CGIAR-CSI SRTM Frequently
Asked Questions (FAQ)
What
is the source of this data, and why is it different
from the NASA data?
The current dataset (Version 2) has been produced
based on the “Unfinished” 3 arc-second
SRTM data released by NASA and distributed by the
USGS through ftp access (ftp://edcsgs9.cr.usgs.gov/pub/data/srtm/version1/).
The original data came with data voids, where insufficient
contrast was available in the radar data to extract
the elevation. These data voids tend to occur over
water bodies (lakes and rivers), areas with snow cover
and in mountainous regions (for example, the Himalayas
has the greatest concentration of no data voids in
the original data). The CGIAR-CSI SRTM dataset has
undergone post-processing of the NASA data to “fill
in” the no data voids through interpolation
techniques (see the Data Processing and Methodology
page for detailed description). The result is seamless,
complete coverage of elevation for the globe.
What formats are available?
The data comes in two formats: arc-formatted ASCII
and GeoTIFF.
Arc ASCII: This is a text-based data array, with
a five-line header providing the information on spatial
extent and resolution.The coordinates in the header represent the lower left corner of the lower left pixel.
No data values are represented by values of –9999.
This data can be imported into ArcView 3.2 through
use of the Spatial Analysis extension and File->Import
Data Source.
Within ArcGIS, there are different methods available
for importing ASCII files. You can access toolboxes
from the ArcToolbox window or the ArcCatalog tree.
The Conversion toolbox contains tools that convert
data between various formats. Select Ascii to raster.
Next a dialog box appears where you supply the parameter
values for the tool, then click OK. Alternatively,
the command line syntax is:
ASCIIToRaster_conversion <in_ascii_file> <out_raster>
{INTEGER | FLOAT}
Many other GIS software can read ARC ASCIIs through
import functions.
GeoTIFF: The GeoTiff data are georeferenced images
(with the geographic data embedded – there is
no .gtif file) in signed 16 bit (-32768 to 32767)
TIF format. They have been produced through export
from ArcGIS 8. In the GeoTIFF files, values of –32768
represent no data regions.
What is the mask data?
The mask data is also available on the web site.
This represents grids of 1’s and no data values
showing the areas where the original data came with
data voids. The areas with values of 1 in the mask
data have been subjected to the interpolation technique,
and therefore contain interpolated elevation values
NOT available originally in the NASA release of the
SRTM data. The error margin within these original
data void areas should be expected to be greater.
When I import GeoTIFF data, values
of –32768 appear. What does this mean?
When the GeoTiff files are imported into some software
programs, it has been reported that values of –32768
result. This is due to the software failing to identify
values of –32768 as pixels with no data. Users
can either search for a solution that fixes this or
should explore the ASCII files that so far have not
presented any similar problems.
This problem can be solved by reclassifying (reclassifying
–32768 as NO DATA) or querying all pixels with
values greater than –32768, and recalculating
the grid based only on those regions. One command
in Arc/Info can also solve this (the user should specify
the name of the grid for “outgrid” and
“srtmdem”):
OUTGRID = setnull(SRTMDEM == -32768, SRTMDEM)
Why do some mountain regions have
peaks significantly lower than they should be?
As mentioned earlier, many original data voids are
concentrated in mountainous areas and in snow-covered
regions. Hence, many peaks in high-mountain areas
are actually interpolated. Without using a co-variable
for the interpolation (like GTOPO30 1km elevation,
planned for Version 3), the interpolation fails to
identify that the data void is actually a peak, and
tends to “flatten” the peak, leading to
underestimates in the true elevation for that region.
This issue is to be addressed in Version 3.
Why do some small islands not appear
in the dataset?
The dataset available online has been cut for shorelines
using the GSHHS - A Global Self-consistent, Hierarchical,
High-resolution Shoreline Database. In some cases
this dataset fails to include small islands, and in
other cases the islands are slightly mis-positioned.
The result is that these small islands, originally
captured in the SRTM data, are lost from the dataset
during the cutting process with shorelines. Uncut
data is available upon request – this includes
large areas of radar backscatter over water bodies.
Why are there negative elevation
values along the coasts?
The coastlines mask used to cut the data does not
have the same level of precision as the SRTM data
itself, hence some areas of land have actually been
lost in the cut, as well as some areas of sea also
entering into the dataset as land. Hence, in the latter
case, some elevation values may be negative.
Why are some of my coastal areas
not included in the data?
As stated in the previous answer, the coastlines
mask that was used to cut the data to only include
land areas does not have the same precision (scale)
as the SRTM data, hence some areas of land (and islands)
have been lost. If this is a problem you are invited
to request the uncut data by emailing Andy Jarvis
(a.jarvis@cgiar.org).
Why do some regions have very smooth
looking topography?
Interpolated regions in the dataset can often be
spotted as they appear to be very smooth, visibly
lacking low-scale detail. These are the areas with
original no data voids, that have undergone interpolation.
When is the next update likely to
be available, and what modifications are expected?
The current release is Version 2. We are currently
working on Version 3 with a release date penned for
February 2006. Version 3 will make a number of refinements
and improvements to the data:
• A better water-bodies mask (the NASA release
SRTM Water-Bodies Database) applied to cut the coastlines
and ensure lakes are flat
• Use of Version 2 NASA release “Finished”
grade SRTM data
• Use of GTOPO30 (1km), NED DTM and ASTER DEMs
(15m) as co-variables in the interpolation of no data
voids
• Improved interpolation techniques that are
applied depending on the topography typology (i.e.
different technique for flat areas than for mountainous
areas)
It is envisaged that CGIAR-CSI SRTM Version 3 will
be our definitive and final release of the 3-arc-second
product.
Can I use this data for commercial
use?
If interested in using this data for commercial purposes
please email Andy Jarvis (a.jarvis@cgiar.org).
|
