A SYNERGETIC USE OF ACTIVE MICROWAVE OBSERVATIONS, OPTICAL IMAGES AND TOPOGRAPHY
DATA FOR IMPROVED FLOOD MAPPING IN THE GULF OF MEXICO AREA
Marouane Temimi1, Naira Chaouch1, Scott Hagen2, John Weishampel3, Stephen Medeiros2, Jesse Feyen4, Yuji Funakoshi4, Reza Khanbilvardi1
1NOAA- Cooperative Remote Sensing Science and Technology (CREST) Center,City University of New York, New York, NY
2Civil, Environmental and Construction Engineering Department, University of Central Florida, Orlando, FL
3Department of Biology, University of Central Florida, Orlando, FL4NOAA / National Ocean Service / Office of Coast Survey / Coastal Survey Development Lab
IGARSS 2011
Overland
Coastal Erosion
Tidal
Wave
Bay
Inundation
Sediment
Shorelines
Tides
Marsh, Oyster & SAV Assessments
Coastal DynamicAssessments
Management Actions
Integrated Models
Field/Lab Experiments
Salinity
Dynamic Results
Societal and Coastal Ecosystem Benefits
Earth Data
Global ClimateChange Scenarios
Management Tools
Biological Biotic
Moving Towards Spatial Storm Surge Model Validation
Apalachicola River
Turkey Point
Shell Point
Apalachee Bay
Cedar Key
Shark River
St. Andrew Bay
Panama City Beach
National Ocean Service Tidal Gaging Stations
Northeastern Gulf of Mexico Study Area
Station °W °N RMS (%) K1obs (m) K1sim (m) O1obs (m) O1sim (m) M2obs (m) M2sim (m)Apalachicola River –84.98167 29.72667 11.5 0.130 0.146 0.112 0.138 0.116 0.114Turkey Point –84.51167 29.91500 8.5 0.172 0.172 0.157 0.161 0.256 0.251Shell Point –84.29000 30.06000 7.9 0.183 0.189 0.162 0.177 0.362 0.362Apalachee Bay –84.17833 30.07833 7.8 0.173 0.190 0.153 0.177 0.356 0.365Cedar Key –83.03167 29.13500 15.1 0.177 0.181 0.163 0.170 0.386 0.361Shark River –82.72333 28.87000 12.7 0.136 0.180 0.150 0.166 0.378 0.373St. Andrew Bay –85.66667 30.15167 12.7 0.141 0.148 0.135 0.146 0.026 0.027Panama City Beach –85.87833 30.21333 10.5 0.145 0.150 0.141 0.148 0.034 0.029
(a) Apalachicola
(b) Turkey Point
(c) Shell Point
(d) Apalachee Bay
Comparison of Simulated and Measured Tidal Signals
Project Sub-Objective
To demonstrate the efficacy of employing high resolution imagery to improve coastal inundation models that are presently employed by NOAA (NWS and NOS), USACE, and FEMA, and those soon to be applied operationally.
Imagery will enable the assessment of wetting/drying algorithms and general spatial validation.
•Radar is sensitive to water, due to its high dielectric constant, and hence valuable in characterizing wetlands
• It differentiates between moist soil and standing water•Standing water interacts with the radar differently
depending on vegetation structure •When exposed to open water without (or submerged)
vegetation, specular reflection occurs.
Double – bounce backscattering Specular scattering
CO-register and re-sample to the same projection and pixel size
Radarsat 1 data LiDAR-derived DEMLandsat 7 image
(low tide)
Speckle filtering High contour line Low contour line
RGB colorcompositing
Flood-prone areasmask
Change detection within flood-prone areas
Flooded / non-flooded areas map
Validation with aerial photography
Acquisition date* 01/20/2003 09/17/2003 03/03/2004 07/25/2004
Water level (m) -0.064 0.24 -0.24 0.278
Wind speed (m/s) 1.1 4.3 1.1 2.3
Radarsat Imagery
*Acquisition time for all the Apalachicola scenes was 11h:40 GMT
Hours
0 2 4 6 8 10 12 14 16 18 20 22 24
Met
ers
Rel
ativ
e to
NA
DV
-0.6
-0.4
-0.2
0.0
0.2
0.4
20 Jan 200317 Sep 20033 Mar 200425 Jul 2004
Historic Observed Water Level (Apalachicola, FL)
(from NOAA Tides & Currents)
MHHW
-0.459
Low-water levelLandsat 7 scene (2/2/03; 15:56)
Radarsat Apalachicola Scene Dates andCorresponding Water Level
Radarsat Scene Color Composites 3/3/04 as low tide conditionRed = change to flooded
backscatter ↓Cyan = change from flooded
backscatter ↑White = unchanged pixels
Intertidal Zone Composites 3/3/04 as low tide conditionRed = change to flooded
backscatter ↓Cyan = change from flooded
backscatter ↑White = unchanged pixels
Franklin County - FLDOT
Comparison with Historic Aerial Photographs
Green – detected by SAR (3/3/04) & aerials
Yellow – detected only bySAR (3/3/04)
Blue – detected only by aerials
Site 101/04/10
12:12
Site 201/05/10
14:53
Agreement between SAR and Aerials
Probability of Detection (POD)
POD = A / (A +C)
A = number of pixels of class X (flooded) which were identified correctly as class X
C = number of pixels of class X which were notclassified as X
Site 1 Site 2
Water Level (m) -0.184 -0.222
POD (%) 58 83.8
Summary and Future DirectionsThe multi-temporal composited SAR images clearly show flooded and non-flooded areas during both high tide and low tide conditions. These results show potential for high resolution remotely sensed imagery to: monitor coastal flooding, delineate inundated areas, and improve hydrodynamic model verification/validation across a variety of coastal landscapes.We will: 1) evaluate model spatial flood predictions and guide improvements in the simulation of the wetting/ drying processes2) extend this approach temporally to include more dates and spatially across the northern Gulf of Mexico coast to include Alabama and Mississippi
NASA Applied Sciences Program
http://games.bio.ucf.edu
Acknowledgments
Support for this part of the project was provided by the NASA Program in Earth Science for Decision Making - Gulf of Mexico Region (Grant #NNX09AT44G) awarded to S. Hagen (PI-UCF).