Use of drifter technology for scalar transport

monitoring in estuaries and rivers

Authors Method and materials Results
Richard Brown, Yanming Feng, Kabir Suara and Charles Wang
Science and Engineering Faculty, Queensland University of Technology
Drifter hull and electronics Eddy diffusivity and
tidal inflow velocity
Helen Fairweather, Roy Sidle and Adrian McCallum
Environmental Engineering School, University of the Sunshine Coast
Michael Borgas, Marine and Atmospheric Research, CSIRO
Hubert Chanson, Civil Engineering, University of Queensland

Introduction
Lagrangian field data in tidal shallow waters are rare, but valuable for the under-
standing of the spatio-temporal structure flow and water qualities, validation and
calibration of hydrodynamic models, and advection-dispersion models for such
systems. Recent improvements in GPS technology have paved the way for the
development of instrumented high resolution Lagrangian drifters capable of
Validation of instruments
being deployed in shallow rivers and estuaries where processes of interest occur
at small temporal (100 seconds) and spatial (few metres) scales. Peak flood tide
Mid-flood tide
Slack water

Purpose
The aim of this project is to develop a novel Lagrangian drifter system - Real-
Time Flow Logging of Water (RTFLOW). This system is capable of monitoring river • The distribution of diffusivity with the water elevation measured by the
and estuarine water velocity and turbulence, dispersion coefficients, water param- drifters and averaged in time is shown above. The estimated diffusivity
eters (initially temperature and salinity. Future versions capable of incorporating a varied between 0.001 – 0.02 m2/s during a 4.5-hour experiment indicating
wide range of sensors including dissolved oxygen, nitrates, phosphates, pH, tur- that constant value of diffusivity used in Advection-diffusion modelling of
bidity, chlorophyll, etc.) and measuring air-water interface flux exchange. • Drifters were deployed repeatedly in a sheltered channel in the vicinity estuaries is untrue.
of acoustic Doppler velocimeters (ADVs), an acoustic Doppler current profiler
With about 50% of the global population living in coastal areas, human activ- (ADCP) and a sonic wind anemometer. • Peak diffusivity was observed at the early part of the flood which corre
ities around shallow water bodies (e.g. rivers, floodplains and estuaries) have sponds to the peak horizontal mean velocity. This suggests the diffusivity in
increased. These water bodies are exposed to pollution, climatic change, tidal • The response of the simple designs of high and low resolution drifters to the models can be scaled by the mean horizontal flow velocity.
and sub-tidal exchanges, which all influence their characteristics and thus pose wind and water flow in tidal shallow water was examined using qualitative
dangers to their ecosystems. In order to put effective safety measures in place, correlation and coherence analyses under moderate wind conditions • The small scale (eddy) diffusivity exhibited strong dependence (R2 > 0.9) on
monitoring of water bodies, estimates of mixing and dispersion of particles/con- (0 – 4 m/s). the horizontal velocity.
taminants , measurement of scalar concentration and flux exchange in Air-water
interface are important. • Subsurface drifter motions in bounded sheltered water are affected by wind
through low frequency induced wind current rather than direct wind drag
when only a small portion of the drifter is not submerged.

• The field validation of both high resolution and low resolution drifters with
Runoff from rainfall
through storm water surface measured velocity from ADCP is good (R2 > 0.9; RSME = 0.04 m/s) in
drain carrying the streamwise direction.
Fresh water runoff material into channel
from domestic usage

Material suspension
through bank erosion
Case studies
Fresh water input from Currimundi Lake:
Lake Kawana  Effect of mouth conditions on dynamics of Intermittently Open and Closed
Lakes and Lagoons (IOCLL)
 Mechanisms responsible for the dispersion and mixing
 Air-water interface flux exchange Our data
Sand bar accumulation Occasional boom in  Prediction of channel response to extreme weather conditions
Ground water run off
biting midges
Tidal exchange due to channel dynamics from recreational
and mouth blockage population Pumistone Passage:
parks
 Small scale (‘eddy’) diffusivity
 Assimilation of Lagrangian data into hydrodynamic models
Tidal in-flow while inlet  Air-water interface flux exchange: field observation and numerical modelling
is opened
Eprapah Creek: For comparison and to examine the similarity between the scaling of small scale
 Response of drifter to driving forces (water and wind velocities)
 Small scale (‘eddy’) diffusivity
mixing parameters in Eprapah Creek and existing bodies of theory, the plot
 Variation of eddy diffusivity with tidal flow of the apparent diffusivity against the length scale of diffusion are presented.
These are presented alongside Okubo’s dye experiments ocean diagram data.

Significance Okubo’s data were obtained from the dye tracer diffusion experiments covering
the time scale in the range of 2 hours - 1 month and length scale 30 m - 100 km.
Despite the difference in geometry, physics of the systems, approach and
This research will enable improved hydrodynamic models to be developed method of estimates, it is clearly observed that the diffusivities scale locally by
through better calibration and improved representation of their governing 4/3 Richardson power law.
processes. These improved models will provide more accurate predictions of
flood and tidal surge levels, pollutant transport, changes in water parameters Our data clearly extend Ocean dispersion diagram to smaller scale for small
due to spills or weather events, erosion and other environmentally important water bodies.
phenomena.

Scales of transport processes

Lengthscale→
Eprapah Creek Experiment Conclusion
Scale <1 cm 1 cm 1m 100 m 1 km 100 km Acoustic
↓Timescale • A new design for drifters with the capability of studying flow and air/water
Doppler
Transport ± 1 cm Velocimeters
quality parameters in tidal shallow waters is evolving.
Arrow ends indicate
caused by
theoretical limit of
the position tracking Mini-scale 10-2 s
molecular Energy ±1m • Validation of drifter measurement in shallow tidal water showed good cor
technologies diffusion, dissipation relation with acoustic measurements next to the surface.
evaporation ± 1 km
etc.
• Eddy diffusivities increase by two orders of magnitude in periods less than
Transport caused by 3D a tidal cycle as against constant values used in hydrodynamic models.
turbulence mixing; wind
Small-scale 102 s mixing, surface waves,
Opportunity • Eddy diffusivity showed strong dependence (R2> 0.9) on the mean
interaction with rough
bathymetry etc.
horizontal velocity.
Mesial-scale 104 s Transport caused by inertial and
resonance oscillation, diurnal and
• Eddy diffusivity scales with integral length to the power of 4/3 similar to
Meso-scale 106 s semidiurnal tidal currents, storms,
river and tributary current etc. Drifter trajectories from case Richardson’s. Therefore, “small scale” eddy diffusivity can be parameter
ised using the classic scaling.

study sites
Macro-scale 108 s Transport caused by seasonal and
Climatic- climatic variations, storms and
• “Small scale” eddy diffusivity from drifter measurements in shallow tidal
109 s water extends ocean diffusion diagram.
scale cross shelf exchanges

Timeline of technology development

Year Tracking Technology Applications Comment Source

References
error
1990s O[1 km] Acoustic Large ocean Still in use Ollitrault and
Rannou (2013)
1996 O[1 m] GPS Coast and Differential George and
estuary mode Largier (1996) Brown, R. J., & Chanson, H. (2012). Suspended sediment properties and suspended sediment flux estimates in an inundat-
ed urban environment during a major flood event. Water Resources Research, 48(11). doi:10.1029/2012WR012381
2003 O[1 m] GPS Surf zone & Non-differential Johnson et.al.
200 m 500 m
lakes mode (2003) George, R., & Largier, J. L. (1996). Description and performance of finescale drifters for coastal and estuarine studies.
Journal of Atmospheric and Oceanic Technology, 13(6), 1322-1326. doi:10.1175/1520-0426(1996)013<1322:DAPOFD>2.0.
2003 O[1 cm] GPS Surf zone Differential Schmidt et al. CO;2
mode (2003)
Johnson, D., Stocker, R., Head, R., Imberger, J., & Pattiaratchi, C. (2003). A compact, low-cost GPS drifter for use in
Now O[1 cm] GPS Estuary RTK in RTFLow project

Data analysis
the oceanic nearshore zone, lakes, and estuaries. Journal of atmospheric and oceanic technology, 20(12), 1880-1884.
differential mode doi:10.1175/1520-0426(2003)020<1880:ACLGDF>2.0.CO;2

Okubo, A. (1971). Oceanic diffusion diagrams. Deep Sea Research and Oceanographic Abstracts, 18(8), 789-802.
doi:10.1016/0011-7471(71)90046-5

Conceptual model vi ( t ) = Vi ( t ) − Vi ( t )
Ollitrault, M., & Rannou, J.-P. (2013). ANDRO: An Argo-based deep displacement dataset. Journal of Atmospheric and
Oceanic Technology, 30(4), 759-788. doi:10.1175/JTECH-D-12-00073.1

Ozcoast (2016), Australian Online Coastal Information, http://www.ozcoasts.gov.au/conceptual_mods/index.jsp

Schmidt, W., Woodward, B., Millikan, K., Guza, R., Raubenheimer, B., & Elgar, S. (2003). A GPS-Tracked Surf Zone Drifter.
1 t + ∆T
∆T ∫t
Vi (t ) = Vi (t )dt Journal of Atmospheric and Oceanic Technology, 20(7), 1069-1075. doi:10.1175/1460.1

Suara, K. A., Brown, R. J., & Borgas, M. (2016). Eddy diffusivity: a single dispersion analysis of high resolution drifters in a
1. Drifter design and calibration 1 T 2
∫
2 tidal shallow estuary. Environmental Fluid Mechanics, 1-21. doi:10.1007/s10652-016-9458-z
vi = vi ( t )dt
L T 0 Suara, K. A., Brown, R. J., Wang, C., Borgas, M., & Feng, Y. (2015). Estimate of Lagrangian integral scales in shallow tidal
Design 2. Integration of water and air quality sensors T water using high resolution GPS-tracked drifters. In 36th IAHR World Congress.

RL (τ) =
∫
0
( vi ( t ) vi ( t + τ))dt
Suara, K.A., Ketterer, T., Fairweather, H., McCallum, A., Vanaki, Sh. M., Allan, C., & Brown, R. (2016) Cluster dispersion of
3. Deployment of the real time data acquisition, v i2 low Cost GPS-Tracked Drifters in a Shallow Water. In 10th Australasian Heat & Mass Transfer Conference, 14 - 15th, July
L
processing and management 2016, Brisbane.
∞
TL = ∫ RL (τ ) dτ Suara, K. A., Wang, C., Feng, Y., Brown, R. J., Chanson, H., & Borgas, M. (2015). High Resolution GNSS-Tracked Drift-
er for Studying Surface Dispersion in Shallow Water. Journal of Atmospheric and Oceanic Technology, 32(3), 579-590.
0

( )
L L = vi2
L
1/ 2
* TL
doi:10.1175/JTECH-D-14-00127.1

Taylor, G. I. (1954). The Dispersion of Matter in Turbulent Flow through a Pipe. Proceedings of the Royal Society of London
∞ A: Mathematical, Physical and Engineering Sciences, 223(1155), 446-468. doi:10.1098/rspa.1954.0130

4. Field deployment
K=2 vi2
L ∫ 0
R L (τ) dτ

Application
5. Assimilation of Lagrangian data into Eulerian-
based models
vi = residual velocity, V = instantaneous velocity from the drifter, over bar signifies
mean, t = time, Subscript ‘i’ = Streamwise and Across stream directions, Acknowledgements
= timelag, RL = autocorrelation function, TL = Lagrangian integral time scale,
The project is supported through Australia Research Council Linkage grant
6. Hydrodynamic and transport modelling LL = Lagrangian integral length scale, K = eddy diffusivity
_ LP150101172 and Sunshine Coast Council.
Variable interval time averaging (VITA) for Vi(t); window size, ∆T= 200 s every 1 s.
Thanks to Queensland Department of Natural Resources and Mines, Australia
This equation suggests asymptotic behaviour of K(t) after decorrelation time. In for providing access to SunPOZ network for reference station data used for
Management of estuarine systems practice, this behaviour may or may not occur. Time of first zero crossing was RTK post processing of the high resolution GPS-tracked drifter.
adopted for upper integral limit.