GLIMS Glacier Database

From OpenStreetMap Wiki
Jump to navigation Jump to search

The Global Land Ice Measurements from Space (GLIMS) is a global project of mapping glaciers. Primary data sources are maps, often quite old and newer ASTER and Landsat images. The data is acquired from many institutions around the world using different methods and different tools and the quality of data therefore varies significantly.

Data access

The GLIMS database is hosted and maintained by the NSIDC, the main entry page there is:

data access including the full database dump can be downloaded from

There is also a separate domain containing various further information like:


The NSIDC has confirmed that GLIMS data is in the public domain:

GLIMS data are public domain so you may use it how you wish. However, we recommend 
the following format for a general citation for the entire GLIMS data collection:  

Richard Armstrong, Bruce Raup, Siri Jodha Singh Khalsa, Roger Barry, Jeff Kargel, 
Chris Helm, Hugh Kieffer. 2005. GLIMS Glacier Database. [indicate subset used]. 
Boulder, Colorado USA: National Snow and Ice Data Center. 

Alternatively, if you download a specific data set, please use the appropriate 
citation(s) listed in the file called 'CITATIONS_nnnn.txt' (where nnnn is a number) 
for the specific GLIMS data you use.

There is also a newer citation from

This file contains the citations for the data downloaded from the GLIMS
Glacier Database.  All data are in the longitude/latitude coordinate system
on the WGS-84 datum.

Field names in the downloaded file are documented at

When citing GLIMS data, please be sure to include all the information in
the examples below, including the analyst's name.

When referring to the GLIMS Initiative in general, please cite

   Raup, B.H.; A. Racoviteanu; S.J.S. Khalsa; C. Helm; R. Armstrong; Y.
   Arnaud (2007).  "The GLIMS Geospatial Glacier Database: a New Tool for
   Studying Glacier Change".  Global and Planetary Change 56:101--110.

For the complete set of GLIMS glacier data (when using it for a global
study of glaciers, for example), please cite

GLIMS and NSIDC (2005, updated 2018): Global Land Ice Measurements
from Space glacier database.  Compiled and made available by the
international GLIMS community and the National Snow and Ice Data Center,
Boulder CO, U.S.A.  DOI:10.7265/N5V98602

For Analysis_IDs in the range 309497--312426, the appropriate citation is

  Cogley, Graham (submitter); Kienholz, Christian; Miles, Evan; Sharp, Martin; Wyatt, F. (analyst(s)), 2015.
  GLIMS Glacier Database. Boulder, CO.
  National Snow and Ice Data Center.

Data structure

The full database dump is a huge shapefile containing ice and internal rock polygons separately. The shapefile also contains many duplicate polygons and the attributes seem partially broken.

Data quality

The quality of the data is strongly varying due to the different sources and analysis methods used. The following is by initial assessment (completeness is only a rough estimate)

region age of the data source completeness quality of the data better than current OSM
Alps (west) 1998 no good except for poor vectorization partly
Alps (east) 1985/~2000 no partly poor vectorization partly
Caucasus ~2000 mostly good yes
Scandinavia ~2000 yes good except for poor vectorization yes
Svalbard 2000-2010 no good except for poor vectorization yes
Iceland ~2000 yes good yes
Greenland ~2000 no bad no
Pamir 2001-2003 no good yes
Hindu Kush 2000-2005 maybe good yes
Tibet 1960-1975 unsure good but not very detailed yes
Karakoram (part of) 2000-2005 no good in parts covered
Eastern Himalaya 1960-1975 no good but not very detailed where covered
Tian Shan 1959-1975 no good but not very detailed yes
Altai 1959/2000-2005 no varying partly
Russian far east 2000-2005 no good yes
Kerguelen 1965/1994/2000 yes varying yes
Patagonia 2007 no good only where covered
northern Peru (small area) 2003 yes good except for poor vectorization yes
cont. US 1950 yes moderate yes
Alaska 2005-2010 no good except for poor vectorization where covered
Canada varying no varying mostly but probably worse than Canvec

Vectorization issues

Some of the better data in GLIMS, notably that in northern Europe and Alaska, is hampered by poor vectorization of a raster data source. To properly import this into OSM would require newly vectorizing it which is especially difficult where there have been manual edits in the GLIMS data. Below is an example from southern Norway

Glims example norway1.png Glims example norway2.png
Original GLIMS data with newly vectorized outline but trying to keep the original internal division