Preprint/Prétirage RADARSAT OBSERVATIONS OF EXTREME WEATHER EVENTS J.A. Cranton, T.J. Pultz, and P.W. Vachon Canada Centre for Remote Sensing 588 Booth St., Ottawa, ON K1A 0Y7 ABSTRACT Recent evidence suggests that natural disasters such as hurricanes and floods are increasing in frequency. The Synthetic Aperture Radar (SAR) onboard the RADARSAT satellite is being used to monitor these phenomena. 1. RADARSAT Canada's RADARSAT, launched in 1995, carries a C-band, HH polarization SAR that can provide imagery during day or night, and for any weather conditions. RADARSAT has several modes (Table 1) that trade swath width for image resolution [e.g. http://radarsat.space.gc.ca]. The application requirements drive the choice of beam mode. Fine, Standard, and Wide modes provide smaller area coverage with higher resolution. These modes are often used for flood monitoring. ScanSAR provides larger area coverage with lower resolution. ScanSAR modes are often used for hurricane monitoring. Table 1: RADARSAT beam modes. Mode Nominal Swath [km] Nominal Resolution [m] Fine 50 10 Standard 100 25 Wide 130 30 ScanSAR Narrow 300 50 ScanSAR Wide 500 100 2. HURRICANES RADARSAT imagery can show a hurricane’s imprint in the ocean surface roughness; surface roughness is a function of wind speed. For the past two years, CCRS has acquired images of hurricanes to evaluate wind retrieval and to address hurricane morphology and propagation. In 1998, RADARSAT images were acquired of Hurricanes Bonnie, Danielle, Georges, and Mitch [Vachon et al., 1999]. Fig. 1 shows Hurricane Georges in the Gulf of Mexico on Sept. 26, 1998. Rain bands, a pre-hurricane squall line, and boundary rolls and gravity waves in the marine atmospheric boundary layer are visible in the image. The image was processed using the Ocean Monitoring Workstation (OMW) [http://www.satlantic.com/omw] wind retrieval module (Fig. 2). The OMW is an automated system that estimates wind speed and direction Figure 1: W1 image of Georges (�CSA 1998). Figure 2:OMW wind product for Georges. -84˚50' -84˚30' -84˚10' -83˚50' -83˚30' -83˚10' -82˚50' -82˚30' -82˚10' -81˚50' -81˚30' Longitude 22˚45' 23˚10' 23˚35' 24˚00' 24˚25' 24˚50' 25˚15' 25˚40' 26˚05' 26˚30' 26˚55' L at itu de Vent OMW Wind MPO SDMM / MEDS DFO Satellite: RSAT-1 Image Type: SGF Mode: Single Beam Beam / Faisceau: W1 Date: 11:35:49 26-09-1998 Filename: dt12225-01 Satlantic Contact: services@ottmed.meds.dfo.ca Production: 18:30:39 05-01-1999 A1 B1 C1 D1 E1 F1G1 H1 I1 J1 K1 L1M1 N1 O1 P1 Q1 R1S1 T1 U1 V1 W1 X1Y1 Z1 A2 B2 C2 D2E2 F2 G2 H2 I2 J2K2 L2 M2 N2 O2 P2Q2 R2 S2 T2 U2 V2W2 X2 Y2 Z2 A3 B3C3 D3 E3 F3 G3 H3I3 J3 K3 L3 M3 N3O3 P3 Q3 R3 S3 T3 from SAR imagery. The OMW wind field has a 180º direction ambiguity (thus, the opposing wind barbs) and shows the circulation around the storm’s eye. Following serendipitous observations in 1998, in collaboration with the Canadian Space Agency (CSA) and NOAA’s Atlantic Oceanographic Meteorological Laboratories (AOML), CCRS initiated a "Hurricane Watch" program to routinely acquire ScanSAR Wide (SCW) imagery during the 1999 Atlantic Basin hurricane season (August through October). CSA made special efforts to enhance RADARSAT scheduling, while AOML attempted to coordinate their P-3 hurricane penetration flights with the RADARSAT pass times. These flights provide the near surface wind field based on navigation data and on board scatterometers and microwave radiometers. Many RADARSAT images were acquired, including five that show the eyes of Hurricanes Cindy, Dennis and Floyd. Fig. 3 shows Floyd as a level four hurricane on Sept. 15, 1999. The eye appears darker due to its lower winds. Rain bands and squall lines are again apparent. Figure 3: SCW image of Floyd (� CSA 1999). 3. FLOOD MAPPING RADARSAT SAR images are excellent for monitoring floods since it is straightforward to differentiate between land and water [Cranton et al., 1999]. CCRS has used these data to monitor flood conditions caused by snow melt during the Red River flooding in 1996 and 1997, by extreme rain events such as in the Saguenay in 1996, and following hurricanes. Fig. 4 is an analysis of North Carolina after Floyd on Sept. 25, 1999 that shows drainage under normal conditions in dark blue, flood extent in light blue, roads in red, and political boundaries and coastlines in green. The flood extent was established with a straightforward image thresholding procedure. Table 2 shows reported river levels in North Carolina for Sept. 25, 1999; the measurements at the time of the RADARSAT acquisition are significantly above normal conditions. Figure 4: SCW of North Carolina (©CSA 1999). Table 2: North Carolina streamflow data following Hurricane Floyd. 4. SUMMARY and OUTLOOK Canada’s RADARSAT acquires SAR imagery that can be used to monitor the earth’s surface during severe weather. RADARSAT provides a straightforward, all-weather means of mapping flood extent. It is also beneficial for mapping the effects of hurricanes through observation of their influence on the ocean’s surface. RADARSAT images may lend insight to hurricane morphology and prediction, and could have a role in hurricane tracking. Continued access to this class of image data is ensured through the upcoming RADARSAT-2 mission, now scheduled for launch in 2002. 5. ACKNOWLEDGEMENTS We thank Satellite Operations at CSA for their efforts during “Hurricane Watch”. We thank L. Wu (Noetix Research Inc.), S. McDonald (Uwaterloo) and J. Wolfe (CCRS) for their help with the acquisition planning and data processing. M. Leahy (CCRS) and V. Bulzgis (CCRS) contributed to the analysis of the flood image. 6. REFERENCES Cranton, J.A., T.J. Pultz, J.L. Sokol, and D. Greenlee, “Integration of Geospatial Data for Applications in Flood Prediction and Management", Proc. ESRI International User Conference, 26 to 30 July 1999, San Diego, California, http://www.esri.com/library/userconf- /archive.html. Vachon, P.W., K. Katsaros, P. Black, and P. Dodge, “RADARSAT Synthetic Aperture Radar Measurements of some 1998 Hurricanes”, Proc. IGARSS’99 28 June to 3 July 1999, Hamburg, Germany, pp 1631-1633. ABSTRACT 1. RADARSAT 2. HURRICANES 3. FLOOD MAPPING 5. ACKNOWLEDGEMENTS 6. REFERENCES