Comparison of Longshore Currents with NGOFS
Predictions
Dell, Larry, Picarazzi, Gabriel, Picarazzi, Francesca and Tissot, Philippe
Texas A&M University-Corpus Christi
Conrad Blucher Institute
Introduction/Background
How accurate is NGOFS for nearshore predictions in the Coastal Bend?
•
Can NGOFS predictions be used instead of actual measurements?
•
If the model is accurate what can we learn from NGOFS predictions?
Description of NGOFS
•
•
The NGOFS model (Wei et al. 2015) was built using the Finite-Volume Coastal Ocean
Model (FVCOM), an unstructured three-dimensional primitive equation based
hydrodynamic model. The NGOFS grid is presented in Figure 1.
Resolution ranges from 10 km in the open ocean to 600 meters close to shore.
Predictions include water levels, water currents, water temperatures, water salinity, and
wind velocity.
RMSE = 0.0856
Count
Count
RMSE = 0.1377
50
0
-1
-0.5
0
0.5
1
Difference (m/s)
Figure 4: Picture of current profiler
at top of I-Beam with winch box.
Figure 3: Sample of NGOFS Grid with
location of bob hall pier.
Time Series Comparing NGOFS Model and Observed Data
1
100
50
0
-0.5
0
0.5
1
Difference (m)
Figure 7: Assessment of NGOFS longshore current and water level
predictions for Bob Hall Pier.
NGOFS Longshore Current Profile
40
35
NGOFS Model
Observed Current
Sigma Layer
•
150
100
Along Shore Current (m/s)
Accurate nearshore measurements and predictions are important for oil spill responders,
development of infrastructure along the coast, rip current warnings, and overall understanding
of coastal processes. Two current profilers and wave sensors were installed on Bob Hall Pier
at the end of Spring 2014 providing such measurements. During the same time frame the
National Oceanic and Atmospheric Administration has made available predictions from its new
Northern Gulf of Mexico Operational Forecast System or NGOFS. This new hydrodynamic
model computes predictions for multiple data sets including water levels, currents, water
temperature and salinity. NGOFS resolution is high and includes the full coast of Texas making
it a promising tool. The goal of this project is to compare predictions and measurements and
answer the following questions:
•
Water Level Difference
(Predicted-Observed)
Longshore Current Difference
(Predicted - Observed)
0.5
0
-0.5
30
25
20
15
10
5
0
-1
6/1/2014
8/1/2014
10/1/2014
12/1/2014
10
20
30
40
50
60
70
80
90
100
Percentage of Maximum Current for Each Time Stamp
2/1/2015
Figure 8: NGOFS model for vertical current profile.
Figure 5: Comparison of measured and NGOFS predicted currents for
Bob Hall Pier Texas (NGOFS values for the element including BHP).
Study Site: Bob Hall Pier, Texas
Longshore Current Comparison
Comparison of the differences between measurements and model predictions were computed
based on the following:
• A data base (Cassandra) was designed and populated with NGOFS nowcasts for the Bob
Hall Pier location.
• Time series of longshore current measurements from a SonTek Argonaut SL-500 current
profiler for a distance of 54 to 65m away from the sensor (bin 5). Bin 5 is the furthest stable
current bin for this experiment (Tissot et al. 2015)
• Average NGOFS longshore current predictions (computed based on orientation of the
coastline and northward and eastward predicted currents) for sigma layers 8 through 33 to
reflect a vertical profiler beam spread of approximatively 4.3m at that distance
• Time series of water level measurements at the NWLON station referenced to mean sea
level (msl datum) and NGOFS water level predictions referenced to the same datum.
NWLON
Stilling Well
NWLON Wind
Sensors
Predicted Current (m/s)
Figure 1: NGOFS Grid.
Methods
Discussion/Conclusions
1
•
0.8
•
0.6
0.4
R-Squared = 0.725
RMSE = 0.0515
0.2
y = 0.4*x + 0.0016
•
•
0
-0.2
All data
Linear Correction Model
Wind Over 20 mph
y=x line
-0.4
-0.6
•
•
•
-0.8
-1
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
Observed Current (m/s)
Figure 6: Assessment of NGOFS longshore current predictions with
study measurements.
Current Profilers
Winch Boxes
•
Water Level Comparison
1
0.8
Predicted (m)
Solar Panels
0.6
R-Squared = 0.834
RMSE = 0.0683
0.4
y = 0.75*x + 0.03
References
• Wei, E., Z. Yang, Y. Chen, J.G.W. Kelley, A. Zhang (2014) The Northern Gulf Of Mexico Operational Forecast
System (NGOFS): Model Development and Skill Assessment. NOAA Technical Report NOS CS 33,
190pp.[Available online at
http://www.nauticalcharts.noaa.gov/csdl/publications/TR_NOS_CS33_FY_14_02_Eugene_NGOFS_report.pdf.]
• ​Tissot, P., L. Dell, J. Rizzo, D. Williams (2015) Nearshore Measurements of Wave Climate and Current Profiles
along the Texas Coastal Bend. Proceedings of the 13th Symposium on the Coastal Environment, AMS 95th
Annual Meeting. [Available online at https://ams.confex.com/ams/95Annual/webprogram/Paper269560.html].
0.2
0
-0.2
Data
Linear Correction Model
y=x Line
-0.4
-0.6
Acknowledgements
-0.8
Offshore Looking Current
Profiler I-Beam
•
NGOFS predictions, longshore currents and water levels were compared with
measurements at Bob Hall Pier, Texas for the period June 1st, 2014 to March 6th, 2015.
Longshore current near BHP is under-predicted by NGOFS but at a consistent rate. A
linear correction to the output could be applied, yielding more accurate predictions with a
resulting RMSE improvement from 0.14 m/s to 0.05 m/s.
Instances of high wind events do not affect the accuracy of the predictions. Current
during these events tend to stay near the same trend line for all the data.
Agreement between water level predictions and measurements is closer. The predicted
values are highly linearly correlated to the observed values. A smaller correction would
improve the RMSE from 0.09 m to 0.07 m.
The difference between predictions and measurements is partly due to comparing point
measurements to average predictions over a grid cell.
The statistically corrected model could be useful for coastal emergency responders, and
coastal stakeholders in general.
Future work will include assessing the accuracy over longer-term predictions.
The performance of an experimental NOAA model, the Nearshore Wave Prediction
System (NWPS), is in the process of being analyzed. Water level and significant wave
height predictions are compared to measured values at Bob Hall Pier. That model has a
considerably lower (square) grid resolution of about 3.76 square miles likely representing
a challenge for nearshore accuracy.
This study will continue at least through July 2015 and will provide a more complete data
set for assessments of these models and overall description of the nearshore processes
at this important location. Accuracy low winds vs high winds.
Across-shore Looking
Current Profiler I-Beam
Figure 2: Overall experimental setup at Bob Hall Pier including two current
profilers collocated with the Corpus Christi NWLON station.
-1
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
Observed (m)
Figure 7: Assessment of NGOFS water level predictions with NWLON
measurements.
1
Funding for the work presented in this paper is provided as part of a R&D 2014-2015 Funding
Cycle grant from the Texas General Land Office (TGLO) and is gratefully acknowledged. The
views expressed herein are those of the authors and do not necessarily reflect the views of
TGLO.
The authors wish to acknowledge members of the Conrad Blucher Institute field crew and
information technology teams for efficiently and creatively resolving problems throughout
the project.ng observed for water currents.