Soil moisture retrieval using ALOS PALSAR

Transcription

1 Soil moisture retrieval using ALOS PALSAR T. J. Jackson, R. Bindlish and M. Cosh USDA ARS Hydrology and Remote Sensing Lab, Beltsville, MD J. Shi University of California Santa Barbara, CA November 6, 2008

2 Project Background Radar remote sensing provides very high spatial resolution data that can be used to estimate soil moisture. Robust retrieval methods have not been developed. Limitations of previous satellites systems have been the wavelength and the number of independent radar parameters. ALOS PALSAR includes a low frequency and multiple polarizations that may lead to improved soil moisture retrievals. Scaling from PALSAR to SMAP

3 Outline Soil Moisture Active Passive (SMAP) Mission Mission description Role of PALSAR Cloud Land Surface Interaction Campaign (CLASIC) PALS/PALSAR Soil moisture retrieval SMAP Validation Experiment 2008 (SMAPVEX08)

4 Soil Moisture Active Passive (SMAP) Satellite

5 Soil Moisture Active Passive Mission NASA One of the first missions resulting from the NRC Decadal Survey L-band: microwave radar (1-3 km), microwave radiometer (40 km) 40 degrees incidence angle Polarimetric Three day global coverage Launch 2013 (SMAP)

6 SMAP Mission Concept Radiometers and radars each have advantages and disadvantages for soil moisture. Limited spatial resolution of the radiometer and the robustness of the radar algorithms SMAP will attempt to overcome these with: A technology that provides a large aperture low mass/volume antenna (deployable mesh) Enhanced resolution by combining high accuracy radiometer retrieval with high resolution radar (10 km product)

7 SMAP Science and Applications Decadal Survey Panels Water Resources and Hydrological Cycle Climate / Weather Human Health and Security Land-Use, Ecosystems, and Biodiversity Cited SMAP Applications 1. Floods and Drought Forecasts 2. Available Water Resources Assessment 3. Link Terrestrial Water, Energy and Carbon Cycles 1. Longer-Term and More Reliable Atmospheric Forecasts 1. Heat Stress and Drought 2. Vector-Borne and Water-Borne Infectious Disease 1. Ecosystem Response (Variability and Change) 2. Agricultural and Ecosystem Productivity 3. Wild-Fires 4. Mineral Dust Production Societal benefit and applications will impact mission development and implementation Applications that utilize SMAP (and SMOS) observations will be of great interest to NASA and supported under a variety of programs *Material from SMAP Science Team

8 SMAP Status NRC Decadal Survey Report (Jan. 2007) NASA prioritization (Feb. 2008) SMAP Project and Program (March 2008) Science Team Selected (October 2008) Launch 2013

9 SMAP Synergy With Other Missions/Applications A great decade for low frequency microwave remote sensing! SMAP provides continuity for L-band measurements of ALOS, SMOS, and Aquarius, and synergy with GPM and GCOM-W 1.2 SMAP 1 GPM 0.8 GCOM-W 0.6 Aquarius 0.4 SMOS 0.2 ALOS 0 Estimated Mission Timeline *Material from SMAP Science Team

10 Evolution of L-Band Remote Sensing Resolved Temporal Scales Day Week Month Aquarius SMOS Climate Applications Radiometer Evolution of L-Band Sensing SMAP Radar-Radiometer Weather Applications Radar Carbon Cycle Applications ALOS SAR 100 km 10 km 1 km Resolved Spatial Scales *Material from SMAP Science Team

11 Relevant Issues (PALSAR/SMAP) RFI information for design Data resource for algorithm design and validation. How do high resolution SAR data (PALSAR) scale to 1-3 km (SMAP)? Can we use PALSAR to simulate SMAP radar? Field experiments with SMAP simulator and concurrent PALSAR coverage. (challenge) CLASIC SMAPVEX08

12 CLASIC June 2007 Southern Great Plains Oklahoma (included the LW) Intensive ground sampling of soil moisture over a one month period Aircraft coverage with the Passive Active L-band System (PALS) (SMAP Simulator)

13 Single beam Fixed angle (40 degrees) Spatial resolutions ~ 0.3(Altitude) First deployment of new antenna

14 PALS Experiment Design over the Little Washita Watershed 8 lines in eastwest direction Spatial resolution ~400 m 11 flights Predominantly pasture and dry winter wheat

15 CLASIC Radar Coverage June-July FC, LW LW ALOS FS FS LW FC, LW FC, LW FC FC, CF FC FC, LW ALOS LW FC, LW LW FC, LW FC, LW LW Legend: Blue=PALS, Green=ALOS and PALS

16 PALS Radiometer Mapping L-band H-polarization over Little Washita Watershed June 8 June 11 June 12 June 19 June 23 July 1 July 3 July 4 July 5 July Brightness Temperature (K)

17 PALS Radiometer and Radar Mapping L-band H-polarization and HH over Little Washita Watershed June 8 June 11 June 12 June 19 June 23 July 1 July 3 July 4 July 5 July Brightness Temperature (K) June 8 June 11 June 12 June 19 June 23 July 1 July 3 July 4 July 5 July 6 July Backscatter Coefficient

18 PALSAR FBD Little Washita July 1, 2007 Bandwidth 14 MHz Polarization HH+ HV Incidence Angle Swath 34.3 deg 70 km hh=r,b hv=g

19 Comparison Between PALS and PALSAR Cross-validation of radars and scaling. Averaged PALSAR observations for every PALS footprint (400 m resolution). Note that Backscatter coefficient decreases with decrease in incidence angle - PALSAR measurements should be slightly smaller than PALS. The two sensors also have different azimuth angles.

20 Comparison Between PALS and PALSAR PALSAR Sigma0 Cross-validation of radars and scaling. Averaged PALSAR observations for every PALS footprint (400 m resolution). Note that Backscatter coefficient decreases with decrease in incidence angle - PALSAR HH (SEE=1.80) measurements should be slightly HV (SEE=1.91) smaller than PALS. The two sensors also have different azimuth angles. Some of the scatter can be attributed to the sub-pixel variability and noise. Resolution : PALSAR (12.5 m), PALS (400 m). Good agreement between the two sensors without significant bias indicates comparative calibration and possibly linear scaling PALS Sigma0

21 Soil Moisture Retrieval-Preliminary Next phase will retrieve soil moisture from the radiometer and radar of PALS and PALSAR. Current approach employs the Dubois Model, however, There is no vegetation parameterization Requires hh and vv can be directly applied to PALS data For PALSAR an adaptation for using hv instead of vv is needed

22 SMAP Validation Experiment 2008 (SMAPVEX08) Support for SMAP science issues Primary study area will be the Delmarva peninsula (Choptank Watershed). Multiple flights over a two week period (Sept. 29-Oct. 13, 2008) Two aircraft; four sensors Spatial resolutions of 1 and 3 km (more typical of SMAP than in CLASIC)

23 SMAPVEX08 Choptank Site 100 km

24 SMAP Validation Experiment 2008 (SMAPVEX08) Support for SMAP science issues Primary study area will be the Delmarva peninsula (Choptank Watershed). Multiple flights over a two week period (Sept. 29-Oct. 13, 2008) Two aircraft; four sensors Spatial resolutions of 1 and 3 km (more typical of SMAP than in CLASIC)

25 Sensors and Platforms PALS Twin Otter L-band, fully polarimetric radar/radiometer 40 degrees backward looking Spatial resolution ~ 0.3(Altitude): selected altitude will yield a ground resolution of ~1000 m Payload will include alternative RFI mitigation GSFC Digital Beam-forming Radar on P-3 L-band, fully polarimetric +/- 45 degrees cross track Altitudes (40 degree spatial resolution); 2000 m (1600 m) 5000 m (4200 m) Comrad Ground Based L-band, fully polarimetric radar/radiometer Multiple angles Spatial resolution ~ 50 m MAPIR Radiometer on P-3 L-band passive microwave 40 degrees conically scanned Altitudes (Spatial resolution); 2000 m (1000 m) 5000 m (2200 m) PALSAR ALOS L-band FBD HH, HV 34.3 degrees Sparial resolution of 12.5 m Limited dates Oct. 8

26 SMAPVEX08 PALS and COMRAD

27 Sensors and Platforms PALS Twin Otter L-band, fully polarimetric radar/radiometer 40 degrees backward looking Spatial resolution ~ 0.3(Altitude): selected altitude will yield a ground resolution of ~1000 m Payload will include alternative RFI mitigation GSFC Digital Beam-forming Radar on P-3 L-band, fully polarimetric +/- 45 degrees cross track Altitudes (40 degree spatial resolution); 2000 m (1600 m) 5000 m (4200 m) Comrad Ground Based L-band, fully polarimetric radar/radiometer Multiple angles Spatial resolution ~ 50 m MAPIR Radiometer on P-3 L-band passive microwave 40 degrees conically scanned Altitudes (Spatial resolution); 2000 m (1000 m) 5000 m (2200 m) PALSAR ALOS L-band FBD HH, HV 34.3 degrees Sparial resolution of 12.5 m Limited dates Oct. 8

28 Radar Scaling (RSCALE) Design Establish the scaling of PALSAR (12.5 m) to SMAP radar resolution via high altitude PALS (1 km) Establish the scaling of PALSAR (12.5 m) to SMAP radar resolution via high altitude DBSAR (3 km) Establish the scaling of PALS (1 km) to DBSAR (3 km) radar resolution P3 Line DBSAR 40 o 3 km Twin Otter Line TO Line PALS 40 o 1 km Portion of July 8, 2008 PALSAR IMAGE PALSAR 34.3 o

29 PALSAR FBD Coverage of the Choptank Site May 23, 2008 July 8, 2008 Aug 23, 2008 hh=r,b hv=g

30 PALSAR FBD Coverage of the Choptank Site October 8 data acquisition was cancelled due to a re-tasking a few days before. Our request was the default mode for this cycle. No disaster we could identify. We only found out by checking acquisition planning schedule never contacted. Asked for a reversal... There is a need for improvement in the communications and prioritization. Future plans.

31 SMAP Major Field Campaigns Year/ Quarter Australia GCOM-W SMAP 2 SMOS Aquarius Australia SMAPVEX10 SMAPVEX08 Australia Europe/Canada Europe SMAPVEX10 SMAPVEX13 SMAPVEX14 Satellite Launch in Red 3 ver. 10/08 4 Australia SMAPVEX11 SMAPVEX08 High priority design/algorithm issues Australia one-week campaigns to span four seasons Aircraft Radar/Radiometer Separate SMOS validation Europe (and Canada) SMOS validation launch delays possible No radar SMAPVEX10:CLASIC Spring-Summer 2010 Oklahoma SMOS and Aquarius available Focus of algorithm validation SMAPVEX11 Focus on different problems; FT, regions, seasons SMAPVEX13 and 14 SMAP product validation