Analysis of Mean Velocity and Turbulence Measurements With Adcps

See
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/268689707
Analysis of mean velocity and turbulence

measurements with ADCPs
ARTICLE in ADVANCES IN WATER RESOURCES MAY 2015
Impact Factor: 3.42 DOI: 10.1016/j.advwatres.2014.11.006
CITATIONS
READS
189
2 AUTHORS:
Francesca De Serio
Michele Mossa
Technical University of Bari, Italy
Politecnico di Bari
49 PUBLICATIONS 129 CITATIONS
202 PUBLICATIONS 403 CITATIONS
SEE PROFILE
All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.
SEE PROFILE
Available from: Francesca De Serio

Retrieved on: 07 January 2016
Advances in Water Resources xxx (2014) xxxxxx
Contents lists available at ScienceDirect
Advances in Water Resources

journal homepage: www.elsevier.com/locate/advwatres
Analysis of mean velocity and turbulence measurements with ADCPs

Francesca De Serio , Michele Mossa
Department of Civil, Environmental, Building Engineering and Chemistry DICATECh, Technical University of Bari, Via E. Orabona 4, 70125 Bari, Italy
a r t i c l e
i n f o
Article history:
Available online xxxx
Keywords:
Log law velocity prole
Variance method
Reynolds stress
Turbulent kinetic energy
Bottom drag coefcient
Wind drag coefcient
a b s t r a c t
The present study examines the vertical structure of the coastal current in the inner part of the Gulf of
Taranto, located in the Ionian Sea (Southern Italy), including both the Mar Grande and Mar Piccolo basins.
To this aim, different measuring stations investigated by both a Vessel Mounted Acoustic Doppler Current
Proler (VM-ADCP) and a bottom xed ADCP were taken into consideration. Two surveys were carried
out in the target area on 29.12.2006 and on 11.06.2007 by the research unit of the Technical University
of Bari (DICATECh Department), using a VM-ADCP to acquire the three velocity components along the
water column in selected stationing points. The measurements were taken in shallow waters, under
non-breaking wave conditions, offshore the surf zone. Due to the recording frequency of the instrument
time-averaged vertical velocity proles could be evaluated in these measuring stations. Water temperature and salinity were also measured at the same time and locations by means of a CTD recorder. A rigidly
mounted ADCP, located on the seabed in the North-Eastern area of the Mar Grande basin, provided current data relative to the period 1020 February 2014. Set to acquire the three velocity components with
higher frequency with respect to the VM-ADCP, it allowed us to estimate the turbulent quantities such as
Reynolds stresses and turbulent kinetic energy by means of the variance method.
Therefore, the present research is made up of two parts. The rst part examines the current pattern
measured by the VM-ADCP and veries that, for each station, the classical log law reproduces well the
vertical prole of the experimental streamwise velocities extending beyond its typical limit of validity
up to the surface i.e. reaching great heights above the sea bed. This behavior is quite new and not always
to be expected, being generally limited to boundary layers. It has been convincingly observed in only few
limited experimental works. In the present study this occurred when two conditions were met: (i) the
ow was mainly unidirectional along the vertical; (ii) the interested layer was non-stratied.
The second part of the research studies the turbulent statistics derived from the beam signals of the
xed ADCP by means of the variance method. This technique had the advantage of being able to measure
the time evolution of the turbulent mixing throughout the entire water column, thus making it possible
to perform a detailed study on momentum transfer and turbulence. The deduced vertical proles of the
Reynolds stresses and of the turbulent kinetic energy TKE showed an increasing trend toward the surface,
in agreement with previous results in literature.
New data-sets of mean velocities and shear stresses, coming from eld measurements, are always
needed. In fact they represent the rst step to derive reliable reference values of coefcients and parameters for the implementation and calibration of the used mathematical hydrodynamic models.
Consequently, an effort was made to evaluate consistent bottom drag and wind drag coefcients, on the
basis of the calculated bottom and surface shear stresses, respectively.
2014 Elsevier Ltd. All rights reserved.
1. Introduction
The open North-Eastern part inside the Gulf of Taranto in the
Ionian Sea, along the Southern Italy coast (Fig. 1), is considered a
vulnerable and sensitive area, affected by massive chemical and
Corresponding author. Tel.: +39 080 5963557; fax: +39 080 5963414.
biological pollutant discharges due to the presence of heavy industry. The natural assimilative capacity of the sea, accomplished by
initial mixing and successive dispersion, as well as sediment transport, are phenomena strictly dependent on the magnitude and
directions of the current [1]. Hence, a knowledge of the coastal current pattern is desirable, being a useful support for the local
authorities in the planning and management of the coastal area.
This purpose may be achieved with both eld measurements and
E-mail address: francesca.deserio@poliba.it (F. De Serio).

http://dx.doi.org/10.1016/j.advwatres.2014.11.006
0309-1708/ 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: De Serio F, Mossa M. Analysis of mean velocity and turbulence measurements with ADCPs. Adv Water Resour (2014),
F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx
numerical model results. Collecting a large amount of data in widespread areas is challenging because of technical and economic limitations. Numerical models represent a more rapid and a less
expensive solution, simulating the hydrodynamics of extended
areas with the desired level of accuracy and in a relatively short
time. Nevertheless, in order to be accurate, models need to be calibrated and successively validated by eld measurements [2,3].
These procedures often require as input condition the vertical laws
obeyed by the current velocity in order to derive parameters
involved in the model equations, i.e. the bottom roughness length
or the bottom drag coefcient.
For this reason, the rst aim of the present paper is to analyze
the eld velocity data measured by means of a VM-ADCP and to
reproduce the vertical velocity proles testing the applicability of
the classical log law. The data were acquired during two cruises,
carried out by the research group of the DICATECh Department
of the Technical University of Bari on 29.12.2006 and on
11.06.2007 respectively. It is worth pointing out that the log law
was deduced experimentally in simple pipe or channel congurations with uniform ows and it should represent accurately the
velocity prole in the inner region of free surface ows [4]. Experimentally, it was proven that this distribution may be extended to
the entire ow in some particular cases, referring to open-channel
ows, where the maximum velocity value is observed close to the
free surface [5] and also referring to tidal ows [6]. The use of the
log law in a complex marine scenario is not a foregone conclusion.
Moreover, in the examined case, the applicability of the log law far
beyond its typical limit of validity was proven, provided that the
ow was almost unidirectional and not stratied [7].
As a second aim, the present research intends to estimate the
vertical distributions of the turbulent quantities (using a single
measuring station, taking advantage of the rapid sampling of a
xed ADCP) and to compare them with trends in literature. In
the quest to measure turbulence parameters the use of ADCPs
has become common practice in recent years [810]. Also a validation by comparing ADCP Reynolds stresses with estimates from
other instrumentation has been carried out [11]. More recently
Wiles et al. [12] used the structure function method to estimate
the TKE dissipation. In the present case, the variance method was
used to compute the turbulent shear stresses and the turbulent
kinetic energy TKE [1315]. The xed ADCP was installed in Mar
Grande basin in December 2013 as part of RITMARE project funded
by PON R&C 20072013.
The investigated area (Fig. 1), with composite topography and

exposed to urban and industrial discharges, includes the open
sea, the Mar Grande and the inner basin called the Mar Piccolo.
The Mar Grande covers an area of 35 km2 with an average depth
of about 15 m and connects with the open Ionian Sea through
two openings. The Mar Piccolo has a surface area of about
21 km2 and is structured in two embayments, with an average
depth of about 12 m. The area where measurements were carried
out is westward of the Mar Grande, in the open sea, with depths
in the range of [15 m to 30 m].
It is worth pointing out the importance of this kind of study,
because accurate estimates of mean velocities and shear stresses
are essential for the study of ow structures [14]. Moreover, laboratory experiments may represent a stringent test of the ability of
ADCPs to measure mean velocities and turbulence. Because length
and time scales of the turbulence in a laboratory ume are smaller
than those in a river or coastal area, ADCPs may resolve the turbulence better in the eld, even if ow inhomogeneity could affect
measurements. Therefore, it is worth doing further tests of ADCPs
on site [15].
The paper is structured as follows. Section 2 briey describes
the theoretical background for both the logarithmic law and the
variance technique. The survey equipment and procedure are illustrated in Section 3, together with the acquired velocity, temperature and salinity data. Section 4 illustrates and discusses the
applicability and validity of the log law, while Section 5 analyses
the turbulent behavior of the current in the xed station. Finally
in Section 6, the bottom stress coefcient and the wind coefcient
are estimated by means of friction velocity and shear stress.
2. Theoretical background
2.1. Log law velocity proles in turbulent ows
Two regions with distinct scalings characterize wall bounded
turbulent ows: the inner region, in which the viscous effect prevails, and the outer region, with prevailing ow inertia. On the
Fig. 2. The Janus ADCP conguration, from Dewey and Stringer [20], showing
relation of transducer beams to coordinates axes and angles.
Table 1
Main characteristics of the VM-ADCP system and of the CTD probe.
Fig. 1. Location of the target area where the survey was carried out. Source: Google
map.
Probe
Type
Value
VM-ADCP
Acoustic frequency
Velocity range
Velocity accuracy
600 kHz
10 m/s horizontal; 5 m/s vertical
1% of measured value 0.5 cm/s
CTD
Pressure range
Pressure accuracy
Temperature range
Temperature accuracy
07000 m
1
5 to 35 C
5
mainly unidirectional, the log law well reproduced the vertical distribution of the velocity also beyond the inner region, but failed
when stratication started to affect the water.
basis of the mean momentum equation, the velocity and length

scales in the inner region are taken to be respectively the shear
or friction velocity U and the viscous length m/U, being m the kinematic viscosity. In the outer region, the velocity scale continues to
be U, while the length scale is assumed to be the boundary layer
thickness d or a scale related to it. The limit between the inner
and outer region has been experimentally xed at z/h = 0.2 in the
case of open channels [4], z being the vertical distance from the
bottom and h the channel depth.
In literature [16,17] a debate is still open about the most suitable law, between the power law and the log law, to reproduce
the velocity proles in channels and boundary layer ows. With
the exception of few objectors, the log law is generally accepted
as the correct scaling for the inner region of a turbulent boundary
layer. Hence, in the inner region, on the basis of the dimensional
analysis and with the assumption of Prandtls mixing length, the
mean velocity prole U(z) can be expressed with inner variables as
Uz 1 U z
2:3
U z
ln
B
log
B
U
k
k
t
t
2.2. Variance technique

When current measurements are acquired by means of an
ADCP, an important limitation for turbulence measurement arises,
because the three velocity components are computed by combining velocities measured along acoustic beams oriented in different
directions and measured in distant uid particles. Therefore it is
necessary to evaluate all the statistical moments along each beam.
For this reason, in applying the variance method, it is assumed
that: (i) turbulence is horizontally homogeneous, i.e. turbulence
statistics are the same at all four beam locations; (ii) turbulent statistics are steady over the averaging interval [15,19].
The basic methodology of the variance method uses the alongbeam velocities from a four-beam ADCP in the Janus conguration
(as is the case, see Fig. 2) to compute proles of the vertical Reynolds stress in the two horizontal directions [14,9,20]. Theoretically, equations for the vertical stresses at a given depth can be
derived from the along-beam velocity equations from the two
opposing beam pairs.
The along-beam velocities, positive toward the ADCP head, are
respectively:
where k is the von-Karmans constant equal to 0.41 and B is a function of ow properties and wall roughness, derived from integration
[7]. As written, the log law should represent accurately the velocity
prole in the inner region of free surface ows, up to z/h 0.2. More
recently, Anwar [18], Lueck and Lu [6] and De Serio and Mossa [7]
stated that the log-layer height can reach many meters above the
bottom. De Serio and Mossa [7], analyzing some velocity eld data
recorded during marine cruises, observed that, when the ow was
b1 u sin h w cos h
b2 u sin h w cos h
4488000
6
4486000
5
4
4484000
latitude [m]
8
9
4482000
3
2
4480000
depth from
surface:
4478000
10
4m
6m
8m
10 m
12 m
14 m
16 m
18 m
velocity scale
0.1 m/s
4476000
2694000
2696000
2698000
2700000
2702000
2704000
2706000
longitude [m]
Fig. 3. Maps of the horizontal velocities measured at some selected depths during survey S1. Gauss Boaga reference system used.
Table 2
Coordinates (Gauss Boaga reference system) and depths of the investigated stations.
Station
Latitude (m)
Longitude (m)
Investigated depth h (m)
2
3
4
5
6
7
8
9
10
A
B
C
D
E
O
4480764.327
4481394.107
4483733.500
4484128.685
4486055.929
4485265.881
4483311.498
4482696.431
4479062.819
4484671.597
4486122.866
4483632.316
4483129.511
4481024.285
4481072.540
2705232.721
2701484.235
2700887.939
2697868.626
2698246.621
2695249.971
2696278.078
2698611.702
2699292.504
2697889.231
2698193.836
2696217.672
2698434.854
2698654.767
2707747.910
7
14
13
29
19
22
24
25
25
18
17
19
19
20
21
the component w02 or, alternatively, to make an assumption on the

anisotropy ratio [15]. Nevertheless, by taking the sum of the variances in Eqs. (8)(11) some terms related to TKE can be obtained.
For fully isotropic turbulence, the following equation is derived:

q2 3 02
02
02
02
b1 b2 b3 b4
8
2
14
which represents the lower bound for the TKE evaluated in the case
of anisotropic turbulence.
It is worth pointing out that the variance technique used in the
present work has the benet of measuring the time evolution of
the turbulent mixing throughout the entire water column, thus
allowing a detailed analysis of turbulence.
3. Equipment and methods used in the experiments
b3 v sin h w cos h
3.1. Surveys description
b4 v sin h w cos h
As mentioned above, during two cruises carried out by the

research group of the DICATECH Department of the Technical University of Bari on 29.12.2006 and on 11.06.2007, a Nortek AWAC
(Acoustic Wave And Current) VM-ADCP was used to measure the
sea three-current-velocity components. It was connected to a gyro
and a DGPS in order to take into account the vessel velocity and
thus to acquire the current velocity with respect to the seabed.
Moreover, the DGPS was used to locate the pre-determined stationing points. The main features of the AWAC current meter system are shown in Table 1. Its standard conguration has three
beams 120 apart, slanted at 25, and one vertical, whose opening
angle is 1.7.
During all the surveys, the measurements of the ow were
assessed with an acquisition frequency of 0.5 Hz, therefore only
where u, v, and w are the x, y, and z velocities in a right handed coordinate system (the earth system, in the present case) and h is the
angle of the beams away from vertical.
Combining equation (2)(5), the three velocity components are
derived:
8
b1 b2
>
>
< u 2 sin h
b4
v b23sin
h
>
>
:
b1 b2 b3 b4
w 4 cos h
All the velocities can be split into a mean part and a uctuating
part using the Reynolds decomposition, where the overbar denotes
the time averaged value and the prime symbol denotes the turbulent uctuation:
8
u u u0
>
>
>
< v v v0
>
w w w0
>
>
:
0
bi bi bi with i 1; 2; 3; 4
Consequently, the along-beam velocity variances can be

expressed as:
02
02
02
10
02
11
b1 u02 sin h 2 u0 w0 sin h cos h w02 cos2 h

b2 u02 sin h 2 u0 w0 sin h cos h w02 cos2 h
b3 v 02 sin h 2 v 0 w0 sin h cos h w02 cos2 h
b4 v 02 sin h 2 v 0 w0 sin h cos h w02 cos2 h
The difference between Eqs. (8) and (9) provides the component
u0 w0 of the Reynolds stress, as well as the differences between Eqs.
(10) and (11) gives the component v 0 w0 :
02
02
sxz
b1 b2
u0 w0
q
4 sin h cos h
12
02
02
syz
b3 b4
v 0 w0
q
4 sin h cos h
13
where q is the water density.

The
turbulent
kinetic
energy
TKE,
dened
as

.
2
02
02
02
q =2 u v w
2, cannot be measured by a four beam
ADCP, because it is necessary to have an independent estimate of
Fig. 4. Examples of time moving average of the measured eastern u, northern v and
vertical w velocity components (top) and of their variances (bottom) at station 5, at
6 m from the surface.
0
0.9
0.9
0
0.8
0.8
0
0.7
0.7
0
0.6
0.6
0
0.5
0.5
z/h
z/h
0
0.4
0.4
st
st
st
st
st
st
st
st
st
0
0.3
0
0.2
0
0.1
2
3
4
5
6
7
8
9
10
0
st 2
st 3
st 4
st 5
st 6
st 7
st 8
st 9
st 1 0
0.3
0.2
0.1
0
1
12
13
14
15
16
17
T [C]
18
19
20
3
37.5
38
S [psu]
38.5
39
Fig. 5. Vertical proles of measured temperature T and salinity S in the investigated stations during survey S1.
4488000
B
4486000
A
4484000
latitude [m]
C
D
4482000
4480000
4478000
depth from
surface:
4m
6m
8m
10 m
12 m
14 m
16 m
velocity scale
0.1 m/s
4476000
2692000
2696000
2700000
2704000
2708000
longitude [m]
Fig. 6. Maps of the horizontal velocities measured at some selected depths during survey S2. Gauss Boaga reference system used.
the mean velocities were deduced and are examined in the present
paper. The measurements were acquired starting from 4 m below
the water surface, as it was the blanking distance of the instrument. The bin size and number depended on the investigated local
depths and are specied in the following subsections. The measurements of all the surveys were assessed by anchoring the boat
and acquiring the velocities for a total time equal to 10 min on

average and not less than 5 min for each investigated station. This
duration was chosen following the sampling times used in
literature in similar cases [10,12]. Moreover, it was also taken into
account that this time period had to be both sufciently short
to consider the ow steady and sufciently long to satisfy the

1
0.9
0.9
0.8
0.8
0.7
0.7
0.6
0.6
0.5
0.5
z/h
z/h
0.4
0.4
0.3
0.3
st
st
st
st
st
0.2
0.1
A
B
C
D
E
st A
st B
st C
st D
st E
0.2
0.1
0
18
19
20
21
22
23
24
25
26
37
37.5
T [C]
38
38.5
39
S [psu]
Fig. 7. Vertical proles of measured temperature T and salinity S in the investigated stations during survey S2.
Fig. 8. Location of the xed ADCP in Mar Grande basin. Google Earth source.
well-known acquisition procedure of an ergodic system. A CTD

recorder system by Idronaut Srl (Table 1) was used to measure
the water temperature and salinity along the water column during
the same time interval. More detailed characteristics of the instruments used are described in Mossa [1].
3.1.1. Survey of 29.12.2006 S1
From 9.30 a.m. till 13.30 p.m. (local time), for each investigated
station the current measurements were collected along the water
column with a vertical resolution of 2 m. Stations 3, 4, 6 and 7
(Fig. 3) were located along the coast following the bathymetric line
of 20 m; stations 5, 8, 9 and 10 were in the open sea, approaching
the 30 m bathymetric line; station 2 was inside the Mar Grande
basin. Their geographical coordinates and depths are summed up
in Table 2. Fig. 3 plots a map of the measured horizontal velocity
for depths in the range 4 m to 18 m starting from the sea sur-
Table 3
Main characteristics of the xed ADCP.
ADCP conguration
Model
Frequency
Beam angle h
Transducer
Bin size
Blank
Sample interval
Heading
Tilt 1 (pitch)
Tilt 2 (roll)
Velocity accuracy
Velocity resolution
Velocity range
Workhorse monitor ADCP

600 kHz
20 relative to vertical
4 beam convex
0.5 m
1.6 m
0.5 s
252
0 (beam 34 axis)
0 (beam 12 axis)
0.3% of water velocity 0.3 cm/s
0.1 cm/s
5 m/s
Fig. 9. Vertical proles of the measured streamwise velocity in the investigated stations during survey S1, with error bars.
Fig. 10. Vertical proles of the measured streamwise velocity in the investigated stations during survey S2, with error bars.
face, in order to show the principal pattern of the current. The

shape of the gulf seems to constrain the sea current to follow the
shoreline particularly in stations 7, 6, 4, 3 and 10. Thus, the coastal
ow has a mainly anticyclonic trend and the current bends toward
east and southeast. The vertical velocity component can be disregarded when compared to the horizontal one, for each examined
stationing points. In Fig. 4, referring to one of the investigated stations, the time moving average and the variance of the acquired
velocity signals are plotted for each velocity component, showing
that the acquisition time was sufcient enough to have a stationary

measurement, as already written. During the survey time, the wind
data showed an average intensity equal to 4 m/s and an average
direction of N310E.
The vertical proles of temperature T and salinity S measured by
the CTD recorder system are shown in Fig. 5. In all the examined
stations, the temperature values are almost invariant along the vertical and stabilize around 15 C, with the exception of stations 2 and
3, which are the nearest to the Mar Grande and are characterized by
Fig. 11. Measured streamwise velocity distributions tted by the log law, with error bars.
T = 14 C. In stations 6 and 7 a decreasing trend of T affects the

supercial layer, where a small reduction of temperature of 2%
occurs. The vertical homogeneous trend is also evident for salinity,
whose value is 38.10 psu on average for stations 2 and 3 and
38.35 psu on average for all the other stations. Only in stations 4
and 7, which are more affected by industrial and urban discharges,
the salinity respectively increases and decreases from the
surface with increasing depth, but with a rate of 3%. Therefore, in
this survey the ow is not inuenced by stratication along the
whole depth, also considering that the value of the calculated
Richardson number is less than 1, thus assuring the absence of

stratication.
3.1.2. Survey of 11.06.2007 S2

Field data were collected on 11 June 2007 from 10.00 a.m. till
2.00 p.m. (local time), in ve stations, with a vertical resolution
of 1 m. The current speeds ranged from about 0.02 m/s to about
0.10 m/s and the wind data showed an average intensity equal to
3 m/s and an average direction of N200E.

Table 4
Calculated parameters of the log law.
Station
u (m/s)
R2
5
6
7
8
10
A
B
C
D
E
0.018
0.028
0.012
0.013
0.024
0.032
0.018
0.025
0.018
0.019
26.787
25.925
16.844
23.473
25.409
30.775
27.847
28.503
26.474
22.107
0.76
0.87
0.73
0.73
0.73
0.76
0.85
0.71
0.90
0.79
It was a four-beam ADCP, looking upward in a Janus conguration

with an angle relative to the vertical equal to h = 20, while the
beam width was 15 (Fig. 2). It was rigidly mounted, so that the tilt
angles were constantly equal to 0 during all the investigated period. Velocities were sampled along the water column with 0.50 m
vertical bin resolution and a 1.60 m blanking distance. The ADCP
instrument was operated in a standard dual proling and waveburst sampling mode. Mean current velocity proles were collected continuously at 1 h intervals, using an average of 60 measurements acquired every 10 s. Instead, wave bursts of 2400
measurements at 0.5 s intervals (i.e. a 20-min-long record with a
frequency of 2 Hz) were collected each hour. The wave bursts sampled ve selected depth cells for all four beams. These selected
depth bins were distributed over the water column to examine
the data variation along the vertical. Burst samples were collected
at bins 1, 18, 36, 37 and 38 (counted starting from the transducer),
corresponding to heights above the seaoor of 1 m, 9.5 m, 18.5 m,
19 m and 19.5 m, respectively. The ADCP transducer head was
located approximately 0.50 m above the seabed. The ADCP conguration and setup are summarized in Table 3. The analysis carried
out with the eld data measured by the rigidly mounted ADCP
focused on the estimates of the turbulent statistics of the ow (typically mean, root mean square and correlations of the velocity uctuations, whose calculations are possible thanks to the data
recording frequency). In fact, the acquisition frequency of 2 Hz,
which seems low in itself, is absolutely comparable with the
instrument acquisition frequency used by previous researchers in
other similar investigations [14,15,19,21].
4. Streamwise velocity proles
Fig. 12. Experimental B as function of Ustr/U, with error bars. The tting line by De
Serio and Mossa [7] (named DSM 2014) is shown, together with the new tting line
derived also for data of the present experiment.
Station B was the nearest to the coast, while stations A, C, D and

E are further from the coastline (Fig. 6). Their geographical coordinates and depths are summed up in Table 2. In Fig. 6 the map of the
measured horizontal velocity is illustrated for some selected
depths in the range 4 m to 16 m starting from the sea surface.
On average a south-westward ow is observed in the target area,
so that the surface wind induced current seems to be strongly
inuenced by the topography.
Fig. 7 plots the vertical trends of measured T and S and highlights that up to z/h 0.7 a reduction of temperature of 5% occurs
from the surface towards the bottom, while S is almost constant
and equal on average to 38.24 psu. In the most supercial layer,
for z/h > 0.7, the reduction of temperature with increasing depth
reaches 7%, while an increase of salinity of 2% occurs. Since this
thermohaline variation is less than 10%, also in this survey the ow
was still taken as homogeneous. In fact, also in this case, the calculated Richardson number resulted as less than 1.
3.2. Measurements acquisition by the rigidly mounted ADCP

During December 2013, as part of the RITMARE project, with
funds from PON R&C 20072013 provided by the Italian Ministry
of University, an oceanographic station was installed in the Mar
Grande basin. Its coordinates and depth are written in Table 2 (station O) and its placement is shown in Fig. 8. This station was provided with many instruments, including an ADCP and an
anemometer by Teledyne RD.
The shallow ADCP was on bottom tripods, looking vertically
upward with the transducer head about 0.50 m above the seaoor.
In the present study, the log law was applied to the velocity
measurements acquired along the water column, starting from
the inner zone of the ow. The analysis of the vertical distributions
of the velocities assessed during surveys S1 and S2 was carried out
in the following way. For each measurement station the mean ow
direction along the vertical was detected, consequently the vertical
proles were evaluated for the U velocity component along this
direction, the so-called streamwise velocity. This procedure could
be applied observing that the dispersion of the velocity with depth,
with respect to the mean ow direction, was generally slight
(Figs. 3 and 6). Therefore, the ow was assumed as unidirectional
when the abovementioned dispersion was less than 15%, i.e. for
those stations where both the coast and the bathymetric lines provided a natural track for the current. The vertical distributions of
the streamwise velocity for these stations with unidirectional ow
are shown in Figs. 9 and 10, respectively for S1 and S2.
Equation (1) was applied to the abovementioned vertical proles, initially supposing that it may t the observed data only in
the inner region, near the sea bed. Actually, it was observed that
the best matching between experimental data and log law proles
(i.e. maxima values of the correlation coefcient R2) was obtained
when the log law was applied to the whole water depth where
measurements were assessed. This expected result conrmed
recent studies [14,7], showing that the log law reproduces the vertical trend of the streamwise current also outside the inner region
and up to the surface, if a stratication is absent, as seen in Figs. 5
and 7. In this way it was possible to estimate both the shear velocity U and the coefcient B of Eq. (1) from the tting line of each
graph plotting U as a function of log(z/m) (see Fig. 11). The deduced
values of U and B are summed in Table 4, together with R2. It is
worth noting that the B parameters are negative for all the vertical
streamwise proles, conrming the results of Anwar [18] and De
Serio and Mossa [7]. Particularly, following research by De Serio
and Mossa [7], also in the present study the linear relation between
B and the index number Ustr/U was veried. Ustr represented the
10
Fig. 13. Time series of Hs (top) and Tp (bottom) of the signicative wave during the observed period.
Fig. 14. Stick plot series of the measured wind during the analyzed period (1518.02.2014).
streamwise velocity at the distance from the bottom where the

stratication started to affect the water. Therefore, in the present
case, it was the velocity assessed in the most supercial ADCP cell,
having considered the water homogeneous along the whole depth.
Adding the values of B and Ustr/U derived for surveys S1 and S2 to
the data of De Serio and Mossa [7] and Anwar [18], the linear
behavior was proven again. It is worth noting that the coefcients
of the linear equation found in the present research differ slightly
(less than 4%) from those of De Serio and Mossa [7], as shown in
Fig. 12.
5. Analysis of the proles of the turbulent estimates

5.1. Preliminary treatment of the data
A time period of ten days, from 10 to 20 February 2014, was
taken into consideration for the analysis of the xed ADCP data.
This choice was suggested from the available anemometric data,
which guaranteed that in this period no strong wind episodes were
recorded. Hence also the wind-wave effect on the current could be
hypothesized as being slight.
Fig. 15. Amplitude spectrum of the measured turbulent velocity w0 in the most
supercial bin. Days 10.02.2014 and 15.02.2014 hour 00:02.
Referring to this time period, Fig. 13 shows the time series of

the hourly recorded signicative wave heights Hs (i.e. the average
of the highest 1/3 of the waves present) and peak periods Tp. The
xed ADCP was located at so-called intermediate depth, resulting
1/20 < h/La < 1/2, being La the Airy length of the wave with a period equal to the peak period Tp. Consequently, the inuence of the
waves along the water column could not be excluded a priori. For
this reason, a preliminary analysis was carried out to assure that
the turbulent velocities derived from the instrument did not contain the wave contribution.
11
It is evident (Fig. 13) that during the period from 15 to 18 February minimum values of Hs occurred, so that presumably in this
period the wave effect on the current could be considered small,
compared to previous and successive days. Also the stick plot of
the measured wind (Fig. 14) for this selected time interval showed
a low effect of the wind. In order to verify this hypothesis, the
amplitude spectra of the turbulent signals u0 , v0 and w0 were analyzed, for each day and for each hour, relatively to the three ADCP
bins nearest to the surface. Generally, for those cases in which the
value of Hs was relevant, the amplitude spectra of the turbulent
components showed a peak around the frequency corresponding
to the peak period.
On the contrary, when Hs values were small, the amplitude
spectra of the turbulent components had no peaks. As an example,
referring to the most supercial bin, Fig. 15 illustrates the amplitude spectrum of the turbulent velocity w0 for day 10.02 and for
day 17.02 at time 00:02. The spectrum of day 10.02 highlights a
peak around the frequency of 0.2 Hz, i.e. the inverse of 5 s-period,
which is the peak period in that day, at that time, as can be seen
from Fig. 13. This means that the wave contribution to the turbulent signals in this case cannot be disregarded. On the contrary,
in the spectrum of the day 17.02 no peaks appear, thus proving
that in this case the turbulent signals did not include the wave
contribution.
5.2. Estimates of turbulent velocities, shear stresses and TKE
Prior to analyzing the measured turbulent signals, it should be
remembered that the turbulence scales smaller than the measurement volume of the ADCP obviously cannot be resolved, but the
present study analyses the turbulent quantities such as Reynolds
stresses and turbulent kinetic energy, which depend primarily on
Fig. 16. Vertical proles of the Reynolds shear stress and of the TKE, for day 17.02.2014 at hour 00:02; 06:02; 12:02 and 18:02, with error bars.
12
Fig. 17. Vertical proles of the daily averaged horizontal velocity, Reynolds stress and TKE, for day 17.02.2014, with error bars.
the large scales of the ow [15]. Applying Eqs. (12) and (13), the
Reynolds stress components, apart from q, that is u0 w0 and v 0 w0 ,
were calculated for each hour in the examined period. Successively, their vector sum provided the hourly Reynolds stress acting
in the horizontal plane. These values are plotted as an example for
day 17.02 at hours 00:02, 06:02, 12:02 and 18:02 in Fig. 16. A variability with time in these vertical proles up to z/h = 0.8 is evident,
with a prevailing increasing trend of the absolute values for both
u0 w0 and v 0 w0 . Consequently, the Reynolds stress increases from

the bottom up to z/h = 0.8, while for z/h > 0.8 it decreases at hours
00:02 and at 18:00 and means a jet shape at hours 6:00 and 12:00.
Using Eq. (14) also the TKE was calculated and its vertical trend for
the same day is plotted in Fig. 16 every 6 h, showing an increasing
behavior from the bottom with increasing z. Daily-averaging over
the entire day 17.02 the Reynolds stress and the TKE, a resulting
increasing vertical trend towards the surface for both of them
13
sb qU 2
15
when a log-layer is identied. In any case, also the following quadratic form can be used, considering a reference velocity Ub
sb qC b U 2b
16
where Cb is the non-dimensional bottom drag coefcient and Ub is

usually chosen to be either the depth-averaged velocity or a near
bottom velocity.
Comparing equations (15) and (16), the values for Cb could be
derived, as
Cb
Fig. 21. Ratio of daily averaged Reynolds stress to 2TKE plotted against z/h for days
15.02, 16.02, 17.02 and 18.02 (open symbols) and for day 21.11.2010 and
25.11.2010 studied by De Serio and Mossa [7] (lled symbols).
was obtained, as shown in Fig. 17. In the same gure also the vertical distribution of the daily-averaged horizontal velocity is plotted, showing an increasing trend with increasing z and a jet
shape in the upper part of the water column.
The same approach was followed also for days 15.02, 16.02 and
18.02. Therefore, the daily averaged vertical proles of the Reynolds stress and the TKE were obtained, as illustrated in Figs. 1820, respectively. They conrm increasing values of both the Reynolds stress and of the TKE from the bottom towards the surface.
These ndings are in agreement with previous experimental
results by Nystrom et al. [15] and by De Serio et al. [22] in terms
of both order of magnitude and vertical behavior of the turbulent
estimates. Following Nezu and Nagakawa [4], the coefcient Rk,
which is the Reynolds stress normalized with 2TKE, is plotted
against z/h in Fig. 21. Also data from De Serio and Mossa [7] were
added to this graph. Rk increases from the bottom towards the surface and all the experimental data tend to collapse in a power law,
even if a larger scatter of data is present near the surface. This
trend is typical for the near-bottom and intermediate region of
open channel ows [4], while in their upper region the value of
Rk tends to reduce approaching the surface, as the turbulence
intensity is inuenced by the damping effect of free surface. On
the contrary, in the present study, this damping effect did not
occur due to the presence of the surface waves, therefore an
increasing trend of Rk was observed also in the uppermost layer.
6. Estimation of the bottom drag and wind drag coefcients
The bottom stress sb, i.e. the drag exerted on the ow by the
seabed, can be generally expressed as
U
Ub
2
17
Firstly, using Eq. (17), the Cb drag coefcient was evaluated for
each measuring station of both surveys S1 and S2, where the log
law vertical distribution was veried. The used values of U were
those reported in Table 4 and the Ub velocity was the streamwise
velocities measured in the point nearest to the bottom, for each
station.
Secondly, in correspondence of the rigidly mounted ADCP, Cb
was derived from Eq. (16) for all the examined hours of the investigated period 1518 February. It was deduced directly from the
Reynolds stress calculated in the bin nearest to the bottom. In this
case, the horizontal velocity measured in the bin nearest to the
bottom was used as the Ub comparing in Eq. (16).
Both for Cb values coming from surveys S1 and S2 and for Cb values calculated in the xed ADCP location, it is worth noting that
they are generally one order of magnitude greater than the typical
values in the range 24 103 reported in literature [23,24]. This
result was expected, as also previous researches [7,6,9,25] noted
that wide ranges of the drag coefcient are possible and that they
can be attributed to mechanisms like wavecurrent interactions or
turbulent ow conditions or effects of small scale topography (i.e.
bed forms). Lueck and Lu [6] stated that the drag coefcient
related to form drag can be 410 times larger than that related
to skin friction, on the basis of some previous continental shelf
measurements. Also Apotsos et al. [26] wrote that observations
and theoretical calculations had shown that the bottom drag coefcient can be large by an order of magnitude or more over rippled
sand beds, obtaining a value for Cb equal to 0.028 in the surf zone
from experimental data. The values of the bottom drag coefcients
Cb calculated at all the examined stations during surveys S1 and S2
by means of Eq. (17) are in the range 2060 103. Therefore, they
seem to overestimate the classical bottom drag coefcients, but
could be considered consistent in any case with the presence of
bed forms, e.g. dunes. Greater values are detected in correspon-
14
dence of the xed ADCP, where the average value of Cb reaches 0.2.
Taking into account the complexity of the bottom nature and texture in the investigate region, the results could be considered reliable, even if more experimental conrmation is needed, being Cb
affected by many different factors. For example, Apotsos et al.
[26], also following Barrantes and Madsen [27], wrote that the orientation of bedforms may inuence cross-shore and along-shore
ows differently and that smaller drag coefcients may be estimated in alongshore ows.
An estimate of the non-dimensional wind drag coefcient Cw
was also attempted, using an expression for the wind stress sw at
the sea surface analogous to Eq. (16):
sw qa C w u2w
18
where uw is taken as the wind velocity and qa the air density.

The Cw values were calculated only for the station O, taking into
account that in this position also wind measurements were
recorded (Fig. 14). An averaged wind drag coefcient Cw in the
range 670 103 was derived for wind intensities varying in
the range 0.56 m/s. These values overestimate typical literature
values of Cw which are on average around 0.001 [28]. In any case,
it is worth to note that they are referred to a wind velocity measured by the instrument at the sea surface and not at 10 m above
the sea level, i.e. the distance at which usually uw of Eq. (18) is
referred. Hence, hypothesizing that the wind intensity can be
increased of 1015% at 10 m above the sea surface, the corresponding values of Cw derived from Eq. (18) reduce by about one order of
magnitude and therefore better approximate the suggested values
present in literature.
7. Conclusions
The vertical structure of the coastal current in the inner part of
the Gulf of Taranto, in Southern Italy, is examined in the present
research. Both a vessel mounted ADCP and a bottom xed ADCP
were used to measure the three velocity components along the
water column in some selected stationing points. The time-averaged vertical proles of the streamwise current were analyzed
from the VM-ADCP data, taking into account its low recording frequency. It was proved that the classical log law well reproduced
the vertical prole of the experimental streamwise velocities up
to the surface, when the ow was mainly unidirectional along
the vertical and in absence of stratication. The velocities measured by the xed ADCP at higher frequency were analyzed by
the variance method and were used to derive the vertical proles
of the Reynolds stresses and turbulent kinetic energy TKE in this
location. These proles showed an increasing trend toward the surface, consistent with the presence of waves, thus conrming previous results in literature. Finally, starting from the experimental
Reynolds stresses derived near both the bed and the surface, the
estimates of the bottom drag coefcient Cb and of the wind drag
coefcient Cw were respectively attempted. The derived values of
Cb generally overestimate the classical values proposed in literature, but they could be consistent with the presence of bed forms.
The estimates of Cw are comparable with classical suggested values, particularly when the experimental wind velocity is slightly
increased to resemble values at 10 m above the sea surface.
References
[1] Mossa M. Field measurement and monitoring of wastewater discharge in sea
water. Estuarine Coastal Shelf Sci 2006;68:50914. http://dx.doi.org/10.1016/
j.ecss.2006.03.002.
[2] De Serio F, Malcangio D, Mossa M. Circulation in a Southern Italy coastal basin:

modelling and eld measurements. Cont Shelf Res 2007;27:77997. http://
dx.doi.org/10.1016/j.csr.2006.11.018.
[3] Bruno D, De Serio F, Mossa M. The FUNWAVE model application and its
validation using laboratory data. Coastal Eng 2009;56:77387. http://
dx.doi.org/10.1016/j.coastaleng.2009.02.001.
[4] Nezu I, Nakagawa W. Turbulence in open-channel ow. IAHR Monograph
Series, Balkema, Rotterdam, 1993.
[5] Termini D, Greco M. Computation of ow velocity in rough channels. J
Hydraulic
Res
2006;44(6):77784.
http://dx.doi.org/10.1080/
00221686.2006.9521728.
[6] Lueck RG, Lu Y. The logarithmic layer in a tidal channel. Cont Shelf Res
1997;17(14):1785801. http://dx.doi.org/10.1016/S0278-4343(97)00049-6.
[7] De Serio F, Mossa M. Streamwise velocity proles in coastal currents. Environ
Fluid Mech 2014. http://dx.doi.org/10.1007/s10652-014-9338-3.
[8] Rippeth TP, Williams E, Simpson JH. Reynolds stress and turbulent energy
production in a tidal channel. J Phys Oceanogr 2002;32:124251. http://
dx.doi.org/10.1175/1520-0485(2002)032<1242:RSATEP>2.0.CO;2.
[9] Stacey MT, Monismith SG, Burau JR. Measurement of Reynolds stress proles
in unstratied tidal ow. J Geophys Res 1999;104:1093349. http://dx.doi.org/
10.1029/1998JC900095.
[10] Souza AJ, Howarth MJ. Estimates of Reynolds stress in a highly energetic shelf
sea. Ocean Dyn 2005;55:4908. http://dx.doi.org/10.1007/s10236-005-00127.
[11] Howarth MJ, Souza AJ. Reynolds stress observations in continental shelf seas.
Deep
Sea
Res
II
2005;52:107586.
j.dsr2.2005.01.003.
[12] Wiles PJ, Rippeth TP, Simpson JH, Hendricks PJ. A novel technique for
measuring the rate of turbulent dissipation in the marine environment.
Geophys Res Lett 2006:33. http://dx.doi.org/10.1029/2006GL027050.
[13] van Haren H, Oakey N, Garrett C. Measurements of internal wave band eddy
uxes above a sloping bottom. J Mar Res 1994;52:90946. http://dx.doi.org/
10.1357/0022240943076876.
[14] Lu Y, Lueck RG. Using a broadband ADCP in a tidal channel. Part 2: turbulence. J
Atmos Oceanic Technol 1999;16:155667. http://dx.doi.org/10.1175/15200426(1999)016<1556:UABAIA>2.0.CO;2.
[15] Nystrom EA, Rehmann CR, Oberg KA. Evaluation of mean velocity and
turbulence
measurements
with
ADCPs.
J
Hydraulic
Eng
2007;133(12):13108.
http://dx.doi.org/10.1061/(ASCE)07339429(2007)133:12(1310).
[16] Cheng NS. Power-law index for velocity proles in open channel ows. Adv
Water
Res
2007;30:177584.
j.advwatres.2007.02.001.
[17] Afzal N, Seena A, Bushra A. Power law velocity prole in fully developed
turbulent pipe and channel ows. J Hydraulic Eng 2007;133(9):10806. http://
dx.doi.org/10.1061/(ASCE)0733-9429(2007)133:9(1080).
[18] Anwar HO. Velocity prole in shallow coastal water. J Hydraulic Eng
1996;122(4):2203.
http://dx.doi.org/10.1061/(ASCE)07339429(1996)122:4(220).
[19] Kirincich A, Rosman JH. A comparison of methods for estimating Reynolds
stress from ADCP measurements in wavy environments. J Atmos Oceanic
Technol 2011;28:153953. http://dx.doi.org/10.1175/2009JTECHO682.1.
[20] Dewey R, Stringer S. Higher Moment Flow Parameters from Various ADCP
Transducer Congurations. TRDI ADCPs Users Group Meeting: University of
Victoria; 2005.
[21] Burchard H, Craig PD, Gemmrich JR, van Haren H, Matheiu PP, Meier HEM,
Nimmo Smith WAM, Prandke H, Rippeth TP, Skyllingstad ED, Smyth WD,
Welsh DJS, Wijesekera HW. Observational and numerical modeling methods
for quantifying coastal ocean turbulence and mixing. Prog Oceanogr
2008;76:399442. http://dx.doi.org/10.1016/j.pocean.2007.09.005.
[22] De Serio F, Leuzzi G, Monti P, Mossa M. Prole measurements of turbulence
properties in coastal currents using acoustic Doppler methods, ISSF 2011
Congress, Rome, 2011.
[23] Feddersen F, Gallagher EL, Guza RT, Elgar S. The drag coefcient, bottom
roughness and wave-breaking in the nearshore. Coastal Eng 2003;48:18995.
http://dx.doi.org/10.1016/S0378-3839(03)00026-7.
[24] Feddersen F, Guza RT, Elgar S, Herbers THC. Longshore momentum balances in
the nearshore. J Geophys Res 1998;103:1566776. http://dx.doi.org/10.1029/
98JC01270.
[25] Gross TF, Nowell ARM. Spectral scaling in a tidal boundary layer. J Phys
Oceanogr
1985;15:496508.
http://dx.doi.org/10.1175/15200485(1985)015<0496:SSIATB>2.0.CO;2.
[26] Apotsos A, Raubenheimer B, Elgar S, Guza RT, Smith J. Effects of wave rollers
and bottom stress on wave setup. J Geophys Res 2007;112:C02003. http://
dx.doi.org/10.1029/2006JC003549.
[27] Barrantes AI, Madsen OS. Near-bottom ow and ow resistance for currents
obliquely incident to two-dimensional roughness elements. J Geophys Res
2000;105(C11):2625364. http://dx.doi.org/10.29/2000JC900132.
[28] Wu J. Wind-stress coefcients over sea surface from breeze to hurricane. J
Geophys
Res
1982;87(C12):97046.
JC087iC12p09704.

Analysis of Mean Velocity and Turbulence Measurements With Adcps

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analysis of Mean Velocity and Turbulence Measurements With Adcps

Uploaded by

Copyright:

Available Formats

See

Analysis of mean velocity and turbulence

Technical University of Bari, Italy

49 PUBLICATIONS 129 CITATIONS

202 PUBLICATIONS 403 CITATIONS

All in-text references underlined in blue are linked to publications on ResearchGate,

Available from: Francesca De Serio

Advances in Water Resources xxx (2014) xxxxxx

Contents lists available at ScienceDirect

Advances in Water Resources

Analysis of mean velocity and turbulence measurements with ADCPs

E-mail address: francesca.deserio@poliba.it (F. De Serio).

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

The investigated area (Fig. 1), with composite topography and

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

basis of the mean momentum equation, the velocity and length

2.2. Variance technique

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

Investigated depth h (m)

the component w02 or, alternatively, to make an assumption on the

3.1. Surveys description

As mentioned above, during two cruises carried out by the

Consequently, the along-beam velocity variances can be

b1 u02 sin h 2 u0 w0 sin h cos h w02 cos2 h

where q is the water density.

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

and acquiring the velocities for a total time equal to 10 min on

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

well-known acquisition procedure of an ergodic system. A CTD

Workhorse monitor ADCP

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

face, in order to show the principal pattern of the current. The

that the acquisition time was sufcient enough to have a stationary

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

T = 14 C. In stations 6 and 7 a decreasing trend of T affects the

Richardson number is less than 1, thus assuring the absence of

3.1.2. Survey of 11.06.2007 S2

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

It was a four-beam ADCP, looking upward in a Janus conguration

Station B was the nearest to the coast, while stations A, C, D and

3.2. Measurements acquisition by the rigidly mounted ADCP

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

streamwise velocity at the distance from the bottom where the

5. Analysis of the proles of the turbulent estimates

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

Referring to this time period, Fig. 13 shows the time series of

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

u0 w0 and v 0 w0 . Consequently, the Reynolds stress increases from

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

where Cb is the non-dimensional bottom drag coefcient and Ub is

F. De Serio, M. Mossa / Advances in Water Resources xxx (2014) xxxxxx

where uw is taken as the wind velocity and qa the air density.

[2] De Serio F, Malcangio D, Mossa M. Circulation in a Southern Italy coastal basin:

You might also like