Ask A Cartographer

Missing values in SRTM/ASTER DEM data due to cliffs along coastline

November 23 2011 | 3 comments
Categories: Map Data

I am in the process of creating a map for an area in southern Tierra del Fuego, Chile. However, the available DEM data for this area (both SRTM and ASTER) includes significant missing values along the coastline due to sharp changes in elevation- the area of concern is marked by cliffs that drop to the Strait of Admirantazgo. These values have been confirmed to be missing by local guides for the area that are familiar with this coastline, and by digitizing the coastline in GoogleEarth and comparing it with the data.

The missing values clearly detract from the quality of the data and the contours created for this area dodge the missing values. In an attempt to fix this problem, a python script using the focal stats tool was run to interpolate out null values. However, although this script has worked successfully for other data sets (with a significantly lower number of null values), it failed to alter the coastline in this case.

Do other options exist to clean-up this dataset/interpolate out null/background values? Ideally, I would like to use the 30m ASTER data for this map.

Attached are three jpeg images displaying the problems with the current data. The images of the SRTM data show the missing coastline values as null values within the dataset. In the ASTER data, the missing values are background values.

Mapping Center Answer:

Google Earth has a different accuracy and precision compared to SRTM data, so it is unlikely the data will match if you are digitizing shorelines from Google Earth.

Using the Focal Statistics tool may not fill the void of the large areas in one iteration. You may need to run it a couple of times, which can potentially have the disadvantage of altering your data in other areas.

Another and possibly better option is to fill the voids using lower resolution data such as the newly released GMTED2010 (approximately 225 m resolution) data. You can use the Conditional Evaluation with Con tool and Map Algebra to void-fill the data.

Alternatively, you can download already void-filled SRTM data from the Consortium for Spatial Information web site. The latest version of ASTER data is also available to download from USGS which is a significant improvement over the previous version. You may want to try these datasets and see if that will solve your issue.

Thank you posted by Katie Panek on Jan 10 2012 11:34AM
I am now using the void-filled SRTM data from the Consortium for Spatial Information for these maps and it is working very well. I would recommend this data set for anyone with significant missing values in their data. Thank you for your advice!
Void-filled SRTM data from Esri posted by Aileen Buckley on Jan 10 2012 12:23PM
There is also void-filled SRTM dta on the Esri Maps & Data DVD. Here is a blog post we wrote about this great resource for data:

http://blogs.esri.com/Support/blogs/mappingcenter/archive/2011/06/07/esri-data-and-maps.aspx

If you would like to post a comment, please login.

Contact Us | Legal | Privacy |