1. G. Bryan Bailey, U.S. Geological Survey, USA Michael Abrams, Jet Propulsion Laboratory, USA Hiroji Tsu, ERSDAC, Japan Hiroyuki Fujisada, SILC, Japan Dean Gesch, U.S. Geological Survey, USA The New ASTER 30m Global DEM Practical Steps Toward Global DEM Interoperability Workshop Beijing, China July 2, 2008 The New ASTER 30m Global DEM

2. G. Bryan Bailey, U.S. Geological Survey, USA Michael Abrams, Jet Propulsion Laboratory, USA Hiroji Tsu, ERSDAC, Japan Hiroyuki Fujisada, SILC, Japan Dean Gesch, U.S. Geological Survey, USA The New ASTER 30m Global DEM Alaska Statewide DEM Workshop Anchorage, Alaska July 23, 2008

3. ASTER Global DEM (GDEM) Global DEM, at 30m postings, will be produced from ASTER stereo images. The GDEM will be a standard ASTER data product. It is jointly provided by Japan’s Ministry of Economy Trade and Industry (METI) and U.S. National Aeronautics and Space Administration (NASA), in cooperation with METI’s Earth Resources Data Analysis Center (ERSDAC) and the United States Geological Survey (USGS). GDEM was contributed to, and accepted by, GEO at the November ’07 Summit of Ministers. It will be available to all users at no charge to the user.

4. Available Global DEM Data Sets DatasetCoverageHorizontal Posting Comment GTOPO30Global1 km US Level 2 DTED Global30 m Restricted Access SRTM 60 o N to 54 o S 90 m 30m over U.S. and Its Territories ASTER GDEM 83 o N to 83 o S 30 m

5. GDEM Applications Disaster (flood hazard map) Hydrology (water resource management) Disaster (volcanic hazard map) Energy (oil resource exploration) Precipitation data Snow cover data Volcanic data Geological data GDEM

6. ASTER Advanced Spaceborne Thermal Emission and Reflection Radiometer Flies on U.S. Terra Platform Built for METI in Japan Data Processing in Japan and U.S. Instrument Command and Control by Japan and U.S. Joint Science Team

7. ASTER Instrument VNIR Visible/Near Infrared Radiometer # of Bands: 3 Nadir + 1 Aft-Looking Spatial Resolution: 15 m SWIR Short Wave Infrared Radiometer # of Bands: 6 Nadir Spatial Resolution: 30 m TIR Thermal Infrared Radiometer # of Bands: 5 Nadir Spatial Resolution: 90 m  Launched December 1999  Has 3 sensors and detects 14 bands of spectrum.  Ground resolution is 15m ~ 90m.

8. Instrument Characteristics 60 km swath; <16 day repeat cycle; stereo

9. Along-Track Imaging Geometry of the ASTER VNIR Nadir and Aft-Viewing Sensors 430 37060 0 Begin Acquisition of 3N Complete Acquisition of 3B 645590 Sec 27.7  705 Km Orbit Single Stereo Image 3N 3B Ground Relief Second s Km

10. ASTER Stereo Image Data ASTER 3N Image ASTER 3B Image

11. ASTER Digital Elevation Models ASTER DEM Intensity Image ASTER DEM Shade Relief Image

12. ASTER Data Acquisition

13. ASTER Data Acquisition - Alaska

14. ASTER GDEM Specifications Data Posting: 30 m DEM Output Format: GeoTIFF, signed 16 bits, and 1 m/DN Referenced to the WGS84/EGM96 geoid Geographic latitude and longitude Coverage: 83  N to 83  S Special DN Values : -9999 for void pixels 0 for sea water body Accuracies: 20 m with 95 % confidence for vertical data (elevation) 30 m with 95 % confidence for horizontal data (geolocation) Data Set: 22,895 1  x 1º tiles

15. Methodology ASTER GDEM ASTER coverage (~1.5 million scenes in summer 2008) Deeper red indicates more data accumulated. Generation of a seamless tile of DEM using all ASTER data ever acquired over the tile A seamless Tile of DEM GEOSS Both US and Japan committed to contribute to GEOSS at Capetown Summit 2007. Contribution to Global coverage Automated processing ASTER scene (60km x 60km) Tile (1 arc degree square) 1. Stereo-correlate entire ~ 1.5 million scene ASTER Archive; 2. Cloud mask to remove cloudy pixels; 3. Stack all DEMS & remove residual bad values and outliers; 4. Partition data into 1º x 1º tiles --- SRTM3 coverage

16. Cloud Masking Step Band 3N Cloud-masked DEM

17. Final Product Partitioned as 1  -by-1  Tiles ASTER DEM SRTM3 Mt. Everest Area Elevation: 8,848 m 203 scenes used No holes for ASTER DEM Many large holes for SRTM Geographic Boundaries Latitude: N26 - N27 degrees Longitude:E86 - E87 degrees Mt. Everest Peak Elevation 8815 m

18. Preliminary Evaluation Histogram of Difference between ASTER DEM and SRTM3 Mt. Everest Area Mean: -6.9 m Standard Deviation: 26.0 m

19. Susquehanna Test Site Assessment Results Prototype ASTER GDEM Number of Scenes Used to Produce Prototype ASTER GDEM 10 – 22 Scenes 1 – 9 Scenes

20. Susquehanna Test Site - Vertical Error Tabulations Statistics MINMAXMEANSTDEV Susque NED subtract ASTGTM-1155.0693547.6649-5.395811.7435 Susque NED subtract ASTGTM(Numbers 1 Scene mask)-1155.0693247.1051-5.525210.6168 Susque NED subtract ASTGTM(Numbers 2 Scene mask)-823.6140247.1051-5.43909.9873 Susque NED subtract ASTGTM(Numbers 3 Scene mask)-558.9187145.4936-5.18019.7834 Susque NED subtract ASTGTM(Numbers 4 Scene mask)-129.8367145.4936-4.57849.5178 Susque NED subtract ASTGTM(Numbers 5 Scene mask)-129.8367145.4936-3.75929.2287 Susque NED subtract ASTGTM(Numbers 6 Scene mask)-129.8367140.3742-3.18029.0778 Susque NED subtract ASTGTM(Numbers 7 Scene mask)-129.836773.4538-2.85769.0823 Susque NED subtract ASTGTM(Numbers 8 Scene mask)-129.836766.9593-2.74709.1678 Susque NED subtract ASTGTM(Numbers 9 Scene mask)-129.836766.9593-2.71609.2075 Susque NED subtract ASTGTM(Numbers 10 Scene mask)-129.836766.9593-2.64119.2336 Susque NED subtract SRTM-135.8367551.8665-5.37869.9934 Susque NED subtract L1A Oct 5 2001 Silcast DEM-185.5250134.0040-7.292017.4490 Susque NED subtract L1A May 6 2004 Silcast DEM-144.2590195.07006.420010.6300

21. Susquehanna Test Site – GCP Comparisons Ref. Elev.ASTERNEDSRTM ASTER errNED errSRTM err 250.62248.00250.49252.00 -2.62-0.141.38 353.37345.00352.56351.00 -8.37-0.81-2.37 334.77329.00332.27339.25 -5.76-2.504.49 259.58260.00260.13258.50 0.420.56-1.08 214.00220.00221.66215.00 6.007.661.00 163.33164.00163.96164.25 0.670.630.92 138.07129.00133.65135.00 -9.07-4.42-3.07 188.27193.00191.98189.75 4.733.721.48 Min-9.07-4.42-3.07 Max6.007.664.49 Mean-1.750.590.34 Std Dev5.693.722.43

22. N.W. Washington Test Site Assessment Results Prototype ASTER GDEM Number of Scenes Used to Produce Prototype ASTER GDEM 1 -52 scenes per pixel; Mean: 21.6

23. N.W. Washington Test Site - Vertical Error Tabulations Statistics MINMAXMEANSTDEV Wash NED subtract ASTGTM-258.609292.825-2.64817.232 Wash NED subtract ASTGTM(Numbers 1 Scene mask)-258.609292.825-2.64817.232 Wash NED subtract ASTGTM(Numbers 2 Scene mask)-258.609292.825-2.64817.230 Wash NED subtract ASTGTM(Numbers 3 Scene mask)-258.609292.825-2.65017.224 Wash NED subtract ASTGTM(Numbers 4 Scene mask)-258.609292.825-2.65317.210 Wash NED subtract ASTGTM(Numbers 5 Scene mask)-258.609292.825-2.65217.194 Wash NED subtract ASTGTM(Numbers 6 Scene mask)-258.609292.825-2.63517.172 Wash NED subtract ASTGTM(Numbers 7 Scene mask)-258.609256.636-2.60717.160 Wash NED subtract ASTGTM(Numbers 8 Scene mask)-258.609252.878-2.57317.167 Wash NED subtract ASTGTM(Numbers 9 Scene mask)-258.609252.878-2.58217.151 Wash NED subtract ASTGTM(Numbers 10 Scene mask)-258.609251.896-2.61017.135 Wash NED subtract SRTM (-150 mask)-192.325227.306-7.09313.493 Wash NED subtract L1A June 28 2000 Silcast DEM-127.185334.3899.80616.798 Wash NED subtract L1A June 05 2003 Silcast DEM-116.432323.15113.21713.573

24. N.W. Washington Test Site – GCP Comparisons Ref. elev.ASTERNEDSRTM ASTER errNED errSRTM err 103.9291.0095.9899.50 -12.92-7.94-4.42 104.87110.00111.98107.50 5.137.112.63 87.7389.0086.3190.50 1.27-1.422.77 89.8584.0085.0789.00 -5.85-4.77-0.85 12.590.0010.668.50 -12.59-1.92-4.09 194.87187.00196.07195.00 -7.871.200.13 375.94380.00378.45377.75 4.062.521.81 Min-12.92-7.94-4.42 Max5.137.112.77 Mean-5.47-1.29-0.64 Std Dev7.355.173.13

25. ASTER GDEM Validation Validation is expected to commence in late 2008. USGS and ERSDAC will lead the effort, with support from other agencies and international collaborators. Validation Plan involves two fundamental components. –Conterminous U.S. component.  934 CONUS tiles to be compared to NED and SRTM1 DEMs.  Absolute vertical accuracy to be measured using 13,300 “GCPs on benchmarks” from the National Geodetic Survey. –International component.  Up to 200 globally distributed tiles to be evaluated using SRTM data and other available validation data sets.  Investigators from the international community interested in contributing to ASTER GDEM validation should contact the authors. GDEM distribution sites will make validation results available to the user community.

26. ASTER GDEM Availability Public release of the full global data set is anticipated for Spring 2009. Data will be available at no cost to all users in the spirit of the Global Earth Observing System of Systems (GEOSS). ASTER GDEM will initially be packaged in 1  x 1  tiles, though “seamless” options my be considered in the future. Current plans are for data to be jointly released as a standard ASTER data product from: – U.S. through Land Processes DAAC and EOS Data Gateway. – Japan through ERSDAC Ground Data System.