NASADEM: Creating a New NASA Digital Elevation Model and Associated Products

[Lurker]

Our objective is to provide the scientific and civil communities with a state-of-the-art global digital elevation model (DEM) derived from a combination of Shuttle Radar Topography Mission (SRTM) processing improvements, elevation control, void-filling and merging with data unavailable at the time of the original SRTM production:

• NASA SRTM DEMs created with processing improvements at full resolution
• NASA's Ice, Cloud,and land Elevation Satellite (ICESat)/Geoscience Laser Altimeter (GLAS) surface elevation measurements
• DEM cells derived from stereo optical methods using Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data from the Terra satellite
• Global DEM (GDEM) ASTER products developed for NASA and the Ministry of Economy, Trade and Industry of Japan by Sensor Information Laboratory Corp
• National Elevation Data for US and Mexico produced by the USGS
• Global Multi-resolution Terrain Elevation Data 2010 (GMTED2010) developed by the USGS and the National Geospatial-Intelligence Agency (NGA)
• Canadian Digital Elevation Data produced by Natural Resources Canada

We propose a significant modernization of the publicly- and freely-available DEM data. Accurate surface elevation information is a critical component in scientific research and commercial and military applications. The current SRTM DEM product is the most intensely downloaded dataset in NASA history. However, the original Memorandum of Understanding (MOU) between NASA and NGA has a number of restrictions and limitations; the original full resolution, one-arcsecond data are currently only available over the US and the error, backscatter and coherence layers were not released to the public. With the recent expiration of the MOU, we propose to reprocess the original SRTM raw radar data using improved algorithms and incorporating ancillary data that were unavailable during the original SRTM processing, and to produce and publicly release a void-free global one-arcsecond (~30m) DEM and error map, with the spacing supported by the full-resolution SRTM data.

We will reprocess the entire SRTM dataset from raw sensor measurements with validated improvements to the original processing algorithms. We will incorporate GLAS data to remove artifacts at the optimal step in the SRTM processing chain. We will merge the improved SRTM strip DEMs, refined ASTER and GDEM V2 DEMs, and GLAS data using the SRTM mosaic software to create a seamless, void-filled NASADEM. In addition, we will provide several new data layers not publicly available from the original SRTM processing: interferometric coherence, radar backscatter, radar incidence angle to enable radiometric correction, and a radar backscatter image mosaic to be used as a layer for global classification of land cover and land use.

This work leverages an FY12 $1M investment from NASA to make several improvements to the original algorithms. We validated our results with the original SRTM products and ancillary elevation information at a few study sites. Our approach will merge the reprocessed SRTM data with the DEM void-filling strategy developed during NASA's Making Earth System Data Records for Use in Research Environments (MEaSUREs) 2006 project, "The Definitive Merged Global Digital Topographic Data Set" of Co-Investigator Kobrick.

NASADEM is a significant improvement over the available three-arcsecond SRTM DEM primarily because it will provide a global DEM and associated products at one-arcsecond spacing. ASTER GDEM is available at one-arcsecond spacing but has true spatial resolution generally inferior to SRTM one-arcsecond data and has much greater noise problems that are particularly severe in tropical (cloudy) areas. At one-arcsecond, NASADEM will be superior to GDEM across almost all SRTM coverage areas, but will integrate GDEM and other data to extend the coverage. Meanwhile, DEMs from the Deutsches Zentrum für Luft- und Raumfahrt Tandem-X mission are being developed as part of a public-private partnership. However, these data must be purchased and are not redistributable. NASADEM will be the finest resolution, global, freely-available DEM products for the foreseeable future.

data page:

https://lpdaac.usgs.gov/products/nasadem_hgtv001/

news links:

https://earthdata.nasa.gov/esds/competitive-programs/measures/nasadem

 

[juliusmall]

Ambitious........ To merge SRTM and Aster....

Personally I worked with SRTM and ASTER somewhere in west Africa, near the Ecuador, - 2  latitude area. 5000 sqkm the biggest exploration permit.

ASTER was full of artifacts, something like "small circular spots"  (difficult to understand the origin, perhaps the clouds traces) and lines/shadows in a rectangles grid, probably the connections at the limit of the original data tiles, the swats. Also, big "holes" with   more than 200 m elevation drop, almost vertical, from a mountain ridge.

Not existing in SRTM; because it was a critical area, we checked with a "macete"  team in the the jungle; no such huge hole there.  

Also it was a regular elevation  offset, some goods meters among the datasets.

Joining the grids I obtained largely a DEM showing me the relief, but full of spots (dots) in a regular grid. The dots  were obvious generated by one of the grids, with an vertical offset. 

Edited May 21, 2020 by juliusmall
Adding some words

[rahmansunbeam]

Interesting! 

@juliusmall, have you tried ASTER GEDM v3? The problem you are facing are probably elevation void areas which are reduced in the new version released last year.  

[juliusmall]

@rahmansunbeam

Thanks for the tip. I will check.....

The project is not available now, long market story. But I know the place by heart.....

That hole after the ridge was a nightmare for me because of a critical chain of exclusion ( a pipe, crossing the mountains,  at a limited angle, a maximum elevation, etc).

But as I wrote before, I made some merges of  the two data sets SRTM and ASTER. The result was something like a smallpox affected skin, but with a perfect rectangular pattern. Obvious a offset of the elevations. I observed this on large surfaces.

Sorry to discard all, so I cannot say which one one was up and which down, I think I remember, something of meters order of magnitude, almost 10 m.

 

 

 

[Loxcel]

This link contains our analysis of NASADEM vs SRTM across North America. We used 252 million extremely accurate IceSat-2 terrain height benchmarks for the comparison. The analysis also includes AW3D30 and ASTER DEMs for good measure.

This link is our second analysis of NASADEM across Australia and New Zealand.

NASA also publishes an ellipsoidal height dataset which improves accuracy a wee bit, which can be helpful if you're using bicubic interpolation. It has voids which are easy to correct. Note: its heights are not relative to sea level, and are different than what Google Earth and most other mapping products provide.

In short, NASADEM is a little better than SRTM in forests (less bias), short vegetation, wetland and urban areas, and a bit worse in barren areas and open water. NASADEM is even better in Australia, too. Also take a look at AW3D30 v3.1.

[Loxcel]

Can't find a way to update my post from Jul 30. The link to follow has moved to here