Introducing GWSDAT (GroundWater Spatiotemporal

Data Analysis Tool)

Wayne Jones*, Mike Spence*, Matthew Lahvis**
*Shell Global Solutions (UK). email: Wayne.W.Jones@shell.com , Mike.Spence@shell.com
**Shell Global Solutions (US) Inc.

GWSDAT v2.0

Business Benefits

GWSDAT is a user- friendly Microsoft Excel Add-In that can be used to visualise and
interpret groundwater monitoring data.

Improved data transparency helps design and optimise groundwater monitoring or

remediation programmes (i.e. avoid the collection of redundant data)

Key features include:

Early identification of new releases, migration pathways, need for corrective action and
stable/ declining trends that may aid in closure determinations

Spatiotemporal modelling capability: uses statistical methods to continuously model

variation in groundwater solute concentration as a function of X,Y and time

Automatic generation of concentration contour plots at user specified time intervals,

with the option to overlay groundwater elevation contours and NAPL thickness/ footprint
data

Rapid interpretation of complex data sets from large monitoring networks (e.g. refineries,
terminals)

Get the most out of monitoring data with the click of a mouse

Time series plots of groundwater

elevation, NAPL thickness and solute
concentration

Contour plots of groundwater elevation

and solute concentration
o Concentration contours are generated from a spatiotemporal smoother algorithm

o Generated automatically for any well/analyte

combination

o Groundwater elevation contours are fitted by kriging

o Automatic fitting of linear or non- parameteric trend

overlay functions, with 95% confidence limits

o NAPL thickness and extent is indicated by coloured,

sized circles (bubble plot)

o Graphics can be exported directly to Powerpoint or to

other applications for reporting purposes.

o Site plan can be imported as background image

Toluene in MW8
224480

>5000
5000

224470

GDBH1 03

224460

MW4

800

GDBH 105

400

ND<1

224450

GDBH10 4

200

MW3

MW1

SGS1

100

MW102
MW101
MW8

75

2
MW10

116 00

SGS4 P3

MW2

GDBH101

50

16200
13900
16800
SGS2
MW9

57.6

57.8

1500

MW104

BH1

224440

58.0

1e+04
1e+02

3000

MW103
GDBH102

25

224430

GW Elevation

58.4

Non-Detect Data

58.2

Linear Conc. Trend

Groundwater Elevation (metres)

1e+06

Detectable Data

1e+00

10
BH3
MW7

ND<1

10

ND<1

MW6

MW5

2006

2007

2008

2009

224420

Toluene (ug/l)

ug/l

Toluene : 04-Se p-2006 to 03-Oct-2006

Mann-Kendall P.Value= 0.0995; Half-Life= 201 days

2010

0
551640

Date

551650

551670

551680

551690

551700

NAPL
Th ickness
(m m )

224480

Toluene (ug/l) : 04-Se p-2006 to 03-Oct-2006

Conc. Smoother

NAPL Thickness

Non-Detect Data

95% conf. interval

NAPL Substituted Data

1700
1600
1500

GDBH103

224470

Detectable Data

12

1e+06

TPH in MW8

551660

1400
MW103
GDBH102

MW10 4

10

BH1

1300
1200

224460

N D<1

1100

GDBH104

GDBH10 5

224450

1000
MW3

900

MW1

SGS1

800

MW102
MW101

700
MW8
2
MW10

600

11600

SGS4 P3

MW2
GDBH101

224440

8
6

TPH (ug/l)
1e+04

NAPL Thickness (mm)

1e+05

MW4

500

1 6200
6800
3900
SGS2
MW9

224430

BH3
MW7

300

ND<1
ND<1

1e+03

MW6

200

2007

2008

2009

Well/ boring locations colour- coded red to indicate current/ historical

presence of LNAPL (i.e. footpint), to aid remedial decision making

2010

Date

MTBE : 17-May-2006 to 16-Jun-2006

ug/l
>5000

7500

5000

Capable of handling large datasets for

complex sites

224420

MW5

2006

100
0
551640

551650

551660

551670

551680

Spatiotemporal Prediction

.6
40

1500

.8
40

o Example shows MTBE plume at

manufacturing site

o Generated automatically for any

well/analyte combination

o 40,000 lines of input data in dataset

processed in only 30 minutes.

o Provides rapid visual check on quality of

spatiotemporal model fit to actual
concentration data

800

6500

400
41.2

200
41.4

6000

100
41.6
42
.2

75
50

5500

41.8

25
10

5000

42

5
0
8500

9000

9500

10000

10500

o Processed data file and panel settings can

be saved, thereby avoiding the need to
reprocess the data.
o Monitoring wells can be attributed to
different aquifers, so that multiple models
can be created from the same input data.

o Graphics can be exported directly to MS

Powerpoint or to other applications for
reporting purposes.

10^4

10^3

10^2

10^1

2006

2007

2008

2007

Non-Detect Data
2008

MW9

10^0

o Actual vs modelled concentration data

can also be exported to Excel for further
statistical analysis

551700

Detectable Data
2006

Solute concentration (ug/l)

7000

3000

551690

Spatiotemporal Predictions for Ethylbenzene at GWSDAT Example

Time series plots of measured vs modelled

solute concentrations

MW7
.4
40

41

8000

400
10

2009

Sampling Date

2009