WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 CEOS-WGCV Terrain mapping sub-group (TMSG) : Input for CEOS-WGISS Global Dataset Task-team Jan-Peter Muller, Lorant Czaran* Chair, CEOS-WGCV-TMSG,*Chair, CEOS-WGISS-GDTT MODIS & MISR science team member (NASA EOS) TerraSAR-X and TANDEM-X CoI HRSC CoI (ESA Mars Express)

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Overview Why do we need global topography? Current status (production and validation) Sources for completion from spaceborne data: –X-SRTM+ERS-tandem –SPOT-5 –ASTER –ICESat-GLAS Future sources –ALOS-PRISM –TANDEM-X Way forward : GLOBE 2.0

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Why do we need global topography? Six out of the 9 societal areas in the 10-year Implementation Plan for GEOSS require the highest possible resolution VALIDATED DEM (Digital Elevation Model) Most urgent, short-term need comes from Disasters Georadiometric processing of any land products require global DEM (e.g. global land cover) CEOS Plenary and WGCV have previously agreed that this DEM should be sourced from spaceborne sources and be fully VALIDATED Need to consider the development of global base maps in this context also as ALL 9 areas require this

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Status of spaceborne DEMs - Coarse resolution production and validation USGS-EDC-GTOPO30 and NOAA-NGDC/CEOS-GLOBE1 (30”≈1km) from Best Available Data (primarily US-NGA DTED1/0 and US-NGA- DCW) released in the mid-1990s. Detailed QA performed by NASA EOS- DEM Science WG. GTOPO30 operationally used for NASA-EOS processing. New v2 released in NOT VALIDATED. ERS-derived Radar Altimetry Corrected Elevation (ACE) at 30” (≈1km) developed under ESA funding by P. Berry (de Montfort University). NOT VALIDATED. SRTM30 - merger of unedited SRTM (averaged from 1->3->30”) with GTOPO30. NOT VALIDATED. GETASSE30 - ESA-ESTEC (M. Bouvet) : merger of ACE-SRTM30- EGM96. Used operationally for MERIS data processing. See later for details. NOT VALIDATED. ICESAT: major problems with 2 out of 3 lidars for global data acquisition. Data acquisition limited to 1-2 month acquisitions, 3 times/year. However, significant improvement in polar landmass heights for Greenland and Antarctica and substantial new data on vegetation/biomass. Validation started (Special Issue of Geophysical Research Letters, September 2005)

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 GTOPO30 used by NASA EOS processing chain - source DEMs

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Modified GTOPO30 using ERS Radar Altimetry Corrected Elevation (ACE) used by ESA for processing chain

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Examples of ICEsat-GLAS DEM products over polar landmass

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Status of spaceborne DEMs - Medium Resolution (30-90m) production ERS-tandem IfSAR (raw data acquired primarily in 1995/6) global coverage. Few national DEMs produced (UK-LANDMAP, Switzerland-SARMAP, Italy-Telespazio+XSRTM). Limited by atmospheric WV refraction effects although PS solution feasible if sufficient scenes are available (mostly Europe). No dedicated DEM processing project. Limited validation. SRTM (X-: DLR/ASI; C- NASA/DoD). Near global coverage (80% of landmass). Extensive validation performed and current status reported in AGU-EOS 2 May 06. ASTER. Stereo coverage based on individual requests and limited processing duty cycle. After 5 years, most of the Earth’s surface is covered in cloud-free stereo acquisitions but limited processing capabilities at EDC (2-3 DEMs/day) have restricted available relative DEMs. Increasing number of low-cost ASTER-DEM commercial software. Cost (COFUS) of ASTER level 1 data still issue for large-scale systematic DEM production. Good validation over US but very limited elsewhere. SPOT-5 (and SPOT1-4). IGN/SPOT working on global commercial 10m DEM but no report since 6/04. Publicly available validation limited. ALOS (PRISM). Hopes that WGCV-WGISS Plenary discussions can move this (stalled) process forward for international collaboration on validation. Would like to request that JAXA provides statement on global availability.

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 SRTM component of Unedited SRTM30 coverage Complete 30” (≈1km) DEM can be downloaded from ftp://edcftp.cr.usgs.gov/pub/data/srtm/SRTM30/SRTM30.tar

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006

Source data for ASTER stereo production

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Status of spaceborne DEMs - Medium Resolution (30-90m) validation ERS-tandem IfSAR - validation results in public domain limited to UK-LANDMAP project and TMSG web-site presentationshttp:// SRTM (X-: DLR/ASI; C- NASA/DoD). Consensus that SRTM-DEMs from X- and C- meet DTED-2 specification for height (Zrms≤8m) dependent on radar penetration of vegetation/built settlements. See more details later ASTER. USGS tests indicate that RMSExyz<<30m with 9≤RMSEz≤20m depending on date of acquisition, accuracy of orbital modelling and quality of GCPs. See more details later ICESAT: For flat, non-vegetated areas intercomparison with (6-foot footprint) airborne lidar DEM shows: 0.1±0.22m.

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 DUDES Concept SRTM-X from DLR and ASI ERS Tandem from ESA Merging of DEMs –Horizontal and vertical systematic relative error determination by comparison of the different DEMs –Horizontal and vertical systematic absolute error determination by minimum error norm assumption –DEM fusion: weighted average or wavelet approach

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 > 97 Mkm² of HRS cloud-free stereopairs (red = recent) SPOT 5 HRS DTED2 DEM + orthoimage ( >13M km² ) = HRS DEM in progress JRC-FOMI: RMSEz = 3.4m RMSExy = 5.75m

ALOS-PRISM GCP Database (Shimada) b) Site picture. v Contents of GCP database Latitude, Longitude : ITRF 97 Altitude : GRS 80 ellipsoid height Chip image : Aerial photograph or IKONOS Geographical map image Site location picture etc. * Aerial photo : 42cm/pixel, 1024x1024pixel IKONOS : 1m/pixel, 512x512pixel a) Top image.

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006

Way forward : GLOBE 2.0 Requirements: –DEM available by 2010 at the very latest –DEM must be free of ©, cost and 3rd party issues –DEM should be at least 3 arc-seconds (≈90m) without any gaps –DEM should preferably be 1 arc-seconds (≈30m) Where are we now? –C-SRTM provides ≈80% coverage of region from 60ºN-56ºS –X-SRTM could be employed to fill in many gaps (© and cost issues) –ERS-tandem could be employed to fill in remaining gaps but need remains to correct for atmospheric effects –ASTER stereo could also be employed to fill in many gaps but needs to be corrected with SRTM (where available) or using nearest ICESat- GLAS profile heights or through the use of bundle block adjustment –Unknown status of ALOS-PRISM global topography project described at ISPRS Unknown status of ©, cost and 3rd party IPR issues –TANDEM-X has been approved but DEM at 10m will be ©, very high cost and many 3rd party issues

WGCV-WGISS Plenary, Budapest, Hungary, 11 May 2006 Moving forward : GLOBE 2.0 GDTT produce a requirements (or “gap”) document in association with GEO Secretariat GDTT issue a call for proposals for participation in GLOBE 2.0 with participants bringing their own funding and providing this freely without © restrictions or any 3rd party issues GDTT to find a champion who will provide co-ordination of GLOBE 2.0 including promotion of the fusion of these different input data sources, develop documentation, provide final fused DEM, develop OGC-compliant distribution TMSG to provide validation, with suitable resourcing of individual input and fused products