Using SRTM Data for SE Asia

We have imported the preliminary SRTM data for Eurasia and have filled the data holes.


This is the piece that we are using for the Mekong. First the one-degree tiles (You can see square holes where no tiles were produced for the ocean.) were assembled into 10-degree blocks, and the small voids (Large voids are ocean.) were filled. Parts of six big blocks (The NE and SW blocks were not needed.) were assembled to cover the Mekong Basin. Problems with voids at the junction of tile boundaries were fixed. You see that there are surface elevations for much of the ocean. The come mostly from the raw SRTM files, though more ocean elevations were added by the void-filling algorithm. A good shoreline cannot be derived from the DEM. We tried filling sinks and running hydrological analysis, but ran into problems where 30" data was used to fill voids in the Three Gorges area. We hope to fix this problem soon. The DEM has 693380352 cells, so operations that require intermediate files or arrays of 4 bytes/cell will exceed the two-gigabyte limit and fail.


For the current exercise, we have clipped the DEM to a shoreline polygon.

Shaded relief image from the SRTM DEM. It does not look its best at this scale.

This is a blowup of the previous iamges.

Another blowup approaches maximum resolution. Data spacing is 3 arc-seconds, approximately 90 meters.

This shows calculated streams. All DEM cells with a contributing area greater than 1000 pixels are shown.
Questions: Harvey Greenberg