PDFs between Q vs SOI,

white noise process. The evaluation is based on (i) and Q vs NAO shown distinctive multimodatity

lag-1 autocorrelations, (ii) standard deviation of indicating the nonlinear response of Q to these

day-to-day temperature changes, (iii) characteristics phenomenona or at least a weak direct relationship.

of persisting extreme events (heat and cold waves),

and (iv) spatial correlations.

It is shown that the downscaling with variance First indication of anthropogenic climate change

inflation reproduces persistence characteristics of in local wave conditions in the North Sea, Arnt

minimum temperature correctly, whereas overesti- Pfizenmayer (Institute for Coastal Research,

mates persistence of maximum temperature. The GKSS Research Centre; Geesthacht, Germany)

addition of a white noise process results in a severe

underestimation of persistence (overestimation of In the central North Sea we observe an increase in

day-to-day variations) in both seasons. Similar the frequency of eastward propagating waves in the

results apply to spatial correlations. recent four decades. To assess the significance of

We argue that although the variance inflation is not this change the wave statistic for this century was

correct from a physical point of view, it results in reconstructed with a statistical model. With a linear

much less biased temporal and spatial autocorre- multivariate technique (redundancy analysis)

lations than if a white noise process is added, and monthly mean air pressure fields over the North

therefore, inflation should be given preference if Atlantic and Western Europe were downscaled on

one is concerned with temporal and spatial structure the intramonthly frequency of directional wave

of downscaled quantities. propagation. When compared against this reference,

the recent change appears statistically significant at

the 5% level.

Simulating meteorological data by block re- In order to investigate the reason for this local cli-

sampling, Anders Nordgaard, Asa Forsman, matic change, reconstruction was compared with

Anders Grimvall (Department of Mathematics, the downscaled results of a control and transient

Linkping University, Linkping, Sweden) GCM scenario (ECHAM4-OPYC3) and with re-

sults obtained in a high resolution time slice experi-

Validation of statistical models used to simplify ment with increased concentrations of greenhouse

complex mechanistic models needs large-scale data gases and aerosols. Both estimates are qualitatively

generation. In this paper we use a block resampling consistent with the changes observed in the last

method to generate ''new'' series of data from a long decades. We suggest that the recent increase of

series of multivariate meteorological observations. eastward propagation is a local manifestation of

The method is based on the matched-block anthropogenic global climate change.

bootstrap method suggested by Carlstein et al

(1998) where blocks of data are resampled accor-

ding to a Markov chain. The transition probabilities Ensemble Predictions with a Limited-Area Meso-

are computed using a Mahalanobis distance mea- scale Atmospheric Model - Implications for the

sure. Interpretation of its Direct Model Output,

Susanne Theis, Andreas Hense (Meteorologisches

Institut der Universitt Bonn, Bonn, Germany);

Ulrich Damrath, Volker Renner (Deutscher

Wetterdienst, Offenbach, Germany)

47

The atmosphere is a chaotic system. In principle, solation to maximum daily temperature, (2) para-

this aspect has always to be taken into account meters of the third order autoregressive (AR3)

whenever an atmospheric forecast is produced and model fit to the residuals from the insolation model,

interpreted. One possibility of dealing with the (3) variance of the residuals of the AR3 model; and,

intrinsic uncertainty of the forecast is the generation (4) mean and variance of temperature range.

of ensemble predictions. The principle of ensemble Stochastic structure of daily precipitation is repre-

prediction is to forecast the impact of the uncer- sented by: (1) probability of wet, trace, and dry

tainty of initial conditions on the uncertainty of the days, and (2) parameters (mean and variance) of the

forecast. This is usually done by running several (log-normal) distribution assumed for daily preci-

numerical forecasts from slightly different initial pitation and conditioned on the estimated proba-

states. Certainly, one could also imagine to vary bility of wet days.

other parameters than the initial conditions, for The KSU LCM suite of independent variables com-

example the boundary conditions at the ground, prises surface, upwind, and downwind variables

such as roughnesslength or albedo. This method is computed using CCM3 boundary conditions. Sur-

especially interesting in the context of limited-area, face variables represent local topography and ra-

high-resolution atmospheric models which are nes- diation balance. Upwind variables are: radiation

ted within a global model. balance air temperature and its meridional gradient,

The increase of forecast resolution is substantially average gradient in saturation vapor pressure,

governed by the quantitative specification of the speed, div and curl of velocity as well as orographic

boundary conditions at the ground. Many of these controls. Downwind (30 km) variables represent

boundary conditions cannot be quantified with blocking effects.

downright certainty. We have produced ensemble The transfer function was calibrated (using 434

predictions by slightly perturbing the rough- stations in an elevation- and climate-division-

nesslength and the orography of a limited-area stratified sample) and tested (using 1290 stations)

mesoscale weather forecast model (horizontal for climate of the western U.S. and Mexico using a

resolution: 7km). The results reveal that small-scale control run of CCM3 for boundary conditions.

features of only a few gridboxes' size are not This new model allows high temporal (one day)

deterministically predictable. This implies that and spatial (1km2) resolution solutions of climate

small-scale predictions by deterministic regional for much of western North America for any time

climate models might be subject to a considerable period for which CCM3 solutions are available, e.g.

amount of uncertainty, as well. Whenever the direct 6ka paleoclimate and future climate scenarios.

output of a deterministic forecast model contains a Similar solutions can be used for other GCMs when

non-negligible amount of stochastic information, it calibrated with their control runs.

must be interpreted in a statistical way. Ensemble

predictions, however, are not always realizable,

because they demand an enormous amount of com- Numerical experimentation with regional atmos-

puter capacities. As a temporary solution, we have pheric models, Hans von Storch, Ralf Weisse

developed a method which postprocesses the direct (Institute of Coastal Research, GKSS Research

model output statistically. The method requires the Centre, Germany)

data of one numerical simulation only. It will be

validated against experimentally conducted en- While the need for determining the climatic noise

semble predictions. with the help of multiple runs in numerical ex-

perimentation with global models is acknowledged

since the 1970s, regional atmospheric modelers in

Modeling stochastic structure of local climate as the past have designed experiments usually as a

downscaling of GCM fields, 1,2Marina simple pair of "control" and "experimental" simu-

Timofeyeva and 1Rachael Craig. (Department of lation. The effect of the experimental modification

Geology, Kent State University, Kent, Ohio; is then considered as being described by the

UCAR/NOAA/National Weather Service, Silver difference between the two simulations. This

Spring, Maryland) approach assumes implicitly that the state of the

regional atmosphere in such models would be de-

The Kent State University Local Climate Model termined by the boundary conditions, that no

(KSU LCM) is a statistical downscaling procedure "noise" would be generated on regional scales.

that relates a suite of dependent variables (stochas- However, this assumption is invalid as is shown by

tic structure of local climate) using canonical examples. For long-term differences the noise level

correlation analysis to a suite of independent varia- is indeed rather small, but for episodes large dif-

bles (derived from insolation, a DEM and GCM ferences in the synoptic configuration (e.g. with

velocity, geopotential heights and SST). Our me- respect to the speed of cyclones) intermittently

thod differs from the other techniques of down- emerge.

scaling (e.g., Zorita and von Storch, 1999) as it uses The problem is demonstrated with the case of the

a semi-Lagrangian approach to represent the effect of an interactive ocean surface wave field on

regional climate controls at the point where climate the formation of extratropical storms. In previous

is modeled. studies, this interaction was claimed to cause

The stochastic structure of daily temperature is storms to intensify (through increased latent heat

represented by (1) parameters of the best fit of in- fluxes) or damped (through increased momentum

48

flux) cyclogenesis, but a ensemble simulation complex orography and parameterization scheme

indicates that the found effects are mostly random. for convection and land surface processes in the

model.

Anthropogenic Climate Change over the past 1400 years, K. K. Andersen and P. D.

Ditlevsen, (Department of Geophysics,

Copenhagen, Denmark)

Exploiting the emerging signal of anthropogenic

climate change in probabilistic climate fore- In the accumulation zone of the Greenland ice sheet

casting, Myles Allen (Space Science & the annual accumulation rate is registered in the firn

Technology Department, Rutherford Appleton and subsequently in the ice. The annual layer

Laboratory, Chilton, Didcot, UK) thickness may be identified through identification

of the annual cycle in the isotopic climate signal

and other annually varying parameters. The snow is

Significance of Anthropogenic Radiative Forcings however subject to wind drifting which introduces

in Simulation of Temporal Pattern of Surface Air rather large 'glaciological noise' to the signal.

Temperature in Monsoon Climate, Ajay Singh We developed a method to estimate the noise in

(School of Management, Indian Institute of accumulation records derived from ice cores. This

Technology- Bombay, Powai, India) method was then applied to the annually resolved

accumulation records from four independent deep

Augmentations in greenhouse gas concentrations ice cores. In this way we have extracted a pre-

since pre-industrial era have led to a positive cipitation record for Greenland covering the period

radiative forcing of climate, tending to warm the from 555 AD to 1973 AD with decadal resolution.

surface and to produce other changes of climate.

Tropospheric aerosols have led to a negative direct

forcing and possibly also to a negative indirect Lessons taught from the Homogenization of 200

forcing of a similar magnitude. Primarily the US First Order Stations, Auguste C. Boissonnade,

negative forcing is focused in particular regions and Ph.D.1 and Larry Heitkemper2 (1Risk Management

subcontinental areas, it can have continental to Solutions, Inc. Menlo Park, California, USA;

hemispheric scale effects on climate patterns. All 2

EarthSat Corporation, Washington D.C., USA)

India Summer Monsoon plays very critical role in

many of socio-economic activities in the Indian Weather conditions have an enormous impact on

Subcontinent. Surface air temperature in the agricultural, energy, and indirectly retail

monsoon climate is important for number of rea- businesses. Study of historical weather

sons including its significance in agriculture, observations is one of the means for assessing the

industry and planning. An accurate and reliable likelihood of potential impact weather could have

portrayal of change in the temperature due to cli- on these businesses. The most common weather

mate change is therefore very important in the index, Cooling degree-day or Heating degree-day,

perspective of impact of climate change on socio- used in assessing weather impact is derived from

economic system and environmental system. the daily minimum and maximum temperature at a

The ability of current generation of global climate given well-manned station over a 5 to 7-month

models, in their long-term simulations, to replicate period. New financial instruments called weather

the observed atmospheric behavior on a wide range derivatives, based on the likely future behavior of

of spatial and time scales provides support in the weather index, have been developed and are

applying these models to the regional climate currently traded in financial markets for hedges

change projections induced by anthropogenic against the weather risk. All these instruments are

radiative forcings. The ability of a coupled atmos- very dependent on historical and future

phere ocean climate model (ECHAM3 + LSG) at observations of weather data at given weather

T21 resolution to simulate the temporal pattern in stations.

area-averaged annual mean surface air temperature Large volatility is present in such assessment of

over Indian subcontinent in three different numeri- weather risk due to the large uncertainty in pre-

cal experiments viz.; control, GHG and GHG plus dicting weather beyond a few days but also in the

aerosol is examined. The significance of anthro- uncertainty on the quality of recorded observations

pogenic radiative forcings in the simulation of for the purpose of weather risk assessment. His-

temporal pattern of the temperature has been exa- torical temperature data may contain missing or

mined. The model has superior skill in reproducing erroneous values. Furthermore, changes in location

the observed trend and interannual variability in and/or instrumentation for most weather stations

area-averaged annual mean surface air temperature have caused increase or decrease in temperature

during the past century in GHG plus aerosol experi- observations. Such discontinuities can make his-

ment. The models ability to simulate the observed torical data misleading relative to how weather is

pattern in the temperature over the area of interest being recorded under present day-conditions.

could be improved by better representation of A 1 1/2 year joint project between Risk Mana-

meso-scale forcings such as monsoon trough, gement Solutions and Earth Satellite Corporation

49

was undertaken for homogenization of time series in the November/January signals seems to be more

for more than half of the available First and Second related to Arctic atmospheric processes such as 70

Order stations in the US over the time period days oscillation identified in the geopotential height

(1950-1999). Results of the homogenization study at 500 mb level. The 70 days oscillation might be

were reviewed with the local meteorologists in triggered or locked up by the seasonal cycle in the

charge. special context of orographic conditions and air-

This paper presents findings of this project, a land interaction over high latitudes areas.

comparative analysis of several methods for de-

tecting and assessing the magnitude of discon-

tinuities and present conclusions regarding the What influences decision-makers in drought

quality of available meta databases. In addition, a prone areas when allocating water resources?

summary of the impact of instrumentation changes Bernard Cornlis (Dpartement de Gomatique

over the past three decades is presented. Read- Universit de Lige)

justment of climatological data for particular loca-

tions is also presented. Emphasis on the importance At an operational level, computerised decision

and problems of homogenization of such time support systems implement Bayesian statistics,

series in weather risk is made. Suggestions for fuzzy logic, neural networks, genetic algorithms,

further research with potential benefits in a better data based or model based management systems,

assessment of weather risk are also outlined. Nevertheless, these concepts barely leave the tech-

nical and scientific circles. When it comes to hu-

man decision makers, the results of these processes,

North Atlantic osccillation projection on regional if they reach them, is just one more information

climate predictability, Roxana Bojariu and which can be taken into account or not. In the field

Daniela Paliu (National Institute of Meteorology of climatology, just like in major environmental

and Hydrology, Bucharest, Romania) issues, lack of information are important, and

induce a great uncertainty. The four techniques

NAO is the dominant mode of variability over the used to appreciate uncertainties of consequences

Atlantic-European region. A number of studies are probabilities, imprecise probabilities,

have shown that most of cooling in the northwest possibilities and unknown possibilities. In this case

Atlantic and the warming across Europe and over as well, not all the decision makers have access to

Eurasia since the mid 1970s results from changes in such systems or trust them.

the NAO. However, a more detailed regional pro- This contribution can be seen as an attempt to for-

jection of NAO related climate fluctuations and malise the way current decision makers deal with a

predictability is still needed. potential climatic event. Different types of private

In the present study, a NAO related behavior is and public decision makers, ranging from farmers

identified when the predictive potential of winter is to Members of the European Parliament, were

assessed using a multifield analog technique based asked during an interview to allocate water re-

on large and local climate state vectors. The sources to their territory of action. Through this

hindcast experiment have been done for the interval pragmatic exercise, several ideas were tested:

1962-1989 using seasonal means of temperature would the spatial characteristics of their territory

and precipitation anomalies from 62 Romanian influence the repartition; would their decision

stations, geopotential heights at 500 mb over the behaviour change under uncertainty; would the

Atlantic-European sector and sea surface tempera- integration in time modify their answer. After com-

tures (SSTs) from the North Atlantic. The autumn pleting the allocation scenarios, they were asked to

thermal anomalies in a large part of Romanian comment their decisions. Besides, several neutral

territory are significantly well predicted from the questions relating to the use of water, to drought

previous winter. A larger scale perspective of this and to their territory were asked. Analysis of this

behavior is investigated by performing canonical survey shows that, instead of basing their decisions

correlation analysis (CCA) for air surface tem- on math models and measured info, lay decision

perature over Europe, in November and sea level makers base their decisions on their moral values

pressure (SLP) over Northern Hemisphere in Ja- and their representation of the issue.

nuary. Temperature and SLP data for the interval

1961-2000 are extracted from NCEP re-analyses.

This source of autumn predictability over Ro- Assessing uncertainty in anthropogenic signal

manian regions seems to be related to the inter- estimates using space-time and trend-based

annual persistence of one NAO, more pronounced optimal detection approaches, Nathan Gillett,

in the latest decades of 20th century. Gabi Hegerl, Myles Allen

Many studies have shown that both phases of NAO

are associated with basin-wide changes in the The optimal fingerprinting' approach to the de-

intensity and location of the North Atlantic jet tection of anthropogenic climate change involves

stream and storm track. Changes in the intensity the regression of some observed diagnostic onto the

and location of North Atlantic jet stream may be same diagnostic derived from perturbed inte-

caused by processes such as changes in sea surface grations of a GCM. Both time/space and trend-

temperature and/or in the extent of ice sheet at high based diagnostics have been used in published

latitudes. The NAO physical mechanism involved studies. While results from both approaches ge-

50

nerally indicate that the influence of greenhouse We ask whether these indices provide similar mea-

gases and sulphate aerosol is detectable in global sures of cyclone activity and what each index

surface temperature over the past fifty years, they suggests about trends in cyclone activity during

have lead to different estimates of the amplitude of 1949-1999. The indices are Eady growth rate at

these `signals'. The goal of the present work is to 500 hPa, temperature variance at 500 hPa, the 95th

understand these differences in results. A first step percentile of surface wind speeds, and counts of

is to transform those results that are based on both storms and intense storms. Eady growth,

stepwise regression to multiregression. By elimi- temperature variance, and extreme winds are rea-

nating other differences in the analysis, relating to sonably well-correlated over the storm-track re-

the data used and filtering, we establish the relative gions and northern parts of North America and

importance for detection of the linear trend and Eurasia, but poorly correlated elsewhere. These

residual components of the response to the forcings. three indices also show moderately-strong corre-

An additional reason for differences in results based lations with the two storm count indices over much

on different model simulations relates to the differ- of the storm-track regions, but only when the storm

ing indirect sulphate aerosol response (that due to index is offset 10 degrees to the north of the other

changes in cloud properties) in the ECHAM4 and index. We analyze simple linear regressions of the

HadCM3 models. By examining the spatial and indices against time using the False Discovery Rate

temporal characteristics of this response in each (FDR) procedure. Eady growth, temperature va-

model, we better characterise the causes of dif- riance, and extreme winds show significant in-

ferences between published results on the detection creases over the mid-latitude storm-track regions as

of anthropogenic sulphate aerosols, and assess well as over parts of Eurasia, while temperature

which differences are sensitive to the detection ap- variance shows decreases over the polar latitudes of

proach used. the North Pacific. Extreme winds at the sigma 995

level of the reanalysis model and geostrophic winds

calculated from sea level pressure show unex-

Applications of Statistical Climatology in Finance, plained jumps in the 1950s that account for most of

Stephen Jewson, Anders Brix (Risk Management the increases seen in the wind index. Poisson

Solutions) regressions of the two storm count indices against

time are suggestive of increases over the oceanic

Statistical models of meteorological phenomena are storm-tracks, but it is difficult to distinguish these

used widely in the fields of banking, insurance and increases from statistical chance. Given the diffi-

re-insurance to help estimate both climatological culties with the extreme winds index, it is difficult

and state-dependent financial risk. An overview is to draw conclusions about human impacts from our

given of methodologies used in this area, including results, apart from the evidence for increased

very large ensemble simulations, optimal combi- cyclone activity over the storm-track regions.

nations of climatological and forecast information, Results from Global Climate Models (GCMs) that

and final metric optimisation of forecasts. compare control and enhanced CO2 scenarios

A number of examples are given including: suggest poleward and downstream shifts in the

1)Real-time hurricane loss forecasting: the need for storm-tracks with increased CO2. The models also

ensembles of size in excess of 10,000 and how suggest the possibility of fewer but stronger

these ensembles can be created; how deterministic cyclones. The Eady growth and temperature va-

forecasts can be used to condition these ensembles. riance indices provide some evidence for increased

2)Extratropical wind-storm risk estimation: the downstream cyclone activity over the storm-track

distribution of numbers of wind-storms and the regions. The storm count indices do not suggest a

importance of seriality. decrease in the number of events, but provide weak

3)Weather derivative pricing methods: daily and in- evidence for an increase in more intense events

dex models for weather derivative pricing and the over the storm-tracks.

pros and cons of different methods for incorpora-

ting forecasts.

Signal analysis of the atmospheric mean tem-

perature 500/1000 hPa north of 55 N between

Trends in northern hemisphere winter cyclone 1949 and 1994, Heiko Paeth, Andreas Hense

activity based on a comparison of cyclone indices, (Meteorologisches Institut, Universitt Bonn,

Christopher J. Paciorek (Department of Statistics, Bonn, Germany)

Carnegie Mellon University); James S. Risbey

(Department of Engineering and Public Policy, The lower tropospheric mean temperature 500/1000

Carnegie Mellon University); Richard D. Rosen hPa is examined in the northern hemisphere high

(Atmospheric and Environmental Research, Inc. latitude region north of 55 ffi N with regard to a

Cambridge, Massachusetts); Valerie Ventura (De- climate change signal due to anthropogenic climate

partment of Statistics, Carnegie Mellon forcing as a supplement to previous studies which

University) concentrated on near surface temperatures. An

observational data set of the German Weather

We investigate northern hemisphere extratropical Service (DWD) is compared with several model

winter cyclone activity using a variety of indices simulations including different scenarios of green-

calculated from the NCEP/NCAR reanalysis data. house gas and sulphate aerosol forcing derived

51

from the two recent versions of the coupled climate A Maximum Entropy Approach for Extracting

model in Hamburg, ECHAM3/LSG and ECHAM4- Forecast Probabilities from Finite Ensembles,

/OPYC. The signal analysis is based on the optimal M.S. Roulston1, L.A. Smith1 and C. Ziehmann2

fingerprint method by Hasselmann (1979,1993), (1Mathematical Institute, Oxford University, UK;

2

which supplies a detection variable with optimal Institute of Physics, Potsdam University, Ger-

signal-to-noise ratio. The natural variability mea- many)

sures are derived from the corresponding long-term

control experiments. From 1970 onward, we find A maximum entropy method is presented for in-

high trend pattern analogies between the observa- corporating prior information, in the form of a

tional data and the greenhouse-gas induced model climatological probability distribution, into a pro-

simulations. The fingerprint of this common tem- babilistic forecast generated by an ensemble predic-

perature signal consists of a predominate warming tion system. The approach is to find the probability

with maximum over Siberia and a weak cooling distribution that is as close as possible to the clima-

over the North Atlantic reaching an estimated sig- tology, in the relative entropy sense, while being

nificance level of about 99 %. The additional for- consistent with the ensemble according to the chi-

cing by sulphate aerosols decreases the correlation squared test. The result is that low-probability

of this climate change simulation with the ob- events, too rare to be expected in an N member en-

servations. The natural variability constitutes about semble, are assigned finite probabilities based on

30 % of the conforming trend patterns. The signal- climatology.

to-noise ratio is best over the oceans while the An information theoretic skill score is also intro-

tropospheric temperatures over the land masses are duced. This `ignorance' score is equal to the number

contaminated by strong noise. There is evidence bits that are required to describe truth in a data

that the temperature signal is independent of the compression scheme designed around the probabi-

initial conditions and the model version. listic forecast. It is also related to the rate of return

achieved by a gambler betting optimally (propor-

tionally) on possible outcomes.

Expert Assessment of Uncertainties in Detection The maximum entropy method is illustrated in se-

and Attribution of Climate Change, James Risbey veral chaotic systems. Both perfect and imperfect

(University of Utrecht, Department of Science, ensembles are considered, the latter both under per-

Technology and Society, Utrecht, The fect and under imperfect models. It is shown that

Netherlands) this method can reduce the variability in forecast

skill obtained when ensembles are small. The

The issue of whether, and to what extent, human- average forecast skill, as reflected by Brier Score or

induced increases in greenhouse gases have caused by the ignorance is largely unaffected. The re-

climate changes is contentious. Studies of this issue duction in variance is achieved by systematically

have focused first on `detecting' climate change improving the worst forecasts. This reduction

against the background of natural variability, and occurs both with perfect and with imperfect en-

further on `attributing' any detected signal to in- sembles.

creases in greenhouse gases or other possible cau-

ses. Communication of the science and role of

uncertainties on this issue has been hindered by the Coupled sea ice and atmosphere variability in the

lack of explicit formal approaches for making over- Weddell Sea, Antarctica, Silvia A. Venegas

all conclusions on detection and attribution. We (Danish Center for Earth System Science

have developed a protocol to quantify uncertainties (DCESS), Niels Bohr Institute for Astronomy,

in each component of the detection and attribution Physics and Geophysics, University of

process and to provide a structured way to make Copenhagen, Denmark); Mark R. Drinkwater

overall conclusions. Here we present results on (Oceans/Sea-Ice Unit, ESTEC Earth Sciences Di-

detection and attribution of climate change from vision (APP-FSO), Noordwijk, The Netherlands)

analysis of expert judgements obtained using the

protocol, highlighting areas of convergence and A frequency-domain singular value decomposition

divergence among experts. We show that, among a is performed on 20 years (1979-1998) of satellite

broad spectrum of experts, probabilities of detec- observations of sea ice concentration and drift

tion are uniformly very high (> 95%) for some lines together with sea level pressure reanalysis data in

of evidence and high (> 85%) for others. Further, the Weddell Sea, Antarctica, in order to determine

there is near consensus that greenhouse forcing is the dominant timescales of variability in the region.

responsible for more than half the warming in The interannual timescale (periods of 3-4 years) is

global mean temperature this century. For the other found to dominate the ice and atmospheric fluc-

lines of evidence examined, greenhouse forcing tuations. This variability is a regional manifestation

makes smaller contributions with more spread a- of the Antarctic Circumpolar Wave in the Weddell

mong expert assessments. sector of the Southern Ocean.

Interannual variations of sea ice concentration and

drift are driven largely by anomalous atmospheric

patterns, which alternatively enhance or reduce the

meridional wind circulation to the east of the

Antarctic peninsula. Thermodynamic processes (ra-

52

pid ice growth in open water areas), dynamic pro- climate change indices. The approach is particularly

cesses (ice ridging/rafting due to ice convergence) appropriate for indices whose statistical uncertainty

and geographic constraints all play significant roles can be estimated only subjectively, such as sea le-

in the formation of summer ice concentration vel, sea ice, glaciers, extremes, etc. As example we

anomalies in the southern Weddell Sea every 3-4 have chosen indices based on annual means and the

years. These ice anomalies are exported northwards annual cycle of surface temperature and precipi-

out of the Weddell basin approximately one year tation and the diurnal cycle of temperature, as well

later, and then advected eastwards before diffusing as ocean heat content. The results are compared

in the ice margin. with the conventional D&A approach.

A low-frequency signal (presumably a quasi-deca-

dal oscillation) is also detected, although its signi-

ficance is fairly speculative in a 20-yr record. Ice Changes in statistical properties of daily

variability on this timescale results from a change precipitation in transient climate change

in the shape of the Weddell gyre circulation which experiments with a coupled atmosphere-ocean

occurred around 1990, likely associated with large- GCM, V.A. Semenov and L. Bengtsson (Max

scale variations of the ocean-atmosphere system. Planck Institute for Meteorology, Hamburg,

Germany)

Change Detection and Attribution, Klaus bution are analyzed using data of the coupled at-

Hasselmann, Reiner Schnur (Max Planck In- mosphere-ocean ECHAM4/OPYC3 general circu-

stitute of Meteorology, Hamburg, Germany) lation model transient climate change experiments.

Two different climate forcing mechanisms are

In the conventional optimal fingerprinting approach applied. In the GHG experiment the forcing is due

to detection and attribution (D&A) of anthropo- to the changing concentrations of the greenhouse

genic or other externally forced climate change, the gases only. In the GSDIO experiment in addition to

observations are projected onto a low-dimensional the GHG the direct and indirect effects of the

phase space spanned by one or a few expected cli- anthropogenic sulfur aerosol and changes in the

mate change signal patterns which are rotated away tropospheric ozone are taken into account. Statis-

from high natural climate noise. A detection test is tical properties of daily precipitation for different

performed by assessing whether the null hypothesis seasons are described by parameters of a simple

that an observed value in this reduced space is with- statistical model based on the gamma distribution.

in the interval of natural variability can be rejected The efficiency of the gamma distribution model is

at a certain confidence level or not. This test is most estimated. Global trends of the mean precipitation,

powerful in a one-dimensional sub-space. For the precipitation intensity, wet day probability and the

attribution of a detected change one tests whether gamma distribution parameters are analyzed for the

the detected climate change is consistent with the present (1950-2000) and future (2000-2050) cli-

predicted signals of alternative candidate mecha- mates.

nisms. The number of signal patterns must be kept Particular attention is paid to the extreme preci-

small to achieve statistical significance. The un- pitation events. Simulated characteristics for the

certainties of the predicted signals enter only in the present climate are compared to the available ob-

attribution test. servations. For the future climate the differences of

Here, an alternative, Bayesian approach is presen- the hydrological cycle and daily precipitation distri-

ted which leads to a basically different optimal fil- bution changes due to different climate forcing

tering strategy. One considers the impact of evi- mechanisms are investigated.

dence (i.e. observed data) on the subjectively

defined prior probability that the hypothesis of an

externally forced climate change signal is true. The Near Two Centuries Monsoon Rainfall Variations

resulting posterior probability depends on the ratio across India: Reconstructed Past and Predicted

of the likelihoods of observing the data for the com- 10-year Future, N.A. Sontakke (Indian Institute of

plementary cases that the hypothesis is true or false. Tropical Meteorology, Pashan, India)

There exists no longer a formal separation between

detection and attribution, and the statistical un- Long period rainfall reconstructions are vital for

cer-tainties of the predicted signal play an equally numerous purposes such as climate variability and

important role as those associated with natural va- prediction, cross comparision with climatological

riability. There is also no need to project onto a proxies and so on. Owing to high spatial variability,

strongly reduced sub-space spanned only by the the representativeness of the summer monsoon

signal patterns of the hypotheses being tested. rainfall for the entire country is limited. In order to

Nonetheless, a technique is presented to weight the develop an effective system for monsoon rainfall

observed data such that the impact of the evidence studies the country has been divided into six zones

on the posterior probability is maximized. This i.e. North West India, (NWI), North Central India

leads also to a dimensional reduction, but one (NCI), North East india (NEI), West Peninsular

which is normally less severe. india (WPI), East Peninsular India (EPI) and South

Results are presented for the simultaneous appli- Peninsular India (SPI). Empirical Orthogonal Func-

cation of the Bayesian D&A framework to several tion (EOF) and Cluster Analysis of ubiquitous 306

53

raingauges has been carried out for this classifi- It can be seen, that the total explained and the

cation. By applying an objective technique 'selec- GHG-attributed variance is maximum in case of

ting a subset of few gauges whose mean showed the global averaged temperature data and is reduced in

highest CC with the all gauges mean series' total less spatial averaged and other than temperature

116 optimum number of gauges have been fixed for data (and is minimal in case of European pre-

real time updating of the different zonal series and cipitation).

reconstruction from 1871. The different zonal series The detection of the anthropogenic greenhouse sig-

are extended as far as possible backward prior to nal suceeds in the global mean temperature data in

1871 by applying the objective technique on the 1973 with a probability of 99.9%. In European data

limited meager available gauges. The NEI and EPI NAO is dominant. It is an open question whether

series are extended back to 1848, the NWI to 1844, there exists a GHG-NAO coupling on the European

the NCI to 1842, the WPI to 1817 and SPI to 1813. scale, which aggravates the detection of the GHG

The all-India series from 1871 onwards is prepared signal here. On the other hand the detection of the

from the are-weighted mean of the six zones and GHG signal suceeds much better in the area mean

well agrees with the series prepared with all the data on the global scale. In 42 out of 72 areas the

gauges. The series is extended back to 1813 by probability for a climate change exceeds 95%.

objective technique on the avalilable gauges. The Solar forcing reveals to be weak in most cases.

examination of these series up to 1999 in detail

reveals interesting results in relation to the chan

pattern of monsoon circulation in the present Trends in the Diurnal Temperature Range in the

century with respect to global warming. CCCma Coupled Model, Daithi A. Stone, Andrew

The low frequency smoothed series up to 1999 J. Weaver (School of Earth and Ocean Sciences,

obtained through binomial low-pass filters plus the University of Victoria, Canada)

series from 2000-2009, filled with the long term

mean is subjected to Variable Harmonic Analysis Recent analyses of the global surface temperature

(VHA). VHA is a modification of classical har- record have revealed that late 20th century in-

monic analysis. An objective technique is applied creases have been characterised by the daily mini-

to select few harmonics whose linear combination mum temperature increasing at a faster rate than the

showed a correlation coefficient of ~ 0.85 with the daily maximum temperature. These differential

smoothed series. The rainfall amount generated trends have resulted in a decrease in the diurnal

from the selected harmonics predicts the rainfall temperature range (DTR), the magnitude of which

fluctuations from 2000-2009. The rainfall series is comparable to the increase in mean temperature.

actual and predicted from the earliest to 2009 will Despite the importance of these trends in the con-

be presented. text of climate change, they have not yet been

extensively examined in numerical simulations of

coupled general circulation models due to technical

Statistical detection of the anthropogenic green- limitations. The Canadian Centre for Climate Mo-

house signal in observed global and regional cli- del-ling and Analysis first generation coupled gene-

mate data fields, T. Staeger, J. Grieser, C.-D. ral circulation model has recently been integrated

Schnwiese (Institut fr Meteorologie und under a variety of anthropogenic forcing scenarios.

Geophysik, J.W. Goethe Universitt Frankfurt An examination of the temporal trends in the DTR

a.M.) in these integrations, their spatial characteristics,

their similarity to observations, and their relation to

This study deals with the analysis of climate ob- trends in clouds, specific humidity, aerosols, and

servational data of the last 100 years and their radiative forcing, will be presented.

attribution to anthropogenic and natural forcings.

The data include annual, seasonal and monthly

means (totals) of surface air temperature, sea level Detection of the causes of recent trends in

pressure and precipitation. The spatial resolution radiosonde temperatures, P.W.Thorne1,

reaches from global means over area means on a P.D.Jones , S.F.B.Tett , M.R.Allen , D.E.Parker2,

1 2 3

global scale (only temperature and pressure data) to T.D.Davies1, and T.J.Osborn1 (1Climatic Research

grid points on the European scale. Unit, UEA, Norwich, UK.; 2Hadley Centre for

First an EOF analysis of the data fields is per- Climate Prediction & Research, Meteorological

formed. Thereafter the PC timeseries are used as Office, Bracknell, Berkshire, UK; 3Rutherford

target variables in a sucsessive/multiple regression Appleton Laboratory, Chilton, Didcot, UK.)

analysis, where the following forcings are used as

potential regressors: Greenhouse gases (GHG), Previous climate change detection and attribution

tropospheric sulfate aerosol, explosive volcanism, studies have considered either changes in surface

solar activity, ENSO and NAO. temperatures, or zonally averaged radiosonde tem-

For each forcing and target variable tempo-spatial peratures. We contend that in so treating the radio-

fields of signal amplitudes and signal-to-noise sonde temperatures these latter studies have ignored

ratios are obtained. The unexplained variance was valuable information, as the observed temperature

split up into structured and unstructured compo- changes are zonally heterogeneous. In considering

nents, the latter being an estimation for climate the full fields we hypothesise that the power to

noise. discriminate between competing forcing mecha-

54

nisms will be enhanced. Here initial results are pre- cings and slowly varying internal dynamics, and

sented from an optimal detection study that in- therefore are more potentially predictable than

cludes information on longitudinal as well latitu- those derived by the conventional EOF analysis, ro-

dinal variations in upper air temperatures. This tations of EOFs or low-pass filtering. The proposed

study uses the Hadley Centre's globally gridded modified EOF analysis is particularly suitable for

HadRT radiosonde dataset and compares this to two analysing regional atmospheric fields in the

versions of the Hadley Centre's GCM, HadCM2 extratropics.

and HadCM3. A number of input diagnostics are

utilised in our analysis. We comment upon which

signals are detected and implications for intra- and Extratropical cyclone variability in the Northern

inter-model consistency considerations arising from Hemisphere winter from the NCEP/NCAR Reana-

our analysis. lysis data, Olga Zolina, Sergey Gulev, Sergey

Grigoriev (P.P.Shirshov Institut of Oceanology,

RAS, Moscow)

Signal Analysis and Detection Studies Using

Neural Networks, Andreas Walter (Institute of The winter climatology of Northern Hemisphere

Meteorology and Geophysics, Johann-Wolfgang cyclone activity was derived from 6-hourly NCEP/-

Goethe University, Frankfurt, Germany) NCAR Reanalysis data for the period from 1958 to

1999, using software which provides a better

The climate system is considered to operate in a accuracy of cyclone identification in comparison to

complex, non-linear way. The higher the complexi- numerical tracking schemes. Cyclone characte-

ty of a system is, the more complicated statistical ristics over the Kuroshio and Gulfstream are very

investigations of such a system are. Because of different than those over the continental North

their capability to model multiple non-linear input- America and the Arctic. Analysis of Northern

output relations Neural Network Models (NNM) Hemisphere cyclones shows secular and decadal-

are an appropriate tool for investigations of the cli- scale changes in cyclone frequency, intensity, life

mate system. We apply the most widely used NNM time and deepening rates. The western Pacific and

architecture, the Backpropagation Network (BPN), Atlantic are characterized by the increase in

with and without an additional Momentum-Term to cyclone intensity and deepening during 42-year

analyze the dynamics of the climate system on a period, although the eastern Pacific and continental

global, hemispheric and regional scale and compare North America demonstrate opposite tendencie s in

these results to those carried out by a traditional most of cyclone characteristics. There is an increase

multiple linear regression (MLR). The NNM is able of the number of cyclones in the Arctic and in the

to explain up to 88% of variance and the residuals western Pacific and a downward tendency over the

can be treated as noise. The extracted natural Gulfstream and subpolar Pacific. Decadal scale

(Volcanism, El Ni~no, solar activity) and anthro- variability in the cyclone activity over the Atlantic

pogenic (CO2, Sulfate and combined influence) and Pacific exhibits the south-north dipole-like

signals show reasonable time structures and ampli- patterns. Atlantic and Pacific cyclone activity

tudes, e.g. BPN global: CO2: +1.11 K, SO4--: -0.36 associated with NAO and PNA is analysed.

K, combined CO2 & SO4--: +0.79 K. In a second Atlantic cyclone frequency demonstrates a high

step these modeling techniques are applied to the correlation with NAO and reflects the NAO shift in

EOF-transformed representation of global spatio- the mid 1970s, associated with considerable

temporal temperature anomalies and signal analysis changes in European storm tracks. The PNA is

are perfomed on this sacale, too. Using the signal- largely linked to the eastern Pacific cyclone

to-noise ratio a detection variable is defined and it frequencies, and controls cyclone activity over the

is shown that NNM find a significant higher Gulf region and North American coast during the

anthropogenic influence in larger areas than MLR last two decades. Assessment of th e accuracy of

does. the results and comparison with those derived using

numerical algorithms, shows that biases inherent in

numerical procedures, are not negligible.

Potentially Predictable Patterns, Xiaogu Zheng1

and Carsten S. Frederiksen2 (1 National Institute

of Water and Atmospheric Research, Wellington,

New Zealand; 2Bureau of Meteorology Research

Centre, Melbourne, Australia)

long-range potentially predictable patterns of

meteorological seasonal mean fields, that is, the

patterns arising from slowly varying external for-

cings and slowly varying internal dynamics. Using

500-hPa heights over the North Pacific/North

America region, it is demonstrated that the spatial

patterns derived by the modified EOF analysis are

more closely related to slowly varying external for-

55

being represented on a grid ponits 300 km apart.

Poster contributions to the 8 Inter- One particular interest concerns the simulation of

national Meeting on Statsitical possible changes in rainfall and surface air tem-

perature due to an asummed increase of greenhouse

Climatology gases. However , the models yield the climatic

proyections on grid points that in most cases do not

An Explanation for the Multidecadal, Winter correspond to the sities of major interest. To

Rainfall Decline Over South-Western Australia, achieve local estimates of the climatological va-

Bryson C. Bates1, Stephen P. Charles1, James P. riables, methods like the one known as statistical

Hughes2, and Edward P. Campbell3 (1CSIRO downscaling are applied. In this article we show a

Land and Water, Wembley, WA 6913, Australia; case in point related to monthly precipitation by

2

Department of Biostatistics, University of applying canonical correlation analysis (CCA) to

Washington, Seattle WA 98195, USA; 3CSIRO some regions in the Northwestern of South America

Mathematical and Information Sciences, (Colombian Region).

Wembley, WA 6913, Australia)

There has been a decline in 'winter' (MayOctober) Better Quality of StatisTical Downscaling Models

rainfall over south-western Australia (SWA) since Without Preceding PC-Analyses For Predictands?

about the middle of the 20th century. Since the mid Example: Rainfall Assessments for Namibia U.

1970s, this decline has had important consequences Beyer (Institute of Geography, University of

for water resources management and agriculture Wuerzburg, Wuerzburg)

and forestry. The links between the rainfall decline

and changes in atmospheric circulation are inves- This contribution presents results of regional rain-

tigated using the NCEP-NCAR reanalysis dataset fall assessments for Namibia under conditions of

that spans 195898, observed daily rainfall series man-made enhanced greenhouse warming, obtained

for 30 sites across the region, and a nonhomo- in the context of the DFG-post-graduate program

geneous hidden Markov model (NHMM) fitted to Joint Geoscientific Research in Africa. Relations

the period 197892 and validated for 195877. The between large-scale atmospheric circulation and

NHMM defines stochastic conditional relationships Namibian summer-rainfall are established by multi-

that link daily multi-site precipitation to a discrete ple regression analyses. For this purpose PCs of

set of weather states (patterns). The transition from geopotential heights (300,500,1000hPa) and month-

one state to another is conditioned on a relatively ly rainfall data of Namibia (84 stations) are linked

small number of atmospheric variables (predictors) by stepwise multiple regression analyses for every

derived from the reanalysis dataset. station or alternatively every PC of the station data,

on a monthly base (November March) during a

Our analysis consisted of three parts: (1) the use of 30-year calibration period. After verifying these

a nonparametric jump detection algorithm for i- statistical relations in an independent period, sta-

dentifying any spurious jumps in the predictor tions with regression models of sufficient quality

series; (2) comparisons of the predictor series and (significant correlations r>0.4 between observed

NHMM weather state sequence for 1958 to mid and modelled data) are selected and used to assess

70s with those for mid70s to 1998; and (3) a sen- local rainfall for greenhouse gas-scenarios from

sitivity analysis to investigate whether the rainfall simulated ECHAM4tr-T42 geopotential heights

decline since the mid70s can be attributed to the data. In case of preceding PCA for rainfall data no

behaviour of a single predictor. selection is possible all PCs have to be included

Our results indicate that: (1) there are no spurious to calculate future rainfall. Information about model

jumps in the reanalysis data used; (2) the atmos- quality exists, but can not be used to get higher

pheric predictor series for the mid70s to 1998 are reliability in this downscaling approach. Indepen-

quite different to those for 1958 to the mid70s; (3) dent of the procedure, results for climate conditions

there is very strong evidence of changes in the of increasing CO2-concentrations compared to a

mean probability of occurrence of two out of six reference period (1961-90) show a light increase of

weather states; and (4) the changes in weather state rainfall in parts of Namibia for December, only

frequency, and hence rainfall, are due to a com- small changes in January and significant increases

bination of changes in several atmospheric varia- in February especially in the eastern regions, but a

bles rather than any one predictor. Possible mecha- distinct decrease in March all over the country.

nisms underlying the change in atmospheric circu- These findings point to an intensified, more accen-

lation will be discussed. tuated rainy season, however the amount of rainfall

remain more or less the same under conditions of

enhanced greenhouse warming.

Statistical Downscaling to Monthly Precipitation

in the Northwestern of South America (Colom-

bian Region), Nestor R. Bernal S. (Nacional

University of Colombia)

models of Atmosphere (GCMAs) enable one to

simulate the global climate, with the field variables

56

Multiple Non-linear Statistical Model for Esti- two main groups meteorological complexes,

mating Mean Maximum Urban Heat Island in a whereas the variability of meteorological para-

Medium-Sized Town Szeged, Hungary, Zsolt meters and their mutual relations throughout the

Bottyn, Jnos Unger (Department of Climatology year are determined by two main factors: advection

and Landscape Ecology, University of Szeged, factor joining T1, T2 and e variables and baric-

Szeged, Hungary) radiational - P, S, Q. The portion of dispersion

formed by the both factors in the variability of

This study examines the spatial and quantitative meteorological elements make up on the average

influence of urban factors on the surface air tem- 70-85% of their total dispersion. In cold seasons

perature field of the medium-sized city of Szeged, (November-February) the advection factor is the

Hungary using of mobile measurements under dominating one, forming 44-54% of the total

different weather conditions between March 1999 dispersion, whereas the importance of the second

and February 2000. This city of about 178 000 is factor is 1,5-2 times smaller. In the rest of the year

situated in a low, flat flood plain along the Tisza the influence of baric-radiational factor increases

River. (forms 30-40% of dispersion in variables) and

Efforts have been concentrated on the determi- almost equals the importance of advection factor.

nation of spatial distribution of mean maximum ur- The remaining 15-25% of dispersion in meteo-

ban heat island (UHI) intensity and modelling it in rological elements are, presumably, formed by local

space. meso- and microfactors. The both meteorological

During our work we have been determined correct, complexes, as it has turned out during ortho-

existing relationship between rural parameters and gonalization of factors in FA, exist almost inde-

mean max. UHI intensity by means of multiple pendent of each other (in cold seasons in parti-

non-linear regression analisys. cular), i.e., it is possible to distinguish two most

The results indicate strong relationship exists be- important types of climate-forming processes in the

tween urban thermal excess and land-use features eastern part of the Baltic Sea region. One of them is

such as built-up ratio, gradient of built-up ratio, advection process related with the input of various

sky-view factor and distance from citys centre as air masses whose features are best reflected by air

well. On the other hand the role of water surface is temperature and humidity. The other process

negligible in this case. mostly takes place within one air mass: air mass

The model is based on three non-linear regression transformation, vertical mixing, formation of clouds

equations which are calculated in the studied peri- and related precipitation and input of solar

ods distinguishing heating and non-heating seasons radiation. Thus, FA is not confined only to

within the one-year period in the studied area. systematisation of input data. It is helpful in

Thus our model is very useful because the predicted objectivization of climate-forming peculiarities and,

seasonally spatial distribution of mean max. UHI in this particular case - in confirmation of the

intensity field differences less than 0.2 C on theory about climate-forming processes in the

average from the measured UHI field . eastern in the eastern part of the Baltic Sea region

In the interest of practical application on the basis and quantitative evaluation of the importance of the

of our previous results - we have been developed climate-forming processes themselves.

this model to predict seasonally mean max. UHI in-

tensity field considering only the geographical po-

sition within the studied area. A New Approach of Interpolation and Prediction

The reliability of this model is studying connection of Climate Time Series, Hong-Xing Cao, Xiao-

with another city which has similar invariable urban Jing Jia (Chinese Academy of meteorological

parameters, like Szeged. Sciences, Beijing ,China)

Application of factor analysis for quantification of modeling has been proposed. Firstly, a nonlinear

climate forcing processes in the Baltic Sea region, ordinary differential equation can be retrieved from

Arunas Bukantis Dept. Of Hydrology and Clima- the data series based on a criterion of bilateral

tology of Vilnius University, Vilnius Lithuania) difference forecasts, secondly taking the retrieved

differential equation as a dynamic kernal, a model,

The article represent a discussion of methods based which is called data-based mechanistic self-memo-

on factor analysis (FA) and used for identification ry model (abbreviated as DAMSM), can be estab-

of climate-forming factors and quantitative eva- lished on basis of the self-memorization principle.

luation of their importance. The work was per- The DAMSM can be applied to an temporal

formed using the data from seven meteorological interpolation of climatic observed data in default of

stations in the eastern part of the Baltic Sea region: an observation and to a climate prediction.

average monthly values of different meteorological Modeling of the inter-annual forecast of the preci-

parameters (average maximal and average minimal pitation in summer over Yangtze Delta was carried

values of air temperature T1 and T2, atmos- out, By utilizing the seasonal precipitation for 1951

pheric pressure P, wind speed in the height 10 m to 1998, the retrieved differential equation was

v, monthly precipitation Q, duration of sun- obtained, which is

shine S and partial pressure of water vapour e).

FA has revealed that the initial variables joined into

57

Dx/dt=-0.638x(t-1)+0.569x(t-3)-0.512x(t- records is commonly assessed by analysing the

4)+0.281x(t-5)+0.334x(t-7) +0.277x(t-5)2 difference series of neighbouring stations with

statistical tests such as the Buishand, Alexan-

The hindcast of the DAMSM of the summer dersson or Easterling-Peterson test. These approa-

precipitation over the Yangtze Delta was computed ches generally suffer from the lack of readily

,its RMS ( root mean square) error is 75.3 mm available nearby reference stations and hence, these

which is better than that by the autoregressive test procedures were often applied to original

(AR) model. surface ozone records without differencing with a

Besides, the models of the aviation transport and nearby reference station.

sunspot were established too ,the prediction accu- Using surface ozone data at Jungfraujoch (JFJ,

racies are all quite satisfactory. 3580 m asl) and Zugspitze (ZUG, 2960 m asl) in

the Alps, we show that high-elevation surface o-

zone records display significant serial correlations

Wind Field Retrieval From Scatterometer Data, which may seriously affect the absolute and relative

Dan Cornford, David J. Evans and Ian T. Nabney homogeneity testing. Hence, we review some

existing parametric and nonparametric

It is well known that radar backscatter can be used homogeneity tests and present their performance on

to estimate the near surface wind vector over the simulated surface ozone records, with inclusion of

ocean. In recent years several satellites have been multiple outliers and level shifts.

deployed carrying scatterometers to provide data It is evidenced that none of these tests perform

for retrieval of wind fields and subsequent satisfactory, neither for absolute nor for relative

assimilation into numerical weather prediction homogeneity testing.

models. This paper addresses the question of Structural models in state space form provide a

whether there is sufficient information in the suitable and sufficiently flexible statistical frame-

scatterometer observations alone to uniquely deter- work for the unified analysis and homogeneity tes-

mine the "true" surface wind vector field. This is ting of time series records. Using surface ozone

difficult because the retrieved local wind vectors data at JFJ and ZUG, we show how to estimate, test

have several ambiguous solutions for a given and validate structural models in state space form

scatterometer observation. We use a Bayesian including multiple structural breaks. The per-

approach to combine the local scatterometer formance of this modelling approach is assessed on

observations with a global wind field prior. The simulated surface ozone series with inclusion of

prior model is a Gaussian process with parameters multiple outliers and level shifts.

estimated from several years of numerical weather

prediction model data. We show how direct local

inverse models, based on mixture density networks, Statistical Analysis of Modern Climate Changes

can be used efficiently in the retrieval procedure. observed in the West and East Antarctic,

The posterior distribution of the wind field has high Jagovkina, S. V.1, Lagun, V. E.2 (1Main Geophy-

dimension, with several distinct modes. We show sical Observatory, St.-Petersburg, Russia; 2Arctic

how to sample from these distributions and illus- and Antarctic Research Institute, St.-Petersburg,

trate the performance of several sampling algo- Russia)

rithms on toy problems. Many of the existing algo-

rithms are shown to perform badly on the toy The statistical analysis of temperature, pressure

problems, because the modes of the posterior are fields, precipitation and wind speed is executed for

well separated in the high dimensional space. We Russian Antarctic stations located in the East

show that in most cases the posterior distribution of (Mirny) and in the West (Bellingshausen) Hemis-

the wind field has two dominant modes one of pheres. This analysis was made on the base of

which is typically close to the numerical weather multiyear current registered meteorological data for

prediction model wind field (which is not used at all period of instrumental observations (1956-2000

any point in the retrieval procedure). We also for Mirny and 1968-2000 for Bellingshausen).

discuss the effect of increasing the number of local Mirny Observatory is situated on the coast of Cape

observations considered in the retrieval, which may Davis at a small protrusion of Mirny Peninsula.

improve the chance of retrieving the "true" wind Mirny Observatory is the main base of Russian

field. studies in the Antarctic. The Bellingshausen station

is situated on the Fildes Peninsula in the south -

western part of King-George Island, being a part of

Homogenization and Quality Control in State the South Shetland Isles.

Space Models, Thomas K. Friedli, Evi The tendencies of interannual climatic changes are

Schuepbach, Tania Steiner, Jrg R. Hsler identified with account of annual, seasonal, month-

(Department of Mathematical Statistics and ly and synoptic variations of considered parameters.

Actuarial Sciences, University of Berne, Bern, The traditional methods of regressive and corre-

Switzerland) lative analyses as well as methods of stochastic

analysis of the ensemble of several periodically

Long-term surface ozone records may provide correlated random processes are used for inter-

essential information on changes in lower tropos- annual oscillation description and low-frequency

pheric ozone over time. The homogeneity of such trend determination. Obtained results are compared

58

with NCAR/NCEP Reanalysis (1958-1998) data set month, for preceding month, for more preceding

and hemispheric temperature anomaly time series. month, and so on, to month about 11 months pre-

Statistical features of climatic parameters variation ceding the concrete month) for every grid point.

in Western and Eastern parts of Southern Cluster analysis was computed for every our

Hemisphere are demonstrated. Possible causes of monthly temperature set separately and results

determined climatic tendencies are discussed in were illustrated in maps.

connection with available data about cloudness For some one of these month was number of

regime, eddies synoptic activity and sea ice dyna- clusters clear, for some one was determination of

mics. the number more problematic. Just analysis of the

polemic cases is main aim of the report.

American Stationary Systems (SASSY), Charles A new look at the temporal structure of a time

Jones1 and Leila M. V. de Carvalhol2 (1 Institute series, Ken-Chung Ko (Department of Geography,

for Computational Earth System Sciences National Kaohsiung Normal University,

(ICESS), University of California, Santa Barbara, Kaohsiung, Taiwan)

CA; 2 Dept. of Atmospheric Sciences, University

of So Paulo, Brazil) This study introduces a method in analyzing the

temporal structure of quasi-periodic oscillations.

The South Atlantic Convergence Zone (SACZ) is The idea of the moving spectra will be adopted to

subjectively described as a convective region exten- study the temporal variation of a time series. The

ding from the Amazon basin, passing over objective of this study is to look at different aspects

southeast Brazil and running towards the South of the spectral methods in order to detect the

Atlantic. The SACZ is characterized by a wide periodicity that could be hidden in regular spectra.

range of temporal variability from synoptic to intra- Results show dramatic changes after applying

seasonal and interannual scales. Our recent study normalization procedures and statistical tests to the

has objectively identified the occurrence of South moving spectra. Four idealized cases tested by the

American Stationary Systems (SASSY) during moving spectral method reveal that the moving

austral summer. SASSY events are characterized by spectral method can transform the time series into

persistence greater than two days and large geo- the temporal spectral structure.

graphic extension over southeast Brazil. This pre-

sentation will discuss a statistical method to fore-

cast the large-scale feature of convective activity in Extreme value analysis applied on precipitation

the SASSY region. Daily Outgoing Longwave time series in the Czech Republic, Kveton, V.

Radiation (OLR) from the period January 1979 (Czech Hydrometeorological Institute, Prague,

through December 1999 is used to develop and Czech Republic)

validate the statistical model. Examples of real-time

forecasts will also be discussed. Yearly extremes of 1- to 7-days precipitation totals

both from 8 long time series and 50 shorter time

series (1961-1998) in the Czech Republic were ana-

Using of cluster analysis in processing of corre- lysed. Different observation period and different

lation coefficients, Stanislava Kliegrov, Ladislav estimation method were used. Results are com-

Metelka (Czech Hydrometeorological Institut, pared. Obtained experiences are discussed.

Hradec Krlov Svobodn Dvory, Czech

Republic)

Statistical climatology and interannual variability

The satisfactory predictors are looked for the long- of the extratropical cyclone parameters in the

term (seasonal) statistical forecasts for the Czech global atmosphere, Victor E. Lagun (Air-Sea

Republic territory. The relationships among sets of Interaction Department, Arctic and Antarctic

temperature (monthly value) in the Czech Republic Research Institute, St. Petersburg, Russia)

and sets of SST (Sea Surface Temperature, in grid

points 2x2) in part of the Atlantic were described The calculation of the global distribution and of the

by the help of correlation coefficients. For simpli- time variability of cyclone synoptic eddies para-

fication the cluster analysis was used in next pro- meters, including the space and time frequencies

cessing, when the regions of the ocean with similar estimations, the depth, the linear size, the area, the

behaviour of significant correlation coefficients velocity, the trajectory variability is made on the

were found. Just possibility of using of K-means base of twice - daily mean sea level pressure and

Clustering Method, that moves objects among geopotential fields extracted from NCEP data sets

k clusters with the goal (1) minimize variability for 1980-1998 period. The original methods for

within clusters, and (2) maximize variability among statistical analysis of meteorological fields were

clusters, was explored in the paper. The objects, in used to define the position of cyclones and anti-

our case, were sets of 12 correlation cofficients cyclones on the sphere. These numerical methods

among set of monthly temperature in the Czech are based on traditional definition of synoptic

Republic (for one concrete month) and sets of eddies boundaries as external closed isobars. The

monthly values of SST (stepwise for the same estimations of mean monthly, seasonal, annual and

59

decadal synoptic climatology parameters for both anomaly of SST on a large part of eastern and cen-

hemispheres were carried out. tral Tropical Pacific.) Analyses were carried out

The analysis of different algorithms used for para- using the integral wavelet transform. This paper

meters estimations in synoptic statistical climato- documents 1) The existence of a large amplitude

logy is executed. The obtained results are compared modulation of the Southern Oscillation (AMSO). 2)

with the known estimations for hemispheric para- A change in the relative phase between SOI and

meters distributions of synoptic climatology and the SSTI fluctuations at the decadal and bidecadal

differences in results due to including of small and timescale, which occurred in mid-1930s.

mesoscale disturbances parameters in the statistics Signals in the band of wavelet-period smaller than

of extratropical cyclones are discussed. The tenden- 10 years do not show indication of climate change

cy of the extratropical cyclones time variability throughout the period of study (almost a century-

under global warming conditions is estimated. The long data record). The controlling process on the

results demonstrate that the number of mature variability of decadal component of SOI and SSTI

extratropical cyclones is increasing during the during that period is AMSO (amplitude modulation

current period of global warming, and it is the of the Southern Oscillation). AMSO produces

important indicator of modern climate change. The maximum amplitude of decadal components of

possible reasons of synoptic eddy activity increa- SOI and SSTI in 1913-1917. Thereafter, these com-

sing are discussed. ponents have decreasing amplitude until about

1957. In early-1960s, decadal component of SOI

and SSTI signals starts to amplify. AMSO has a

Spatial Organization of Decadal and Bidecadal major influence on the observed values of SOI and

Rainfall Fluctuation on Southern North America SSTI. The state of AMSO, and contributions from

and Southern South America, Omar Abel Lucero, decadal, bidecadal, and tridecadal component of

Norma C. Rodriguez (Instituto Nacional del Agua SOI and SSTI, account for major changes observed

and National University of Cordoba, Cordoba, in this century in the SOI, and in the SSTI.

Argentina)

The spatial organization of decadal and bidecadal Variability of annual pressure and precipitation

components (fluctuations) of annual rainfall is iden- on Iberian Peninsula, M Dolores Manso Orgaz*

tified in this research for two regions: 1) southern and Liliana Caramelo** (*Depart. of Physics.

South America, and 2) southern North America University of Aveiro, Portugal; ** Depart. of

(conterminous USA, southeastern Canada and Physics University of Trs-os-Montes and Alto

northern and central Mexico). Findings indicate that Douro . Vila Real)

these decadal and bidecadal components have high-

ly-coherent wave-like spatial organization. Two In this study we investigate the space and temporal

types of organization of decadal and bidecadal variation of the annual precipitation and the annual

components of annual rainfall were identified: a pressure (at the sea-level), on the Iberian peninsula

train of propagating fluctuations, and quasi-stan- (IP) in 40 stations of 39 years. Using the technique

ding fluctuations. For decadal components, such of empirical orthogonal functions (EOFs) to obtain

patterns alternate in time. A widespread change in a regional precipitation and regional pressure. In

the spatial organization of decadal component of order to obtain the structure of temporal variations

annual rainfall took place simultaneously in both we are using to spectral analysis of the principal

continents in 1932. The bidecadal component is components (PCs). We analysed the relationship

organized as standing fluctuations in southern between the observed time series and the following

North America; and as travelling fluctuations in teleconnection indices: Oscillation of the North

southern South America The spatial pattern of de- Atlantic (Gibraltar extension NAO-G and Ponta

cadal fluctuations of annual rainfall has 12- and 13- Delgada extension NAO-A) .We found two patterns

year cycles, and the spatial pattern of bidecadal for the annual pressures accounts for 86% of total

fluctuations has predominantly 21- and 22-year cy- variance; and four for the annual total precipitation

cles. account for 74% of the total variance. These

leading PCs are associated with the following

patterns: NAO-G, NAO-A, respectively. The iso-

Interrelationship between Interdecadal Fluctua- baric gradient and the spectra show coherence of

tions in the Southern Oscillation and the Sea Sur- the first EOF of the annual pressure and of the

face Temperature of the Tropical Pacific, Omar annual total precipitation with the trimester of

Abel Lucero, Norma C. Rodriguez (Centro de la winter NAO-G. For the reconstruction of the series

Region Semirida (INA), and National University is provided a statistical model in function of the

of Cordoba.Cordoba, Argentina) more significant oscillations.

cadal timescales on the Southern Oscillation and in

the sea surface temperature of the tropical Pacific

are the focus of this research. Analyses are based

on the SOI (Southern Oscillation Index), and on

Wright's SSTI (The SSTI is defined by the average

60

Reproduction of dry and wet Periods by Stochastic atmospheric trace gases involved in forming ozone

Daily Precipitation Generators, D. Marvan1, M. and nitrogen oxides change rapidly with several

Dubrovsky2 (1Faculty of Math. and Physics, meteorological variables such as wind speed and

Charles University, Prague, Czech Republic; direction, temperature, humidity and solar

2

Institute of Atmospheric Physics, Hradec radiation. Thus, good statistical models which can

Kralove, Czech Republic) handle with the intrinsic variability of

climatological and/or meteorological data are

Stochastic models used to generate the series of wet clearly needed.

and dry days in daily weather generators are mostly In this paper we set up the foundamentals to

based on Markov chains (MCs). Since the first- propose, build and further develop statistical mo-

order MCs cannot satisfactorily reproduce the dels based on semi-parametric approaches. Particu-

dis-tribution of the lengths of wet and dry spells (in larly, a mixed formulation involving a non-parame-

addition, other weather characteristics, such as tric component based on locally weighted regres-

occurrence of hot and cold periods, derived from sion procedures which accounts for a possible

variables generated conditionally on the preci- deterministic trend, and a parametric component re-

pitation occurrence may be distorted too), the alter- pre-senting the purely temporal random variation is

native approaches are employed. These include proposed.

higher-order MCs, hybrid-order Markov chains and The data set analyzed comes from an area with ty-

serial approach models. The present contribution pical mediterranean environmental conditions, so

aims to assess capability of several models (higher that our conclusions can be easily generalized to

order Markov chains, hybrid-order Markov chains general mediterranean polluted environments.

and serial approach model) to reproduce selected

statistical characteristics precipitation occurrence

series. The stress will be put on reproducing the Statistical Methods in climatology: an application

long wet and dry spells. The fit of the synthetic vs. to the longest precipitation series in Portugal,

observed weather series is evaluated in terms of (i) Solange Mendona Leite (Geophysical Centre of

summary measures (Wilcoxon, chi-square and the University of Lisbon, Portugal)

Kolmogorov-Smirnov tests), (ii) maximum length

of wet/dry spells, and (iii) frequency of occurrence Instrumental climatic data are limited to the period

of extremely long dry/wet spells. from mid-twentieth century for the most regions in

In the second part of the contribution, the accuracy Iberian Peninsula, with a few data sets going back

of the generators parameters estimated from the to late 19th century or early 20th century. Portugal,

learning (observed) data will be assessed. This however, has one high-quality instrumental weather

exercise is motivated by the problem, whether it is observation series going back to the middle 19th

better to use simpler model (e.g., first order Markov century.

chain) with less parameters (and greater accuracy) This weather station was selected by an exhaustive

being derived from the learning data, or to use more search of documentation regarding instrumentation

sophisticated model with more parameters, which and observational practices. Statistical tests of the

are, however, derived with lower accuracy from the reliability of observing practices also have been

learning sample. The combined effect of the size of applied in the selection process, as well as tests to

the learning sample, type of the model and detail- detect inhomogeneities. The data set used is the

ness of the representation of the annual cycle of the most reliable available for monitoring precipitation

generator's parameters (one value per year/sea- trends during the last century and a half. With the

sons/month) on the generators' performance will be availability of this long time series it is now possi-

therefore assessed. ble to conduct much more detailed analyses. In this

The study is supported by Grant Agency of the report we summarise these efforts using the weather

Czech Rep., contract 205/97/P159 and by Grant A- station of Instituto Geofsico Infante D. Lus, in

gency of Academy of Sciences of CR, contract Lisbon. A trend analysis over the period 1856-1999

A3042903. indicates a temporal pattern of precipitation variabi-

lity. The statistical characteristics of precipitation

monthly totals are analysed. Maximum entropy

Semi-parametric statistical approaches for the spectral analysis is applied to quantify the relative

analysis of climatological data, J. Mateu1, C. importance of decadal variability. An autorregres-

Alvarez2, T. Sanfeliu2, M.M. Jordan3 (1 sive model of order 45 is fitted to the 144-year

Department of Mathematics; 2 Department of precipitation series and the agreement between

Experimental Sciences, Universitat Jaume I, E- observed and estimated values is judged. The value

12071 Castellon, Spain; 3 Department of Experi- of this model simulation as a source of regional

mental Sciences, Universitat Miguel Hernandez, climate scenarios should be considered in the con-

Elche, Spain) text of other uncertainties in the global climate

change scenarios which are used to drive them.

An understanding of the ozone behaviour near These include the inherent variability of the climate

surface layers is essential for a study of pollution system and the differences in sensitivity and in

surfaces. Controlling possible sources of air pollu- patterns of response between climatic elements.

tion has become a relevant goal in our increasingly

degraded environment. Moreover, concentrations of

61

Spatial patterns of the five-day rainfall all over Data sets consist of observed monthly temperature

North-Western Iberian peninsula, M. Andrs and precipitation amounts from 62 Romanian

(Atmospheric Physics Department, University of stations and indices of NAO, EA, EA/WR, SCA,

Salamanca, Spain); Clemente Toms POL taken from NCEP data. The analyzed time

(Atmospheric Physics Department, University of interval is 1961-1996. Statistical significant rela-

Salamanca, Spain); Fernando de Pablo (Atmos- tionships between large and local scales are

pheric Physics Department, University of Sala- identified mapping the correlation coefficients be-

manca, Spain); Solange Mendona Leite tween the large scale indices and local thermal and

(Geophysical Centre of the University of Lisbon, precipitation anomalies. Also, composite maps of

Portugal) precipitation and temperature anomalies over Ro-

mania territory have been derived depending on the

Precipitation is the climatic element which presents opposite phases of the large-scale phenomena. The

a higher spatial and temporal variability in the large/local scale relationships with predictive po-

Iberian Peninsula. While the analyses of the tem- tential can be further used to develop a regionally-

poral variability of the precipitation data sets is orientated methodology for extreme event risks

quite easy, the spatial variability is much more assessment .

difficult to identify and to interpret, and even more

when orography makes rainfall distributions even

more complex. In this report we present a regiona- Neural networks to predict atmospheric profiles,

lization of Castilla y Len, a region located in Nazario D. Ramirez-Beltran (Department of

north-western Spain, using 5-day rainfall data sets. Industrial Engineering, University of Puerto Rico,

Daily values were available for 119 weather Mayaguez); Jorge Gonzalez (Department of Me-

stations, during a 6-year period, from 1990 to 1995, chanical Engineering, University of Puerto Rico,

considering the rainfall period between September Mayaguez); Felipe Chong Rengifo (Department of

and May. The region presents a characteristic Industrial Engineering, University of Puerto Rico,

orography which affects the final results. Principal Mayaguez)

Component Analysis (VARIMAX rotation) has

been applied as an adequate method of regio- A methodology to predict climatic variables based

nalization. Six components were identified as statis- on radiosonde and satellite observations is de-

tically significant, explaining 82.6% of the total veloped. The proposed methodology will contribute

variance of the field. In order to identify which to mitigate the existing of very limited radiosonde

synoptic systems are related to each component, data for the Caribbean. Satellite and radiosonde

those cases in which the score was positive only in data from 1986 to 1999 were used to create the cli-

one of the components (while the remaining were mate information required for a competitive neural

negative) were selected. Thereafter, the synoptic network to identify the climatic observations

situations associated with that 5-day group were associated to a typical day. A feedforward neural

analysed, looking for the main one and obtaining network was also designed to model the rela-

common patterns for each one of the components. tionships among the inputs and outputs of an

We have also tried to identify the associated atmospheric dynamic system. Meteorological varia-

systems which affect the region when there were bles from two consecutive mandatory pressure

several components with positive values for a 5-day levels were used to train a feedforward neural net-

group. work for predicting the following variables

geopotential height, air temperature, relative

humidity and wind components in the Caribbean

The effects of large scale phenomena on local cli- basin. The designed scheme has the capability of

mate variability over Romanian territory, Daniela predicting five climatic variables from the sea

Mihaela Paliu, Roxana Bojariu, (National In- surface up to 17 km of elevation. The developed

stitute of Meteorology, Hydrology and Water Ma- neural network methodology was successfully

nagement, Bucharest, Romania) tested by using satellite and radiosonde

observations for the year 2000. Climatic

This paper aims to identify the effects of large scale observations and neural network predictions were

variability modes such as North Atlantic Oscillation very close indicating that the proposed

(NAO), East Atlantic Pattern (EA), East Atlantic/- methodology is a reliable tool for predicting atmos-

West Russia Pattern (EA/WR), Scandinavian pheric profiles.

Pattern (SCA) and Polar/Eurasia Pattern (POL) on

the occurrence of extreme temperature and precipi-

tation anomalies over Romanian territory. The large Mid-term prediction of the daily maximum/mini-

scale phenomena are generally responsible for ex- mum temperatures in South Korea using Dynamic

tended range predictability. However, the complex Linear Models, Keon Tae Sohn (Pusan National

orographical characteristics of Romanian area due University, Korea); Jeong Hyung Lee (Dong-A

to the presence of Carpathian mountains which are University, Korea); Byungsoo Kim (Inje

standing against certain types of large scale University, Korea)

circulation and modifies the relation between large

scale phenomena and local variability and predic- The mid-term(from +1 day to +10 day) prediction

tability. of the daily maximum/minimum temperatures in

62

South Korea was performed using the Dynmic the 2001 course. The main course software has also

Linear Models(DLM). Observations, which are 6- been upgraded.

hour-interval data (including the daily maxi-

mum/minimum temperatures) from September 2 in

1999 to May 31 in 2000, and the simulated outputs Changes of large-scale atmosphere circulation

of GDAPS(Global Data Assimilation and Predic- during XX century and their influence on the

tion System) used at Korea Meteorological Ad- climate in Ukraine, V.F. Martazinova, T. Sverdlik

ministration, which are upto 240 hour-prediction, (Ukrainian Hydrometeorological Institute, Kiev,

are used as training data of DLM. The DLM with a Ukraine)

prior distribution consists of two equations and

intial distributions; (1) state equation: The changes of the large-scale atmosphere circula-

Y(t)=F(t)m(t)+v(t), (2) m(t)=G(t)M(t-1)+w(t), (3) tion over Atlantic-Europe region and formed by it

m(0) is a normal-variate with mean a(0) and regional circulation and climate in Ukraine during

variance C(0) where two independent random XX century are considered.

variables v(t) and w(t) have normal distributions The maps of mean monthly sea level pressure

with mean 0 and variances V(t) and W(t), respec- (SLP) for the 1881-1940, 1951-80, 1980-95 are

tively. In addition, V(t) has an inverse Gamma dis- analyzed. The patterns of mean SLP differences be-

tribution, with two parameters n(0) and s(0), as a tween 1881-1940 and 1951-80; 1980-95 and 1951-

prior. Y(t) is an observation at time t, F(t) consists 80 allowed to find identical tendency of the

of input vaiables and m(t) means a dynamic coef- atmosphere circulation transformation during peri-

ficient vector. For estimating W(t) and m(t) dyna- ods of global warming (at the beginning and at the

mically, the discount factor and the well-known end of the century) relatively stable period of global

Kalman filtering are used. temperature course (at the middle of the century).

For the selection of optimal DLM, some procedures The absolute values of the tendency at the end of

are considered as follows; (1) to select initial va- the century in twice as much.

lues, a(0), C(0), n(0) and s(0), randomly. (2) to de- The changes of the large-scale and regional atmos-

termine the transition matrix G(t), (3) to choose the phere circulation during recent decades are defined

components of F(t), (4) to find an optimal DLM by shifting to the east of climate atmosphere centers

under the selected F(t), with changing the value of of action (Azores, Siberian highs and European

discount factor(from 0.01 to 1.00), (5) to determine depression), that forms new weather conditions in

the optimal DLM among the all F(t)'s, (6) to Ukraine. As a result the present weather conditions

perform the sensitivity analysis for the effect of on the territory of Ukraine are formed by eastern

initial values, and (7) interval estimation for fore- periphery of Azores high in all seasons and charac-

casts. RMSE is used as a measure for the optimi- terized by warm and wet winters and cool summers.

zation. These conclusions conform to the results of air

As results, we can find that (1) the proposed DLM's temperature and precipitation time series analysis.

are robust to the initial values from the sensitivity Hence this study has showed that of global and

analysis, (2) the transition matrix G(t) seems to be regional climate changes are connected with chan-

an Identity matrix, (3) for the k-lead forecast, the ges of the large-scale atmosphere circulation.

optimal F(t+k)'s consist of Y(t), the estimate of

Y(t+k) from DLM and the GDAPS(t+k) forecast,

(4) RMSE's are 2.0 - 2.7 (for +1, +2, +3, +4), 2.6 - Validating a Downscaling Model of the Stochastic

3.0(for +5, +6, +7, +8), 3.0 - 3.8(for +9, +10), and Structure of Daily Temperature, Marina

(5) the forecasts using the proposed DLM are re- Timofeyeva1, Rachael Craig2 (1UCAR, NOAA/-

markably better than GDAPS forecasts. NWS, Silver Spring, MD; 2Dept. of Geology, Kent

State University, Kent, OH)

Statistics in Agricultural Climatology adapting a We model daily maximum temperature with seven

workshop for African needs, S. M. Gathara, C. parameters: annual mean, phase shift and amplitude

Oludhe, S. B. Otengi, P. K. Runanu, R. D. Stern from solar insolation, weighted parameters (phis) of

an AR(3) model of the residuals from the insolation

A 10-week workshop on statistical methods in model and residual variance (sigma). The model

agroclimatology (SIAC) was given in Reading UK, was calibrated with a stratified sample of instru-

from 1983 to 2000. In May 2000 SIAC was given mental records from the western U.S. and Mexico

for the first time at the Institute for Meteorological for 1979-1988. Three stations (high, low and

Research and Training (IMTR), in Nairobi, Kenya. middle elevations) from each climate division in the

This paper describes the lessons learned from this U.S. and from each one-degree grid cell in Mexico

venture and the changes that are being made to were selected for calibration; for a total of 434.

tailor the 2001 course more to regional needs. Boundary conditions are defined by CO2 concen-

Many of these changes have followed from a train- tration, sea surface temperature, 500 and 1000 mb

the-trainer workshop run in October 2000. In addi- winds and geopotential height from the CCM3 T31

tion to strengthening the resource team, this has led paleoclimate control run. Thirty-two predictor va-

to a range of practical projects, described in the riables represent regional radiative and orographic

paper, on both data management and applications controls on temperature for up to 1000km upwind

that are being conducted in the 6 months leading to of the calibration locations as well as local and

63

downwind (blocking) effects. The transfer function extreme cold events in the second part of the

defining the downscaling relation was estimated by sampling time.

canonical regression.

The transfer function was validated with an inde-

pendent sample of 1290 instrumental records from Coupled sea ice and atmosphere variability in the

the same area and time period. Six separate analy- Weddell Sea, Antarctica, Silvia A. Venegas

ses comprise the validation. We first tested skew- (Danish Center for Earth System Science

ness and kurtosis of the distribution of residuals (DCESS), Niels Bohr Institute for Astronomy,

(observed minus modeled). All variables but phase Physics and Geophysics, University of

and phi(3) passed both tests. Those two variables Copenhagen, Denmark); Mark R. Drinkwater

showed very slight positive skewness. Secondly we (Oceans/Sea-Ice Unit, ESTEC Earth Sciences

computed paired t-statistics on randomly selected Division (APP-FSO), Noordwijk, The

subsets of size 208 (giving precision in means of Netherlands)

one degree C). We did not reject the null hypothesis

that model and observed means are equal, except A frequency-domain singular value decomposition

for one (of three) tests of phase. The third test is performed on 20 years (1979-1998) of satellite

considered the significance of correlation of observations of sea ice concentration and drift

residuals with each of the independent variables. together with sea level pressure reanalysis data in

One of the seven variables (sigma) passed fewer the Weddell Sea, Antarctica, in order to determine

(89%) than 95% of the 384 tests performed. To the dominant timescales of variability in the region.

detect possible spatial bias we examined (test four) The interannual timescale (periods of 3-4 years) is

the correlation of residuals with each of latitude, found to dominate the ice and atmospheric fluc-

longitude and elevation. All passed. The fifth test tuations. This variability is a regional manifestation

considered a Q-mode correlation analysis of the of the Antarctic Circumpolar Wave in the Weddell

ensemble of independent and dependent variables sector of the Southern Ocean.

at all the stations. Correlation at 92% of the stations Interannual variations of sea ice concentration and

exceeds 0.7. We see no systematic pattern in the drift are driven largely by anomalous atmospheric

spatial distribution of stations with lower cor- patterns, which alternatively enhance or reduce the

re-lation. Finally (test six) we examined the spatial meridional wind circulation to the east of the

distribution of residuals and found no suggestion of Antarctic peninsula. Thermodynamic processes

systematic patterns related to geographic controls. (rapid ice growth in open water areas), dynamic

processes (ice ridging/rafting due to ice

convergence) and geographic constraints all play

Statistical Decomposition of European Tempera- significant roles in the formation of summer ice

ture Time Series, Silke Trmel, Jrgen Grieser, concentration anomalies in the southern Weddell

C.-D. Schnwiese (Institut fr Meteorologie und Sea every 3-4 years. These ice anomalies are

Geophysik, Universitt Frankfurt, Frankfurt /M., exported northwards out of the Weddell basin

Germany) approximately one year later, and then advected

eastwards before diffusing in the ice margin.

One strategy to analyze climate variability in A low-frequency signal (presumably a quasi-de-

meteorological data, used in classical time series cadal oscillation) is also detected, although its sig-

decomposition, is the search for prede ned nificance is fairly speculative in a 20-yr record. Ice

components. We decompose monthly temperature variability on this timescale results from a change

time series into linear as well as nonlinear trends, in the shape of the Weddell gyre circulation which

the season cycle and its variations, harmonic com- occurred around 1990, likely associated with large-

ponents, episodic components, extreme events and scale variations of the ocean-atmosphere system.

noise. In contrast to extreme values our extreme

events are not caused by trends or other well de ned

changes in the mean, but they are unlikely to occur

by chance and are independent from changes in the

parameters of the distribution. It is very unlikely

that they are in accordance with the feature of the

time series. For the decomposition a stepwise

regression is used which converges fast to robust

estimaters. Our results show that the trend over

Europe can be regarded as positive and linear, that

positive phase shifts in the season cycle occur

especially in the western and negative ones in the

eastern Europe. In the north and northwest of

Europe positive changes in the amplitude of the

season cycle dominate, whereas negative ones are

found in the eastern part and in central Europe. The

overwhelming majoritiy of extreme events are

detected in winter and represent extreme cold

events. Besides, we found a signi cant increase of

64

Participants List 8. IMSC e-mail: bauer@pik-potsdam.de

27 February 2001

Thomas BENGTSSON

Ulrich Achatz Geophysical Statistics Project

Leibniz-Institut fuer Atmosphrenphysik National Center for Atmospheric Research

an der Universitaet Rostock e.V. Climate and Global Dynamics Division

Schlossstr. 6 P.O. Box 3000, Boulder, CO 80307-3000

18225 Kuehlungsborn, Germany USA

e-mail: achatz@iap-kborn.de e-mail: tocke@ucar.edu

Dept. of Earth Sciences Ohio State University, Dept. of Statistics

Physical Geography 404 Cockins Hall, 1958 Neil Avenue

Gteborg University Columbus, OH 43210, USA

Box 460 e-mail: mb@stat.ohio-state.edu

40 30 Gteborg

Sweden Nestor R. BERNAL S.

National University of Colombia

Ahmed Adjez e-mail: nbernal@ideam.gov.co

National Climate Center

P. O. Box 45, Hacen Badi Ulrike BEYER

16200 El-Harrach Institute for Geography

Algier, Algeria University of Wrzburg

e-mail: ahmed_adjez@hotmail.com Am Hubland

97074 Wuerzburg, Germany

Nasser ALBALOUSHI e-mail: ulrike.beyer@mail.uni-wuerzburg.de

United nations Development programme

UAE/99/001 Project Peter BISSOLLI

FE24c1

Abdullah Aldhanhani Postfach 10 04 65

United nations Development programme 63004 Offenbach, Germany

UAE/99/001 Project e-mail: peter.bissolli@dwd.de

AARI Arctic and Antarctic Research Institute e-mail: boehm@ziplink.net

Ocean-Atmosphere Interaction Department

199397,Bering st.,38 Dr.Reinhard BOEHM

St.-Petersburg, Russia Zentralanstalt fuer Meteorologie und Geodynamik

e-mail:alexgv@aari.nw.ru Hohe Warte 38

A-1190 Wien, Austria

Hans ALEXANDERSSON e-mail: reinhard.boehm@zamg.ac.at

Swedish Meteorological and Hydrological Institute,

Norrkping, Sweden Dr. Uwe BOEHM

e-mail: hans.alexandersson@smhi.se Universitaet Potsdam

Unstitut fuer Physik

Katrine Krogh ANDERSEN Postfach 60 15 53

Department of Geophysics 14415 Potsdam, Germany

Juliane Maries Vej 30 e-mail: boehm@pik-potsdam.de

2100 Copenhagen O, Denmark

e-mail: kka@gfy.ku.dk Konrad BOGNER

Institute of Hydrology

Sylvia ANTUNES University of Agricultural Research

Instituto de Meteorologia Muthgasse 18

R. C ao Aeroporto 1190 Vienna, Austria

Lisboa, Portugal e-mail: bogner@edv2.boku.ac.at

e-mail: silvia.antunes@meteo.pt

Auguste BOISSONNADE

Judit BARTHOLY Risk Management Solutions Inc.

Dept. of Meteorology Eotvos Lorand University 149 Commonwealth Drive

P.O.Box 32 Menlo Park, CA 94025

Budapest, H-1518, Hungary California, USA

e-mail: bari@ludens.elte.hu e-mail: augusteb@riskinc.com

CSIRO Land and Water Institutul National de Meteorologie si Hidrologie

Private Bag Sos. Bucuresti-Ploiesti 97

PO Wembley WA 6014 Australia 71 552 Bucharest, Romania

e-mail: bryson.bates@per.clw.csiro.au bojariu@meteo.inmh.ro

Potsdam Institute for Climate Impact Research (PIK) Dept. of Earth Sciences

PO Box 601203 Physical Geography

14412 Potsdam, Germany Gteborg University

65

Box 460 Forschungsverbund Berlin

40 30 Gteborg Hausvogteiplatz 5-7

Sweden 10117 Berlin, Germany

e-mail: pcarl@spclim5.wias-berlin.de

Zsolt BOTTYN

Department of Climatology and Landscape Ecology Leila M. Vespoli de CARVALHO

University of Szeged Institute for Computational Earth System Science

University of Szeged (ICESS)

P.O. Box 653 University of California

6701 Szeged, Hungary Santa Barbara, CA 93106

e-mail: zbottyan@solyom.szrfk.hu e-mail: leila@icess.ucsb.edu

Risk Management Solutions Inc. Institut de Statistique

Menlo Park Universit Catholique de Louvain

California, USA Voie du Roman Pays, 20

B-1348, Louvain-la-Neuve

Harold E. BROOKS Belgium

NOAA/National Severe Storms Laboratory e-mail: cebrian@stat.ucl.ac.be

Halley, Circle

Norman, OK, USA Peter CHALLENOR

e-mail: brooks@nssl.noaa.gov Southampton Oceanography Centre

Southampton, SO14 3ZH, UK

Adri BUISHAND e-mail: Peter.G.Challenor@soc.soton.ac.uk

Royal Netherlands Meteorological Institute (KNMI)

De Bilt, The Netherlands Lynda CHAMBERS

Bureau of Meteorology Research Centre

Arunas BUKANTIS Melbourne, Australia

Dept. of Hydrology and Climatology of Vilnius

University Richard CHANDLER

M.K.Ciurlionio St. 21/27 Dept. of Statistical Science

Vilnius 2009, Lithuania University College

e-mail: arunas.bukantis@gf.vu.lt Gower Street

London WC 1E 6BT, UK

Aristita BUSUIOC e-mail: richard@stats.ucl.ac.uk

National Institute of Meteorology and Hydrology

Sos. Bucuresti-Ploiesti 97 Stephen CHARLES

71552 Bucharest, Romania CSIRO Land and Water

e-mail: busuioc@meteo.inmh.ro Private Bag #5 Wembley WA 6913 Australia

e-mail: stephen.charles@per.clw.csiro.au

Ulrich CALLIES

Institute for Hydrophysics Rosane R. CHAVES

GKSS Research Center INPE CPTEC - Rod. Pres. Dutra, Km 40

21494 Geesthacht, Germany C. Paulista - SP - Brasil - CEP:12630-000, Brazil

e-mail: callies@gkss.de e-mail: rosane@cptec.inpe.br

Dept. of Atmospheric Sciences Institute for Numerical Mathematics RAS

University of Buenos Aires 8, Gubkin St.

Argentinia Moscow, 117333, Russia

e-mail: inesc@at.fcen.uba.ar e-mail: chavro@inm.ras.ru

CSIRO Mathematical and Information Sciences Agrometeorology Institute

Private Bag No. 5 Chinese Academy of Agricultural Sciences

Wembley, WA 6913, AUSTRALIA 30 Baishiqiao Road

e-mail: Eddy.Campbell@CMIS.CSIRO.AU Beijing, 100081, China

e-mail: chenym@ns.ami.ac.cn

Hong Xing CAO or: youmin.chen@263.net

Chinese Academy of meteorological Sciences,

46 Baishiqiaolu Jean-Paul CHILES

100081 , Beijing ,China Bureau de Recherches Gologiques et Minires

e-mail : caohx@netease.com BRGM 3, avenue Claude Guillemin

BP 6009, 45060 Orlans Cedec 2, France

Carmen CAPILLA

Universidad Politecnica, Bernard CORNLIS

Depto. Estadistica e Investigacion Operativa Apartado Dpartement de Gomatique Universit de Lige

Apartado 22012-46071 Valencia, Spain Place du 20-Aot, 7

e-mail:ccapilla@eio.upv.es 4000 Lige, Belgium

e-mail: Bernard.Cornelis@ulg.ac.be

Dr. Peter CARL

Institute for Freshwater Ecology Dan CORNFORD

und Inland Fisheries (IGB) Computer Science

66

Aston University Institute for Geophysics, Astrophysics, and Meteorology

Aston Triangle University of Graz (IGAM/UG)

Birmingham B4 7ET, UK Universitaetsplatz 5

e-mail: d.cornford@aston.ac.uk A-8010 Graz, Austria

e-mail: ulrich.foelsche@kfunigraz.ac.at

Joao CORTE-REAL

Univ. Lisbon, Portugal Chris FOREST

e-mail: jcr@fc.ul.pt Massachusetts Instituts of Technology

Dept. Earth, Atmos. And Plan.

Traute CRGER 77 Massachusetts Avenue, Cambrigde,

Meteorological Institute MA 02139-4307 USA

University of Hamburg e-mail: ceforest@MIT.EDU

Germany

Carsten S. FREDERIKSEN

Moses DAMOAH Principal Research Scientist

- c/o Rev Joshua Damoah BMRC Model Evaluation Group

Private-Mail Bag Bureau of Meteorology Research Centre

Abesim Sunyani-B/A Melbourne, Australia

Ghana W/A e-mail: c.frederiksen@bom.gov.au

- Kwaku Damoah

48 Yorkshire Road Thomas K. FRIEDLI

Mitcham, Surrey CR4 1QD, UK Institut fr mathematische Statistik und

Versicherungslehre

Utpal .K. DE Universitt Bern

Jadavpur University Sidlerstrasse 5

Calcutta, India 3012 Bern, Switzerland

e-mail: deutpal@hotmail.com e-mail: friedli@stat.unibe.ch

Department of Physics G. Galilei Dto Fsica de la Tierra II

via Marzolo 8 Universidad Complutense

35131 Padova, Italy Madrid, Spain

e-mail: dezolt@pd.infn.it e-mail: rgarcia@6000aire.fis.ucm.es

Hungarian Meteorological Service Universitt Bonn

1525 Budapest P. O. Box. 38, Hungary Auf dem Hgel 20

e-mail: dobi.i@met.hu 53121 Bonn, Germany

e-mail:c.gebhardt@uni-bonn.de

Mojca DOLINAR, Hydrometeorological Institute of

Slovenia, LJUBLJANA, Slovenia D. GELLENS

Computer Centre

Martin DUBROVSKY Royal Meteorological Institute of Belgium

Institute of Atmospheric Sciences Av Circulaire, 3

Husova 456 180 Brussels, Belgium

50008 Hradec Kralove e-mail: D.Gellens@oma.be

Czech Republic

e-mail: dub@ufa.cas.cz Friedrich Wilhelm GERSTENGARBE

Potsdam-Institut fr Klimafolgenforschung

Salaheddine EL-ADLOUNI POB 601203

Univ. Mohammed V, Facult des Sciences 14412 Potsdam, Germany

Dp. de Mathmatiques et Informatique. e-mail: gerstengarbe@pik-potsdam.de

BP. 1014; Rabat Maroc

e-mail: s_adlouni@hotmail.com Luis GIMENO

Universidad de Vigo, Facultad de Ciencias de Orense,

Geir EVENSEN 32004 Orense, Spain

Nansen Environmental and Remote Sensing Center, Tfno: 988387208

Edvard Griegsvei 3a, Fax: 988387159

5037 Solheimsviken, Norway e-mail: l.gimeno@uvigo.es

e-mail: Geir-Evensen@nrsc.no

Clare GOODESS

Frauke FESER Climatic Research Unit

Institute of Hydrophysics University of East Anglia

GKSS Research Centre Norwich NR4 7TJ, UK

Geesthacht, Germany e-mail: c.goodess@uea.ac.uk

e-mail: feser@gkss.de

Jrgen GRIESER

Gtz FLSER Johann-Wolfgang Goethe Universitt Frankfurt am Main

GKSS Research Centre Institut fr Meteorologie & Geophysik

Geesthacht, Germany Postfach: 11 19 32

e-mail floeser@gkss.de D - 60054 Frankfurt a. M.

e-mail:Grieser@meteor.uni-frankfurt.de

Dr. Ulrich FOELSCHE

67

Jrgen Grieser 53121 Bonn, Germany

Food and Agriculture Organization of the United Nations e-mail ahense@uni-bonn.de

SDRN - C 854 Via delle Terme di Caracalla

00100 Rome, Italy Carme HERVADA i Sala

Fisica i Enginyeria Nuclear

Anders GRIMVALL L'escola industrial (UPC)

Department of Mathematics Colom 1

Division of Statistics 08222 Terrassa, Spain

Linkping University e-mail: carme.hervada@upc.es

SE-58183 Linkping, Sweden

e-mail: angri@mai..liu.se Heike HOFFMANN

Graduiertenkolleg Rumliche Statistik

Nikolaus GROLL Bernhard-v.-Cotta-Str. 2

Institute of Hydrophysics 09596 Freiberg, Germany

GKSS Research Centre e-mail: hoffmann@merkur.hrz.tu-freiberg.de

Postfach 1160

21494 Geesthacht, Germany Greg HOLLOWAY

e-mail: groll@gkss.de Institute for Ocean Sciences

9860 West Sannich Road

Peter GUTTORP Sidney, B.C. V8L 4 B2, Can.

University of Washinton e-mail: hollowayg@dfo-mpo.gc.ca

Seattle, WA, USA

e-mail: peter@stat.washington.edu Ibrahim HOTEIT

Laboratoire de Modelisation et Calcul

Sven HARIG Grenoble, France

Alfred-Wegener-Institut e-mail: Ibrahim.Hoteit@imag.fr

Bussestrae 24

D-27570 Bremerhaven Richenda HOUSEAGO-Stokes

Germany James Rennell Division for Ocean Circulation and

Climate

Knut HARSTVEIT Southampton Oceanography Centre

Pb 43 Blindern Empress Dock

0313 OSLO, Norway Southampton, SO14 3ZH, UK

e-mail: knut.harstveit@dnmi.no e-mail: R.Houseago@soc.soton.ac.uk

Meteorological and Geophysical Agency Dept. of Biostatistics

Jl. Angkasa I No. 2 University of Washington

Jakarta, Indonesia Mailstop 359931

e-mail : urip_haryoko@hotmail.com Seattle, WA 98195

e-mail: hughes@biostat.washington.edu

Klaus HASSELMANN

Max-Planck-Institute for Meteorology Radan HUTH

Bundesstr. 55, 20146 Hamburg, Germany Institute of Atmospheric Physics

e-mail: klaus.hasselmann@dkrz.de Prague, Czech Republic

e-mail: huth@ufa.cas.cz

Gabi HEGERL

Gabi Hegerl and Tom Crowley Knut A. IDEN

Dept. of Oceanography Norwegian Meteorological Institute (DNMI)

Texas A&M P.O.Box43 Blindern

College Station, TX 77843, USA N-0303 Oslo

e-mail: paleo@cox-internet.com Norway

e-mail: knut.iden@dnmi.no

Gabi HEGERL

University of Washinton Valerie ISHAM

Seattle, WA, USA Dept. Of Statistical Science

e-mail: hegerl@atmos.washington.edu University College

Gower Street

Dr. Hermann HELD London WC1E 6BT, UK

Potsdam Institute for Climate Impact Research (PIK) e-mail: valerie@stats.ucl.ac.uk

Telegrafenberg C4, 14473 Potsdam

14412 Potsdam, Germany Shaleen JAIN

e-mail: held@pik-potsdam.de Cooperative Institute for Research in Environmental

Sciences (CIRES)

Jaakko HELMINEN University of Colorado

Finnish Meteorological Institute Campus Box 216

Helsinki, Finland Boulder, CO 80303-0216 USA

Jaakko.Helminen@fmi.fi E-mail: sjain@cdc.noaa.gov

Office: NOAA-CDC/David Skaggs Research Center,

Andreas HENSE, Room 1D-111

Meteorological Institute

University of Bonn Stephen JEWSON

Auf dem Hgel 20 Risk Management Solutions, Inc.

68

Menlo Park Gennady A. KIVMAN

California, USA Alfred-Wegener-Institut for Polar- and Marine Research

e-mail: stephen.jewson@riskinc.com Bremerhaven,, Germany

e-mail: gkivman@awi-bremerhaven.de

Ian JOLLIFFE

University of Aberdeen, UK Richard KLEEMAN

e-mail: itj@math.abdn.ac.uk Courant Institute for Mathematical Science

New York University, USA

Charles JONES e-mail: kleeman@cims.nyu.edu

Assistant Researcher

Institute for Computational Earth System Science Stanislava KLIEGROVA

(ICESS) Czech Hydrometeorological Institut

University of California, Santa Barbara Branch Hradec Krlov

CA 93106-3060, USA Dvorsk 410

e-mail: cjones@icess.ucsb.edu 503 11 Hradec Krlov - Svobodn Dvory, Czech

Republic

Julie M. JONES e-mail: stanislava.kliegrova@chmi.cz

Institute of Hydrophysics

GKSS Research Centre Ken-Chung KO

Postfach 1160 Department of Geography

21494 Geesthacht, Germany National Kaohsiung Normal University

e-mail: jones@gkss.de 116 Ho-Ping First Road

Kaohsiung, Taiwan

Eigil KAAS e-mail: kko@nknucc.nknu.edu.tw

Danish Meteorological Institute

Lyngbyvej 100 Till KUHLBRODT

2100 Copenhagen, DENMARK Potsdam Institute for Climate Impact Research (PIK)

e-mail: ek@dmi.dk Climate System Dept.

Telegrafenberg, PO Box 60 12 03

Tiba KABANDA 14412 Potsdam, Germany

School of Environmental Sciences, University of Venda e-mail: kuhlbrodt@pik-potsdam.de

Private Bag, X5050, Thohoyadou, 0950, South Africa

e-mail: Kabanda@univen.ac.za Ashwini KULKARNI

Indian Institute of Tropical Meteorology

Jaroslava KALVOVA Pashan, Pune 411008, India

Dept. of Meteorology and Environment Protection e-mail : ashwini@tromet.ernet.in

Faculty of Mathematics and Physics

Charles University Venerabilis KULULETERA

V Holeovikch 2 Tanzania Meteorological Agency

180 00 Praha 8, Czech Republic P.O.Box 3056

e-mail: jaroslava.kalvova@mff.cuni.cz Dar es Salaam

Tanzania

S.S. KANDALGAONKAR e-mail: kululetera@hotmail.com

Physical Meteorology and Aerology Division

Indian Institute of Tropical Meteorology Vit KVETON

N.C.L. Post Office, Dr. Homi Bhabha Road Czech Hydrometeorological Institute

PUNE 411 008, India Prague, Czech Republic

e-mail: sskandal@tropmet.ernet.in Na abatce 17

143 06 Prague 4, Czech Republic

Mohammed-Said KARROUK e-mail: vit.kveton@chmi.cz

Universitt Hassan II

Centre de Recherche de Climatologie Frank KWASNIOK

BP 8220 Oasis, MA-20103, Casablanca, Morocco Leibniz-Institut fr Atmosphrenphysik

e-mail:karrouksaid@yahoo.com Schlossstrasse 6

18225 Khlungsborn, Germany

Rick KATZ e-mail: kwasniok@iap-kborn.de

Environmental and Societal Impacts Group

National Center for Atmospheric Research Won-Tae KWON

Boulder, CO 80307 Meteorological Research Institute

Colorado, USA Korea Meteorological Administration, Korea

e-mail: rwk@ucar.edu e-mail: wontk@metri.re.kr

Alfred-Wegener-Institute for Polar and Marine Research Institute of Atmospheric Physics

Bussestr. 24 Prague, Czech Republic

27570 Bremerhaven, Germany e-mail: honza@ufa.cas.cz

e-mail: fkauker@awi-bremerhaven.de

Victor E. LAGUN

Dr A.S.A. KHALIL Air-Sea Interaction Department

Chief Technical Advisor Arctic and Antarctic Research Institute

United nations Development programme Beringstr. 38

UAE/99/001 Project St. Petersburg 199397, Russia

e-mail: lagun@aari.nw.ru

69

Monika LAKATOS Solange MENDONCA LEITE

Hungarian Metrological Service Physics Department

e-mail: lakatos.m@met.hu University of Tras-os-Montes e Alto Douro

PoBox 202

David R. LEGATES 5000-911 Vila Real, Portugal

Center for Climatic Research e-mail: solange@utad.pt

University of Delaware

227 A Pearson Hall Miroslaw MIETUS

Newark, DE, 19716, USA Institute of Meteorology and Water Management

e-mail: legates@udel.edu Maritime Branch

Gdynia, Poland

Piero .LIONELLO e-mail: mietus@imgw.gdynia.pl

Dept. of Material Sciences

University of Lecce Janos MIKA

Via per Arnesano Hungarian Meteorological Service

73100 Lecce, Italy H-1525 Budapest P. O. Box. 38.

e-mail: lionello@pd.infn.it (H-1024 Budapest Kitaibel P. u. 1), Hungary

e-mail: mika.j@met.hu

Ulrike LOEPTIEN

Universitt Kiel Adam MONAHAN

Institut fr Meteorologie Institut fuer Mathematik

e-mail: uloeptien@ifm.uni-kiel.de Humboldt-Universitaet zu Berlin

Unter den Linden 6

Leonid LOPATOUKHIN 10099 Berlin, Germany

State University (Dep.Oceanography),

199178, St.Petersburg, 10 Line, 33/35, Russia M. Granger MORGAN

e-mail: leonid@LL1587.spb.edu Carnegie-Mellon University

5000 Forbes Ave., Pittsburgh

Omar Abel LUCERO PA 15213, USA

Omar Abel Lucero e-mail: gm5d@andrew.cmu.edu

National University of Cordoba

Centro de la Regin Semirida (INA), Argentina Manfred MUDELSEE

e-mail: Omarabellucero@yahoo.com Institute of Meteorology

olucero@com.uncor.edu University of Leipzig

olucero@gtwing.efn.uncor.edu Stephanstr. 3

D-04103 Leipzig, Germany

M Dolores MANSO ORGAZ e-mail: mudelsee@rz.uni-leipzig.de

Depart. of Physics

University of Aveiro, Peter MLLER

Portugal Department of Oceanography

e-mail: mariola@fis.ua.pt University of Hawaii, USA

e-mail: pmuller@iniki.soest.hawaii.edu

Rol MADDEN

UCAR, Boulder, Colorado, USA Brbel MLLER-KARULIS

e-mail ram@ucar.edu Marine Monitoring Centre

Institute of Aquatic Ecology

Ileana MARES University of Latvia

Constantin MARES 8 Daugavgrivas

National Institute of Meteorology and Hydrology 1007 Riga, Latvia

Sos. Bucuresti-Ploiesti 97 e-mail: baerbel@monit.lu.lv

Bucharest, 71552, ROMANIA

e-mail: imares@meteo.inmh.ro Donald E. MYERS

cmares@meteo.inmh.ro Department of Mathematics

University of Arizona

Jorge MATEU Tucson, AZ 85721, USA

Department of Mathematics e-mail: myers@math.arizona.edu

University Jaume I

Campus Riu Sec Philippe NAVEAU

12071 Castellon, Spain Geophysical Statistics Project

e-mail: mateu@mat.uji.es National Center for Atmospheric Research,

Climate and Global Dynamics Division

Istvan MATYASOVSZKY Boulder, CO, USA,

Department of Meteorology e-mail: pnaveau@cgd.ucar.edu

Eotvos Lorand University

1117 Budapest Pzmny Pter stny 1/A Lars NERGER

1364 Budapest, Hungary Alfred-Wegener-Institut

e-mail: istvan.matyasovszky@elte.hu fuer Polar- und Meeresforschung

Rechenzentrum

Norm MCFARLANE Buergermeister-Smidt-Strasse 20

Canadian Centre for Climate Modeling and Analysis 27568 Bremerhaven, Germany

Victoria, Canada e-mail: lnerger@awi-bremerhaven.de

e-mail: norm.mcfarlane@ec.gc.ca

70

Svetlana NIKULINA Wind Science and Engineering Program

Central Asian Research Hydrometeorological Institute Texas Tech University

(SANIGMI) Lubbock, TX 79409 USA

Tashkent, Uzbekistan e-mail: richard.peterson@ttu.edu

e-mail: sanigmi@meteo.uz

Arnt PFIZENMAYER

Carlos NOBRE Institute of Hydrophysics

Centro de Previso de Tempo e Estudos Climaticos GKSS Research Centre

Cachoeira Paulista, Brasil Geesthacht, Germany

e-mail: nobre@cptec.inpe.br e-mail: pfitzenmayer@gkss.de

Statistics Division CGUL, - Centro de Geofsica da Universidade de Lisboa

Department of Mathematics Rua da Escola Politcnica 58

Linkping University 1250 Lisboa, Portugal

SE-581 83 Linkping, Sweden e-mail: cpires@fc.ul.pt

e-mail: annor@mai.liu.se

Faculdade de Cincias

Doug NYCHKA Departamento de Fsica, Edifcio C8

Geophysical Statistics Project Rua Ernesto de Vasconcelos - Campo Grande

National Center for Atmospheric Research 1749-016 Lisboa Codex - PORTUGAL

Climate and Global Dynamics Division e-mail cpires@fc.ul.pt

Boulder, CO, USA

e-mail: nychka@cgd.ucar.edu Henrique Oliveira PIRES

Instituto de Meteorologia

Guillermo O. OBREGON R. C ao Aeroporto

Centro de Previsao de Tempo e Estudos Climaticos Lisboa, Portugal

(CPTEC) e-mail: hop@esoterica.pt

Instituto Nacional de Pesquisas Espaciais (INPE)

Rod. Presidente Dutra, km. 40 Budong QIAN

12630-000 Cachoeira Paulista, Sao Paulo, Brazil ICAT - Faculty of Sciences

University of Lisbon, Portugal

Timothy J. OSBORN e-mail: bdqian@fc.ul.pt;

Climatic Research Unit jcr@fc.ul.pt; chingxu@fc.ul.pt

School of Environmental Sciences

University of East Anglia, Radhika RAMACHANDRAN

Norwich NR4 7TJ, UK Space Physics Laboratory

e-mail: t.osborn@uea.ac.uk Vikram Sarabhai Space Centre

Trivandrum - 695 022

Taha B. M. J. OUARDA INDIA

e-mail:radhika@md3.vsnl.net.in

Christopher J. PACIOREK

Department of Statistics Nazario D. RAMIREZ-Beltran

Carnegie Mellon University Department of Industrial Engineering

Pittsburgh, Pennsylvania, USA University of Puerto Rico

e-mail: paciorek@stat.cmu.edu Mayaguez Puerto Rico 00680

e-mail: na_ramirez@rumac.uprm.edu

Heiko PAETH

Meteorologisches Institut der Universitt Bonn Clarence RICHARDSON

Auf dem Hgel 20 Laboratory Director

53121 Bonn, Germany Grassland Soil & Water Research Laboratory

e-mail: hpaeth@uni-bonn.de 808 E. Blackland Road

Temple, TX 76502, USA

Daniela Mihaela PALIU richards@brc.tamus.edu

National Institute of Meteorology and Hydrology

Sos. Bucuresti-Ploiesti 97 Pedro RIBERA

71552 Bucharest, Romania e-mail: pribera@uvigo.es

e-mail: dpaliu@meteo.inmh.ro

Joao RIO

Ruben PASMANTER Centro de Geofsica da Universidade de Lisboa;

Faculty of Science PORTUGAL

University of Amsterfdam e-mail : jrio@fc.ul.pt

The Netherlands

e-mail: pasmant@wins.uva.nl James RISBEY

University of Utrecht

Sarath PEIRIS Department of Science, Technology and Society

e-mail: rescri@sri.lanka.net Padualaan 14

3584 CH Utrecht, The Netherlands

Richard E. PETERSON, Chair e-mail: james@nwsmail.chem.uu.nl

Professor of Atmospheric Science

Associate Director Torben SCHMITH

Department of Geosciences Danish Meteorological Institute

71

Climate Research Division

Lyngbyvej 100 Dietrich STOYAN

2100 Copenhagen , DENMARK Institut fr Stochastik

e-mail: ts@dmi.dk Bergakademie Freiberg, Germany

e-mail: stoyan@orion.hrz.tu-freiberg.de

Vladimir A. SEMENOV

Max-Planck-Institut fr Meteorologie Daithi STONE

Bundesstr., 55 School of Earth and Ocean Sciences

20146 Hamburg, Germany University of Victoria

e-mail: semenov@dkrz.de PO Box 3055, Victoria

BC, V8W 3P6, Canada

Julien SNEGAS e-mail: stone@ocean.seos.uvic.ca

Centre de Geostatistique

Ecole des Mines de Paris Hans von STORCH

77305 Fontainebleau GKSS Research Centre

e-mail: senegas@cg.ensmp.fr Geesthacht, Germany

e-mail: storch@gkss.de

Amjad Majeed SHAH

Assistant Professor Jin-Song von STORCH

Statistics Department Meteorologisches Institut der Universitt Hamburg

Islamia University Bahawalpur, Pakistan Germany

e-mail: Amjee2001@yahoo.com e-mail: jinsong@gkss.de

School of Management Ukrainian Hydrometeorological Institute,

Indian Institute of Technology 37, Nauki av., Kiev, Ukraine

Bombay,Powai Phone: (044) 265 87 90

Mumbai-400076, India Fax: (044) 265 53 63

e-mail: ajay@cc.iitb.ernet.in e-mail: nigmi2@mail.kar.net

Indian Institute of Tropical Meteorology Climatological and Agrometeorological Division

Dr. Homi Bhabha Road Hungarian Meteorological Service

Pashan, Pune- 411 008, India 1525--Budapest P.O. Box 38, Hungary

e-mail: nsingh@tropmet.ernet.in e-mail: szentimrey.t@met.hu

Centre for the Analysis of Time Series Computa of Cientfica, Petrpolis, Brazil

London School of Economics

London, WC2A 2AE, UK Karl E. TAYLOR

e-mail: lenny@maths.ox.ac.uk P.O. Box 808, L-264

Livermore, CA 94550

Keon Tae SOHN email: taylor13@llnl.gov

Pusan National University, Korea

e-mail: ktsohn@hyowon.cc.pusan.ac.kr Jonathan TAYLOR

Lancaster University, UK

Neelima SONTAKKE

Indian Institute of Tropical Meteorology Susanne THEIS

Pashan, Pune-411008, India Meteorologisches Institut der Uni Bonn

Auf dem Hgel 20

Tim SPARKS 53121 Bonn, Germany

Centre for Ecology and Hydrology e-mail: susanne.theis@uni-bonn.de

Monks Wood, Abbots Ripton

Huntingdon Peter THORNE

Cambridgeshire PE28 2LS, UK Climatic Research Unit

e-mail: ths@ceh.ac.uk or t.sparks@ite.ac.uk School of Environmental Sciences

University of East Anglia

Tim STAEGER Norwich, NR4 7TJ, UK

J.-W. Goethe-Universitdt Frankfurt e-mail: Peter.Thorne@uea.ac.uk

Institut fr Meteorologie und Geophysik

Arbeitsgruppe Meteorologische Marina TIMOFEYEVA

Umweltforschung/Klimatologie National Weather Service

Postfach 11 19 32 1325 East West Hwy

D-60054 Frankfurt a.M., Germany Silver Spring, MD 20910 3283

mail: staeger@meteor.uni-frankfurt.de e-mail: marina.timofeyeva@noaa.gov

National Technical University of Athens Johann Wolfgang Goethe Universitt

Dept. of Naval Architecture & Marine Engineering Institut fr Meteorologie und Geophysik

P.O.BOX 64033 Robert-Mayer-Str. 1

Zografos, 15710 60054 Frankfurt a.M.,

Athens, Greece e-mail: S.Troemel@metteor.uni-frankfurt.deGermany

e-mail: chstef@central.ntua.gr

72

Heikki TUOMENVIRTA Bundesstrasse 55

Finnish Meteorological Institute D-20146 Hamburg, Germany

P.O. Box 503 e-mail: blender@dkrz.de

00101 Helsinki, Finland

e-mail: heikki.tuomenvirta@fmi.fi Abebe YESHANEW

National Meteorological Services Agency

Silvia A. VENEGAS Addis Abeba, Ethiopia

Danish Center for Earth System Science (DCESS) e-mail: a_yeshanew@hotmail.com

Niels Bohr Institute for Astronomy, Physics and

Geophysics Xiaogu ZHENG

University of Copenhagen, Denmark The National Institute of Water and Atmospheric

e-mail: silvia@apollo.dcess.ku.dk Research

301 Evans Bay Parade, Greta Point

Lucie VINCENT P. O. Box 14901

Climate Research Branch Wellington New Zealand

Meteorological Service of Canada e-mail: x.zheng@niwa.cri.nz

4905 Dufferin Street

Downsview, Ontario, Canada Christine ZIEHMANN

M3H 5T4 Universitt Potsdam - Institut fr Physik

e-mail: Lucie.Vincent@ec.gc.ca Am Neuen Palais 10 - Haus 19

Postfach 601553

Aldo J. VIOLLAZ 14415 Potsdam, Germany

Universidad Nacional de Tucuman e-mail: chriss@agnld.uni-potsdam.de

Argentina

e-mail: aviollaz@noanet.com.ar; Olga ZOLINA

aviollaz@herrera.unt.edu.ar. P.P.Shirshov Institute of Oceanology

Russian Academy of Science (IORAS)

Hans Wackernagel 36 Nakhimovsky ave.

117851 Moscow, Russia

Ecole des Mines e-mail: zol@gulev.sio.rssi.ru

Paris, France

e-mail: hans@cg.ensmp.fr Francis ZWIERS

Canadian Center for Climate Modelling and Analysis

Andreas WALTER Meteorological Service of Canada

Institute of Meteorology and Geophysics c/o University of Victoria

Johann-Wolfgang Goethe University PO Box 1700, STN CSC

Frankfurt/Main, Germany Victoria, BC V8W 2Y2, Canada

e-mail: A.Walter@meteor.uni-frankfurt.de e-mail fzwiers@ec.gc.ca

Network Strategies Division, AMWSD

Meteorological Service of Canada

4905 Dufferin Street, Downsview

Ontario, Canada M3H 5T4

Ed WAYMIRE

Oregon State University, Dept. of Mathematics,

Kidder Hall 368, Corvallis, OR 97331-4605, USA

e-mail: waymire@math.orst.edu

Joanna WIBIG

Dept. of Meteorology and Climatology

University of Lodz

Lipowa 81

90-568 Lodz, Poland

zameteo@krysia.uni.lodz.pl

Martin WIDMANN

Institute of Hydrophysics

GKSS Research Centre

Postfach 1160

21494 Geesthacht, Germany

e-mail: widmann@gkss.de

Lawrence WILSON

Environment Canada

e-mail: Lawrence.Wilson@ec.gc.ca

Katja WOTH

Jun-Ichi YANO

Meteorologisches Institut

Universitt Hamburg

73

von Storch, 2, 7, 8, 15, 37, 41

74

- Uncertainties in climate change projections and regional downscaling 2010Uploaded byIrene Torres
- Shelter for Warm-humid ClimatesUploaded byAnn Nambiaparambil
- Jan 15 CSUploaded byAysh
- Application of Data Mining Techniques in Weather Prediction and Climate Change Studies.pdfUploaded bySampath Amarakoon
- Local Perceptions of Climate Variability and Change in Tropical Forests of Papua, IndonesiaUploaded byBachrul Ulum, SP
- Typh0000Uploaded byPauline Tracy Batiles
- One of the Biggest Thresholds Has Already Been ReachedUploaded byDoug Grandt
- Adapting to Climate Change Thresholds, Values, GovernanceUploaded byAnonymous orJf8e1kk
- 5 pidatoUploaded byTamara Fauziyah
- climateregionUploaded byapi-265618999
- Solar EteUploaded byVivekSingh
- Climate ChangeUploaded byTeguhAmbanSantoso
- 2017 Was the Year of the Billion-Dollar Disaster _ Climate CentralUploaded byGustavo Fonseca
- HurricanesUploaded bysashina
- Dispersion ModelingUploaded byAnuradha Chauhan
- States Have No Responsibility to Other StatesUploaded byJoanne Esther
- Revised Geo 1200 Syllabus Fall 13Uploaded byBastianGonzalez
- A Nemo MetroUploaded byFilgueirasJr
- Ao Xiii, 1-2, 2006 PapersUploaded byBurn Bianca
- try out un XIIUploaded byteree_mail8332
- IndiaUploaded byChris H-J
- trsUploaded byzz
- Reliant Roofing Launches New and Innovative Solar and Hurricane Protection PackagesUploaded byPR.com
- Watch+Technical+Report+Number+32+Revisiting+Glace+ +Understanding+the+Role+of+the+Land+Surface+in+Land Atmosphere+CouplingUploaded bytgimran
- A Great Tourist DestinationUploaded byVivian Labelle
- Hurricane ProjUploaded byTonia Darrisaw Parker
- SINGER, Peter Hurricanes’ Unnatural TollUploaded byAnonymous ykRTOE
- Kevin Collins WPE 2009Uploaded byCharles Thompson
- Meteorological HazardsUploaded byjames_chong_16
- Florida climate science letter to Donald Trump: WMNF NewsUploaded byWMNF News

- Agudelo P.a. 2004 - Analysis of Spatial Distribution in Tropospheric Temperature TrendsUploaded byerik
- Alexander D. 1999 - Temperature structure of the quiet Sun X ray corona.pdfUploaded byerik
- Alexander D. 1999 - Temperature structure of the quiet Sun X ray corona.pdfUploaded byerik
- Aires F. 2002 - Remote Sensing From the Infrared Atmospheric Sounding Interferometer InstrumentUploaded byerik
- Alanko-Huotari K. 2007 - Cyclic Variations of the Heliospheric Tilt Angle and Cosmic Ray ModulationUploaded byerik
- Alanko-Huotari K. 2007 - Cyclic Variations of the Heliospheric Tilt Angle and Cosmic Ray ModulationUploaded byerik
- Aguiar-Conraria L 2011 - The Continuous Wavelet Transform a PrimerUploaded byerik
- Akasofu S.I. 1984 - Sunspot Pair Formation by the Photospheric Dynamo ProcessUploaded byerik
- Ajabshirizadeh a. 2008 - Wavelet Analysis of Solar Macro-spicule RecurrencesUploaded byerik
- Allen M.R. 1996 - Monte Carlo SSA_ Detecting Irregular Oscillations in the Presence of Coloured NoiseUploaded byerik
- Akasofu S I 2004 - A New Morphology of Solar Activity and Recurrent Geomagnetic Disturbances the Late-Declining Phase of the Sunspot CycleUploaded byerik
- Alcayde D. 1974 - Long-Term Variations of Thermospheric Temperature and CompositionUploaded byerik
- Akbari H. 2009 - Global Cooling Increasing World-wide Urban Albedos to Offset CO2Uploaded byerik
- Aikin a.C. 1991 - Temperature Trends in the Lower MesosphereUploaded byerik
- Alexander M a 2002 - The Atmospheric Bridge the Influence of ENSO Teleconnections on Air–Sea Interaction Over the Global OceansUploaded byerik
- Enghoff M. B. 2008 - The Role of Atmospheric Ions in Aerosol Nucleation – a ReviewUploaded byerik
- Zhou T. 2006 2006 - Twentieth-Century Surface Air Temperature Over China and the Globe Simulated by Coupled Climate ModelsUploaded byerik
- Chattopadhyay S. 1999 - Sunpsots and Cycles ReconsideredUploaded byerik
- Anexos FinalUploaded byerik
- Graficación en Dos y Tres DimensionesUploaded byerik
- LEY ORGÁNICA DE LA UNIVERSIDAD NACIONAL AUTÓNOMA DE MÉXICO.pdfUploaded byerik
- Teotihuacan 2Uploaded byerik
- Barnett J.J. 1984 - Temperature Comparisons Between the NIMBUS 7 SAMS, RocketRadiosondes and the NOAA 6 SSUUploaded byerik
- Raspopov O.M. 2004 - Periodicity of Climate Conditions and Solar Variability Derived From Dendrochronological and Other Palaeoclimatic Data in High LatitudesUploaded byerik
- Bell B. 1965 - Lunar Eclipses and the Forecasting of Solar MinimaUploaded byerik
- Iniciativa Ley de Participacion CiudadanaUploaded byErik Martinez
- Adelfang S.I. 1970 - Amplitude of the Annual Temperature Cycle in the StratosphereUploaded byerik
- Zhou Y. 1998 1998 - Interannual Polar Motion With Relation to the North Atlantic OscillationUploaded byerik
- Function ListUploaded byerik
- Derecho AccesoUploaded byManuel Limita

- Statically Inde. StructuresUploaded bySalem Albarki
- Master SlaveUploaded byrocket_spirit_mani
- Scientific Method Terms - Heart Rates LabUploaded byjerome
- formal and informalUploaded byapi-246444495
- Computational LexicographyUploaded byDovilė Landauskaitė
- logic gate - -Uploaded byapi-254068557
- History of LinguisticsUploaded byDarie Nicolae
- A 3D linearized method for turbo machinery tone noise analysis with 3D non-reflecting boundary conditionsUploaded bynascosannascosan
- unit 8 VTU format.pdfUploaded byammayi9845_930467904
- the data exploration mini projectUploaded byapi-268991410
- Swipe - Image Segmentation based gesture controlUploaded byTanay
- PrcUploaded byJoseph Jones
- ME 595 695 Introduction No HwUploaded byShail Satak
- [67]_Unbalanced Magnetic ForcesUploaded byrakeshee2007
- Smoothed Particle Hydrodynamics Modelling of Fluids and SolidsUploaded byAaniya Cyril
- Simulated EvolutionUploaded bySauravRoy
- CCP303Uploaded byapi-3849444
- Polygons for KidsUploaded bycircleteam123
- bravo2015.pdfUploaded bySajid Mohy Ul Din
- BossUploaded byhbuddy
- D49996GC11_sg3 Fundamentals I Volume IIIUploaded bydjpunkful
- GESEKANUploaded bykidispur
- 01400_Uploaded bybeo
- 2.a New Compact Dual Band GPS Patch Antenna Design BasedUploaded byPhạm Hoàng
- Bentley Rm Bridge Advanced Detalii Module AditionaleUploaded byphanoanhgtvt
- Pelton WheelUploaded bylimakupang_mat
- Forest inventories by LiDAR data: A comparison of single treesegmentation and metric-based methods for inventories of aheterogeneous temperate forestUploaded byJhon Huacani
- ChengUploaded byDana Skakova
- Vertical Curves WordUploaded byDushan Senarathne
- flowers for algernon multiple intelligences projectUploaded byapi-340486727