SRTM-GeoTIFF

Python snippets to read elevation data from raster-based SRTM files (GeoTIFF) from data sources such as:

USGS EarthExplorer Japan Space Systems NASA ASTER GDEM ALOS AW3D30 OpenTopography

File types supported by this snippet: Public Domain- GeoTIFF, Military- DTED, ESRI- BIL and legacy- HGT

Data type supported by this snippet: raster files with few restrictions on pixel dimensions or aspect ratios. It can handle these file structures: NASA/JSS ASTER GDEM 3601x3601. ALOS AW3D30 3600x3600. USGS 3601x3601, 1801x3601, 1201x1201 and 601x1201.

Snippet makes use of GDAL's GetGeoTransform (Affine Transformation) to translate the given longitude latitude into pixel indices inside the GeoTIFF raster layer. It also handles LZW compressed rasters automagically.

The more time consuming task is to find out which SRTM/GeoTIFF tile to use for a particular lat/lon location. Each data source provider uses a different file naming convention. For personal hobby use, I use GeoTIFF tiles because the file naming convention (with small additional regex manipulations on filenames) is similar to the original SRTM .hgt files. For ALOS's 3600x3600 non-overlapping tiles, a modified algorithm tile_alos.py is necessary to parse a given lat/lon to select the correct filename.

Below are some providers of ASTER GDEM in GeoTIFF tiles:
(ASTER = Advanced Spaceborne Thermal Emission and Reflection Radiometry)
(GDEM = Global Digital Elevation Model)

NASA's Earth Data Filters: Instrument=ASTER; Data Format=Cloud Optimized GeoTIFF

USGS EarthExplorer.
primer on how to download GeoTIFF from USGS EarthExplorer.

Japan Space Systems ASTER GDEM Select tiles using an interactive map.

Snippet to determine which GeoTIFF tile(s) to download/use:

For NASA, JSS and USGS (1-pixel overlapping tiles):

$python3 whichtile.py

For ALOS (non-overlapping tiles):
In words, if the latitude is exactly an integer, then use 1 tile further south compared to NASA/USGS counterpart.

$python3 whichalos.py

After assembling the necessary tiles in a local folder, you can add elevation points to GeoJSON, KML and gpx files.

Adding elevation to GPX file

Reads a GPX file and determines which GeoTIFF tiles to use (on local drive), reads it and extracts the elevation, updates or adds elevation to GPX waypoints, routes and tracks.

Adding elelvation to GeoJSON file

Reads a GeoJSON file and determines which GeoTIFF tiles to read (on local drive), reads it and extracts the elevation, adds elevation to Point, MultiPoint, LineString, MultiLineString, Polygon, MultiPolygon and GeometryCollection.

Adding elevation to KML file

Reads a KML file and determines which GeoTIFF tiles to read (on local drive), reads it and extracts the elevation, add elevation to Point, LineString, Polygon with LinearRing and MultiGeometry.

Example: Grouse Grind is a popular hiking trail in Vancouver, Canada

$sudo apt install gdal-bin
$python3 gpx2elev.py grouse-grind
$python3 geo2elev.py grouse-grind
$python3 kml2elev.py grouse-grind

Manually find the elevation of a single geolocation:

Using Python 3.10.6 and GDAL 3.4.3; Windows Subsystem Linux (Ubuntu 22.04)

>>>import srtm1
>>>srtm1.read( './N46E008.tif', 46.854539, 8.49701)
2691 (Lucern, Switzerland)
>>>srtm1.read( './N49W123.tif', 49.68437, -122.14162 )
618  (British Columbia, Canada)
>>>srtm1.read( './S33W071.tif', -32.653197, -70.0112 )
6926 (Aconcagua, Argentina)
>>>srtm1.read( './S36E149.tif', -35.2745, 149.09752 )
801  (Canberra, Australia)

Some older snippet to determine which GeoTIFF tile to use:

For NASA and USGS (1-pixel overlapping tiles):

>>> import tilename
>>> tilename.find( 49.6, -122.1 )
>>> 'N49W123.tif'
>>> tilename.find( 49.0, -122.1 )
>>> 'N49W123.tif'

For ALOS (non-overlapping tiles): In words, if the latitude is exactly an integer, then use 1 tile further south compared to NASA's/USGS's counterpart.

>>> import tile_alos
>>> tile_alos.find ( 49.6, -122.1 )
>>> 'N49W123.tif'
>>> tile_alos.find ( 49.0, -122.1 )
>>> 'N48W123.tif'

Subtle difference, but critical to avoid index out of range errors if latitude is an integer.