The whole world is now on this site at 3 and 15 arc second resolutions.
Download data: 1" 3" 15" Search Tool
Last revision 14 December 2012. Current work in progress
Data developed and uploaded in Scotland by Jonathan de Ferranti. Interactive coverage maps supplied by Christoph Hormann.
For more information, see the continental links below.
ASIA * NORTH AMERICA * SOUTH AMERICA * ALPS * NORTH * OTHER EUROPE * AFRICA * ANTARCTICA * OTHERS
The downloadable 3 arc second 1°x1° Digital Elevation Models on this site are mainly based on data collected by the 2000 Shuttle Radar Topography Mission. SRTM data, in HGT format, can be downloaded from here, but for some mountain and desert areas there are no-data (void) and phase unwrapping error areas, and there are no SRTM data north of 60°20'N. For the tiles here, these areas have been filled and corrected from the best available alternative sources, using the method described on my voidfill page. They are much more accurate than those created by interpolation, with or without the aid of SRTM30. To see some images created from data downloaded from this page, click here. The accuracy of the data can be judged from these images. See also external links to an independent review, and some photographic comparisons.
ASTER GDEM. The 2009 and 2011 releases of a global ASTER GDEM dataset, with 1" postings, may also have been noticed. I welcome and make use of this new free data source, but I invite readers to read my initial review of these data before getting too excited. Perhaps I am biased, but I still believe that for many areas, the best available data are still on this site!
NORTH AMERICA is downloadable from the world coverage map. For US territory, the 3" data have been downsampled from USGS 1" data. A few errors have been removed, and nil values within US coastline have been amended to +1. Data for Mexican territory are primarily from SRTM/GDEM; Mexican 1" data have helped in some areas of high relief. For Canada, 0.75" Geobase data have been resampled to 3", using the mean of nine nearest 0.75" neighbours (with the exception of the province of Alberta south west of 54°N 113°W, where the Geobase data were found to be unsatisfactory and were replaced with SRTM/GDEM). To the best of my knowledge, higher resolution data are still freely available from US, Canadian and Mexican government sites. Seamless 1" coverage of North America, and near global 3" coverage can be found at rmw.recordist.com. Some 3" data from my site have contributed to this source, but I am not sure to what extent, or what mix of sources were used for the 1" data.
The Eurasian continent and Australia are now complete at 3" resolution and can be downloaded via the world coverage table.
Sources: SRTM, ASTER GDEM, Russian 200k and 100k, Nepal 50k and various others. All 8000m and most 7000m summits and their surroudings have been accurately mapped, but elsewhere accuracy may not be up to SRTM standard. Apart from a few error corrections, most of the areas covered by 2005-2011 SRTM voidfill from topographic sources have been left unchanged. By contrast, new areas covered have substantial input from ASTER GDEM data. Future updates are likely to make use of GDEM to improve the old areas too, but unless the GDEM stacking artefact issue can be resolved, SRTM will remain the default source.
For comparison of real photographs with images generated from these data and Landsat images, see Everest, K2 and Kangchenjunga. To see the original photographs, move the mouse over the images, and note that the virtual images have been partially overlaid with the photographs, and that there is a significant error margin in the camera positions.
SOUTH AMERICA is now complete and can be downloaded via the world coverage table.
Sources: SRTM, Aster GDEM, local 50k and 100k topos. Generally, accuracy is quite good, but is not always up to SRTM standard.
The main source is still SRTM, but where SRTM data are void or anomalous there is input from GDEM, Landsat and various topographic maps. Without these additional sources, I would be struggling to provide adequate SRTM voidfill. My thanks to the late Diego Vallmitjana from, Bariloche, Patagonia for help and encouragement.
Sources: Local 25k and 50k; Russian 100k; SRTM (limited). Most of the data contained in these tiles were generated from work done before the advent of SRTM. The accuracy of most of the data is up to SRTM standard, but there may be some slight local terracing and pockets of inaccuracy, especially in Italy.
3" Resolution are now only available via the world coverage table.
1" Resolution more information
To compare virtual images created from these files with real photographs, click on Switzerland, Austria, Germany, France and Italy.
§ Some parts of the Italian Alps in these areas were improved on 15 March 2008.
* Added or revised in February 2006; * Added or revised in April 2006. * Revised 11 April 2006 *Revised 13 April 2006. * Some visible seams smoothed 5 May 2006.
¹A slight horizontal discrepancy in some areas was drawn to my attention; on 1st January 2007 this was reduced by shifting some areas south by 2".
The above blocks cover the majority of Antarctic mountain areas. If other parts of Antarctica are needed in geographically projecterd HGT format, see:
UTM zones 01-15 (W180-W091)
UTM zones 16-30 (W090-W001)
UTM zones 31-45 (E000-E089)
UTM zones 46-60 (E090-E179)
The files are provided in HGT format for consistency with SRTM data. No source other than 200m NSIDC has been used to create them, so their true resolution is not 3". For realistic rendering, images generated from these should be reprojected to UTM or polar; both 3dem and Global Mapper will do this. I am grateful to Trond Nesøen of Fredrikstad, Norway for the provision of polar to geographic coordinate conversion tables.
December 14 Replaced SL39,SM42,SM58,15-V,15-X(sea artifacts) and J36(artifact at N37°49' E35°15')
December 8 Added Global 15" DEM
November 23 Added A56,A59,G02 (missing atolls)
November 22 Removed some fictitious land strips from SE57(s20e159 per media report) and R35(n71e029).
November 1 Completed the world!
October 15 Completed 7 continents; only islands are still missing.
October 15 Completed Latin America and uploaded data for parts of USA.
October 6 Removed line anomalies from B20 (n05w063, n05w061).
September 30 Extended South America to 8°N.
August 25 Completed South America south of the equator.
August 24 Replaced 4 folders due to remove some line anomalies. Folders replaced: H44(n31e079,n30e081), H46(n30e090,n30e091), I43(n35e074), J43(n38e075).
August 5 Replaced several folders due to incompletion of some previous updates. Folders replaced: O29, L31-33, K31-33, J53, H43-48, G43-48, SE19, SI19, SK59, SL58, SL59.
August 2 Completed Eurasia, Australia and south half of South America.
July 1 Completed Africa and New Zealand.
June 5 Completed India, Middle East, North Africa and South East Asia.
April 11 Completed High Asia.
February 17 Completed Indonesia, New Guinea and Malaysia.
December 30 Completed Philippines.
December 9 Completed Patagonia.
July 17 1" coverage of Sierra de Gredos, central Spain.
May 15 Extended coverage of Patagonia.
March 6 Added several tiles in Mauretania with large SRTM voids.
February 12 New edition of Iceland DEM.
February 5 Added N59E170 in East Siberia. Land area is small but SRTM void.
January 15 Added several tiles in Namibia with large SRTM voids.
January 3 Uploaded new edition of areas of East Siberia uploaded on December 21 (P58, P59, P60).
December 21 Added provisional data for oustanding parts of East Siberia: zones P59, P60 and the outstanding part of P58.
December 08 Added a tile in South China with large SRTM voids.
November 18 Added UTM zones 54, 55 and 56 to East Siberia. The outstanding tiles, P59, P60 and part of P58, will follow by Christmas.
October 31 Added UTM zone 57 to East Siberia.
September 15 Completed West Siberia.
August 21 Added Novaya Zemlya.
August 9 Completed areas P41 and Q41 (east of Ural mountains). Data capture for the remainder of north west Siberia is expected to be completed by 31 August, with uploads to follow soon after.
July 18 Completed Svalbard and Franz Josef Land
June 13 Added provisional data for south and west Svalbard
May 20 Tahiti DEM, South Georgia DEM, Egypt DEM and Kinabalu DEM added; some improvements and additions to East Tibet.
May 20 Extended coverage of Siberia: zones 51,52,53 and part of 44 added
April 29 Taiwan DEM added
April 9 Greenland DEM added
April 8 Added zone 58 to Siberia, and a sample desert tile (n24e012) with high SRTM void incidence
March 14 Extended coverage of Oman as far as Strait of Hormuz
March 5 Extended coverage of Siberia
September 11 Added Antarctica
July 25 Improvements and additions to Yunnan province, China
July 22 Added Chukotka, and more China-Burma border areas
July 10 Added Central Siberian plateau, some high relief areas of Oman, and more China-Burma border areas
March 15 Completed European mainland Russia and the Ural mountains
March 15 Improved part of the Italian Alps
December 30 Added Jan Mayen and Bear Island
December 24 Extended coverage of Northern Europe further into North West Russia
December 9 Added 1" DEM for High Tatra, Slovakia/Poland
October 5 Added Réunion (French overseas territory) and Alpi Apuane, Italy
August 14 North Russia, improved and extended areas P38 and P39
July 25 Improvements and addition of 1" data for Møre og Romsdal, Norway
July 8 Completed Iceland
June 28 Upgraded Scandinavia to version 2, with better contour interpolation
June 13 Completed south east Finland, added west Iceland, upgraded all of Norway's Jotunheimen, and added more of north west Russia.
April 21 Finland and Russian Lapland
February 16 More of Finland (provisional)
January 24 Fixed some tile boundary issues arising out of the January 1 revision
January 1 Corrected slight horizontal discrepancy in parts of the Alps south of 45°N
December 22 Finland (provisional)
November 13 Finnish Lapland (provisional)
October 23 Completion of Caucasus
October 11 More of ex-Soviet Asia and Afghanistan, including Ala Archa
September 30 More of ex-Soviet Asia and Afghanistan
August 22 eastern fringes of High Asian plateau
July 28 southern fringes of Eastern Himalaya
June 30 and July 1 north east of Karakoram (Yarkant, Karakash)
June 5 North Sikkim
May 20 Various High Asia additions, see High Asia section
May 5 Smoothed some seams between N 44°00' E 6°00' and N 46°00' E 8°00'
April 29 Added Santa Marta, Colombia
April 27 Corrected further datum error in Faeroe Islands
April 20 Added High East Africa
April 15 Added Moroccan High Atlas
April 13 Some more border and local terracing removed
April 11 Revised Italian pre-Alps tiles n45e009, n45e010, n45e011 (removed lake noise, fixed inaccuracy at N45°33' E10°50')
April 9 Added Scotland
April 5 Added Alps 1" data
April 2 Added some Italian pre-alps tiles, and repaired some terraces and scratches south of Julian Alps in Slovenia, and an inaccuracy at N 46°00' E 12°10'
March 26 Added Madeira
March 23 Added West Crete
March 22 Added Canary Islands (Palma, Tenerife)
March 20 Added Greece, Balkans and Italy's Corno Grande.
March 4 Corrected some null values and scratches in the Caucasus.
March 3 Added Picos de Europa in Northern Spain.
February 26 Corrected files north of 60° to WGS84.
February 21 Extended the highest summits of the Tien Shan, to between 79°30' and 81°15'.
February 1 Added New Zealand Southern Alps. Data of similar resolution can be downloaded from Geographx but are provided here to fill most of the voids in the geographically projected SRTM data.
February 1 Added some areas south and east of the Cordillera Blanca, Peru.
January 22 Added Vilcabamba and Vilcanota, Peru.
January 14 Added most of the Bolivian Cordillera Real.
January 7 Added the Patagonian Fitzroy and Paine groups.
January 2 Added a new 200km stretch of the Andes, centred on Aconcagua.
December 19 Cordillera Blanca, Peru addded.
December 14 Pamir extended again; Tien Shan extended; Meili and Bogda Shan added; some extensions to Central and Western Nepal; Corsican tile added.
December 5 Polish/Slovak Tatra mountains added.
November 26 Pamir and Eastern High Asia coverage extended.
New areas in the Eastern Himalayas, between Sepu Kangri (Nyainqentanglha East) and Bairiga (Garpo Kangri), have been illuminated. Some areas south of Tirich Mir (Hindu Kush, Hindu Raj) have also been illuminated.
Changes made on October 28
Scandinavia, Pyrenees, Shetland and Faeroes sections added.
Scratches removed from Himalayas section. A few isolated terraces, caused by phase errors in SRTM source data, remain; these will be removed in subsequent revisions.
Topographic Maps: YES For mountainous areas, the best alternative sources are detailed topographic maps, preferably based on material collected from ground surveys. These are many and various. The best general source is the Russian military. Their maps cover most of the world at 200k and much of the world at 100k. The contours on these are correctly placed, with very few exceptions. In some places there are elevation inaccuracies but these are often shown up by surrounding SRTM data and other sources and have been adjusted for. Most 7000m summits are covered by more accurate topos, e.g. Finnmaps of Nepal, Chinese Snow Maps and maps from various German sources. 250k JOG topos from the US military cover much of the world, but most of these have not been released, and the topographic quality of their AMS predecessors is very poor.
Spot elevations on Sketch Maps: YES Some elevations on some Japanese Alpine Club sketch maps covering most of the Chinese Eastern Himalayas that top 6000m were taken from Chinese military maps and were very helpful. But elsewhere in the Eastern Himalaya amd in some parts of the Indian and Chinese Western Himalaya, there are significant uncertainties.
Landsat images: YES In some areas, where I have found insufficient reliable topographic map detail, I have been consulting the shading on Landsat and other imagery used by Google Earth, and finding it to be helpful. More recently (summer 2008) I have using Landsat images more directly by reprojecting them geographically and using them as supplementary underlay. The results were surprisingly helpful, even though Landsat contains no elevation data. Landsat imagery is very accurate and its shades accurately match SRTM generated contour patterns.
GTOPO30 and GLOBE30: NO Even if the resolution of this source were adequate for mountain areas, which it is not, the quality of some of this is such that it is not unusual for two summits separated by 2km and a drop of 600m to show up as single summits. There may be good DTED Level 1 data (resolution 3", based on 250k maps) for some areas, but outside the US this has not been generally released by the US military.
ASTER GDEM: Yes, in some areas uploaded or updated since July 2009. For my general comments about this source, see my detailed review.
How to view shaded color relief images using 3DEM.
Note that SRTM voids can be patched by clicking on "Operation" and selecting "Patch Missing Data", but note also that this only patches the missing data by interpolating and extrapolating existing data. Missing data is not accurately determined and whole mountains may get left out. For best results, patching should not be used as a substitute for downloading tiles from this page.
See my screenshot comparisons page. Data from this page has also been used to generate some of the mountain panoramas on the panoramas page. Some parts of these have been reproduced alongside photographs from the same viewpoints on the panorama gallery page.
A shaded color relief image (1MB) of the Himalayas, from n28e085 to n29e089, has been created by Rafal Jonca of Poland, using 3DEM to convert the data downloadable from this page to .dem files. Click here for a higher resolution version (3.75MB). Surfer was used to color and shade, this is expensive but good results can be achieved with 3DEM too. For more information, see here.
Other DEM sources
Other DEM sources
NASA download website. SRTM data is usually available direct from NASA's website (no longer FTP). However, due to high demand and occasional maintenance, the site is not always available, and users are asked to limit their downloads. If you require a large quantity of data in .hgt format, please contact me instead. My details are at the bottom of my home page.
CGIAR download website. Here you can find SRTM data in other formats, improved by data imported from other sources, including this page. In August 2008 I checked a new version 4 that has been uploaded. It is a significant further improvement on version 3, especially in the Alps and Karakoram. But there are still some areas which have been covered by my files using topographic map data for some time, but which are still only covered by CGIAR with interpolated SRTM data. Also, close examination of some improved areas still shows some steep terracing and deep "trenches", and that all SRTM data that exists has been given priority, even where there are major phase errors in SRTM data. Still, some users may like the CGIAR format, and that they have, in one way or another, filled all the SRTM voids. There has been some smoothing down of mountain data from alternative sources, especially in the Alps. This has both advantages and disadvantages; among the advantages is smoother looking rendering, among the disadvantages is loss of mountain detail including the shoulder on the west ridge of the Matterhorn (shown on my screenshot comparisons page).
See also my virtual earth products review page.
HGT File Format
HGT files can be read and converted to other DEM formats by 3dem.
An HGT file covers an area of 1°x1°. Its south western corner can be deduced from its file name: for example, n51e002.hgt covers the area between N 51° E 2° and N 52° E 3°, and s14w077.hgt covers S 14° W 77° to S 13° W 76°. The fils size depends on the resolution. If this is 1", there are 3601 rows of 3601 cells each; if it is 3", there are 1201 rows of 1201 cells each. The rows are laid out like text on a page, starting with the northernmost row, with each row reading from west to east. Each cell has two bytes, and the elevation at that cell is 256*(1st byte) + (2nd byte). It follows that a 3" HGT file has a file length of 2 x 1201 x 1201. SRTM 3" cells are calculated by calculating the mean of 1" cells and their eight neighbors. It follows that the highest local point is likely to higher than the highest SRTM 3" cell. The difference should vary with the steepness of the local relief.
1" DEM data
NASA has only released 1" SRTM data for US territory. I have created HGT files with a resolution of 1" for some parts of Europe from topographic maps. For most of the Alps and the most rugged parts of the Pyrenees, these were created from the most detailed sources, and should be accurate. Elsewhere, I have 1" DEM data for Great Britain, Ireland, France (excluding Corsica), Germany, Benelux, Denmark, Norway, Sweden and north east Spain. But these were created from Russian maps of scale 1:100,000, which are less accurate than SRTM. Consequently, for these areas, where 3" DEM data is available from this page or from SRTM, they should be generally better than the 1" data, despite the lower resolution.
All the files, including the 1" files, downloadable from this page or otherwise available now conform to the WGS84 coordinate system used by SRTM. Previously, some of the Europe files outside the Alps conformed to the Russian Pulkovo 1942 coordinate system, which varies from WGS84 by about 200 metres.
Only the Alps 1" files are downloadable; I do not have the capacity to upload the remainder, but I am willing to supply them for the cost of reproduction, and I have uploaded two sample areas outside the Alps: P31, covering part of Norway's fjordland north west of 60°N 6°E, NorthCape, covering the area around Norway's North Cape.
Elevations and contour lines are facts that should be ineligible for copyright, but users should still note that significant commercial use of 1" DEM data may just possibly be contested by the authors of the source maps on copyright grounds. The most significant risk of this in the Alps, where there is the most input from local topographic mapping; in Scandinavia, where the source is almost exclusively Russian mapping, the risk is low. However, as of September 2006, despite extensive data use, especially in the Alps, by the flight simulation industry, no copyright problem has come to my attention.
Water bodies are flattened in all the files. In some files, including an alternative set of Alps files, they are defined by multiplying the cells on which they fall by -1, but they are not defined in any of the downloadable files except the above Scandinavian samples. There may be a slight gradient on some smaller lakes. Much of the terracing in Italy and Slovenia has now been fixed, but there may be still be some local terracing. An earlier issue concerning one pixel towers in Scandinavia has also been fixed.
In some areas, especially Scandinavia, the incidence of lakes and islands is very high, and manually checking all of them was impossible. Incidence of some very small islands and lakes may therefore be in error.
Current Work In Progress Last updated 1 November 2012
Current Work In Progress Last updated 1 November 2012