Last modified on Jun 30, 2026 10:09

Contributors: J. Bannwart (University of Zurich), F. Paul (University of Zurich), M. Zemp (University of Zurich)

Issued by: University of Zurich / Frank Paul

Date: 30 Apr 2026 

Ref: C3S2_WP1-DDP-GA-01_202511_PUGS_v7.0

Official reference number service contract: 2024/C3S2_313d_ENVEO/SC1 

History of modifications

List of datasets covered by this document

Acronyms

General definitions

For the glacier area product we define the Randolph Glacier Inventory (RGI) as a Climate Data Record (CDR) and the regionally and temporarily constrained improvements or updates submitted to the Global Land Ice Measurements from Space (GLIMS) database as Interim Climate Data Records (ICDRs). The latter might be integrated in new releases of the CDR. This document is related to the generated CDR.

Brokered data set: A dataset that is made available in the Climate Data Store (CDS) but freely available (under given license conditions) from external sources. In the case of the glacier service, the Randolph Glacier Inventory (RGI) is brokered for the CDS from https://nsidc.org/data/nsidc-0770/versions/7 (last access: 24.11.2025) under a CC-BY 4.0 license.

Debris-cover: Debris on a glacier is usually composed of unsorted rock fragments with highly variable grain size (from mm to several m). These might cover the ice in lines of variable width separating ice with origin in different accumulation regions of a glacier (so called medial moraines) up to a complete coverage of the ablation region. Automated mapping of glacier ice is only possible when the debris is not covering the ice completely in regard to an image pixel.

Glacier complex: A contiguous ice mass that is the result of the binary (yes/no) glacier classification after conversion from raster to vector format. Usually the glacier complexes are divided into individual glaciers by digital intersection with a vector layer containing hydrologic divides derived from watershed analysis of a digital elevation model (DEM).

Hypsometry: The hypsometry of a glacier gives the distribution of area with elevation, i.e. for each elevation interval covered by a glacier the area in this interval is listed.

Executive summary

This document provides a description of the Climate Data Record (CDR) dataset provided by the Copernicus Climate Change Service (C3S) Glacier Service to the Climate Data Store (CDS): The Randolph Glacier Inventory (RGI) containing glacier outlines (shapefile format) with glacier-specific attribute information and hypsometric information (csv file) for the c. 275,000 glaciers on Earth. In this updated version we also describe details of the related raster dataset that has been derived from the RGI at three different spatial resolutions (0.1°, 0.5 °and 1°).

In Section 1, we first provide a description of the RGI and how it was created, before we give an overview of the technical requirements for the product (glacier area) according to GCOS and show some visual examples of the datasets. The next section provides information on data usage, including a description of data formats, file contents, example applications and known issues. Section 2 provides information on data access and Section 3 on user support.

Product Description

The Randolph Glacier Inventory (RGI)

The RGI is a globally almost complete collection of digital glacier outlines (Figure 1) that was for its first version compiled by the science community in a short period of time to serve as a base for global-scale glacier-related calculations in support of the Intergovernmental Panel on Climate Change Fifth's Assessment Report (IPCC AR5) (Vaughan et al. 2013). It was largely based on glacier outlines from the Global Land Ice Measurements from Space (GLIMS) database (https//:www.glims.org, last access: 24.11.2025) and additional outlines derived mostly from satellite data such as Landsat but also Advanced Spaceborne Thermal Emission and reflection Radiometer (ASTER) and since 2015 Sentinel-2. Figure 2 shows the timeline of freely available satellite data. In a few regions the outlines are also based on digitised topographic maps, or nominal data from the World Glacier Inventory (WGI) are used where glaciers are represented by circles of the related size (Pfeffer et al. 2014). The dataset was improved several times and in the new version 7.0 of the RGI all nominal glaciers have been replaced by their real outlines and a large number of glacier outlines were either corrected / deleted / added or freshly created. Figure 1 provides an overview of the reworked outlines in RGI 7.0. More details about the data processing in C3S can be found in the ATBD (Paul et al. 2025b).

Figure 1. Glacier coverage and the 19 first order regions in RGI v7.0 showing regions of improvement in red and unchanged regions in blue. The image is taken from Maussion et al. (2023).

Additionally, internal drainage divides have been adjusted in many regions using more recent digital elevation models (DEMs). Most of the new outlines have been derived from satellite images using automated methods to map clean ice (e.g. band ratio) and manual editing to include omission (debris cover, clouds, ice in shadow) and exclude commission (lakes, seasonal snow, rock in shadow) errors. Examples and further details are described in the PQAR (Paul et al. 2025a). However, in some regions glacier outlines were completely digitized manually (e.g. High Mountain Asia) and in other regions the existing datasets were not updated and are partly based on digitized topographic maps (e.g. USA). Hence, all glacier outlines are quality checked against a reference dataset, but the applied corrections vary with the experience of the analyst and interpretation differences exist (see examples in PQAD, Paul et al. 2023).

A DEM is the key auxiliary dataset used to obtain (a) drainage divides (e.g. Bolch et al. 2010, Kienholz et al. 2013) to separate glacier complexes into individual glaciers, and (b) to calculate topographic information (e.g. min, max, mean and median elevation, slope and aspect, hypsometry) for each glacier (e.g. Paul et al. 2002 and 2009). As a range of different DEMs (e.g. ASTER Global Digital Elevation Model (GDEM), Shuttle Radar Topography Mission (SRTM), AW3D30 and the Copernicus Global DEM) have been used to derive both (a) and (b), there is also some variability in the position of joint glacier boundaries and topographic parameters (Frey and Paul 2012). In any case, the derived glacier outlines are categorical variables representing a Level 3 Earth Observation data processing.

In the present document there will be no details regarding the individual datasets used for RGI 7.0, as the related information can be found in the description of each first-order region in the RGI 7.0 User Guide that is available online (Maussion et al. 2023). This document also lists for all regions covered its contributors, the satellite data and processing methods used, and the history of updates since RGI 1.0. A summarised description of the RGI history and its contents can be found in the study by Pfeffer et al. (2014). This overview paper relates to RGI v3.2 but is also valid for later versions. It also provides details of the applied measures for quality control, the merging of the contributing datasets and known limitations of the dataset. For a thorough description of RGI 7.0 and its new way of creation a publication is in preparation.

Figure 2. Timeline of freely available satellite sensors (1982-2025) used to create glacier outlines. After May 2003, the ETM+ sensor only provided scenes with striping due to failure of the scan line corrector (SLC off scenes). The ETM+ sensor was decommissioned in June 2025. The ASTER SWIR sensor failed in April 2008.

Product Target Requirements

The RGI covers all regions with glaciers. To ease data handling, the entire dataset has been subdivided into 19 first-order regions (Figure 1). The temporal coverage is variable, nonetheless the majority of outlines are from within a 10 year period centered on the year 2003 (77% of the datasets are now from the year 2000 ±5 years, in RGI 6.0 it was 65%). About 60,000 glaciers mostly smaller than 1 km² have been added to RGI v7.0, resulting in a more reasonable size-class distribution compared to theoretical assumptions (Figure 3). Random and systematic errors have not been determined regularly for the provided datasets, but can be found in the PQAD (Paul et al. 2023) for the RGI and in the PQAR (Paul et al. 2025a) for datasets provided by C3S.

Figure 3. Size-class distribution in RGI 6.0 (blue line) and 7.0 (red line). The image is taken from Maussion et al. (2023).

It has also to be noted that systematic errors (omission and commission) are removed by the analyst during manual editing (see Section 1.1). The remaining random errors are related to a wide range of sources (interpretation, geolocation, pixel size, contrast) and a range of methods have been applied to assess them (e.g. Paul et al. 2017). A condensed overview of the error terms and a generalized assessment for the datasets is presented in Pfeffer et al. (2014).

In Table 1 we provide the list of technical requirements for the glacier area product according to the Global Climate Observing System (GCOS) requirements for the Essential Climate Variable (ECV) 'Glaciers' from GCOS (2022). Most important here is the required uncertainty of the glacier outlines (better than 5%) which can only be reached by manual editing as e.g. debris-covered glacier parts are not included by the automated mapping and can easily cover 50% or more of the glacier area. Table 2 shows if the products created (Paul et al. 2025a) meet the GCOS requirements.

Table 1. GCOS requirements for the ECV Glaciers (from GCOS 2022, Section 8.2.1).

Table 2. Compliance of the glacier area product with GCOS requirements. The shading in the column 'Reported value' indicates which threshold requirement (listed in columns 2, 3, and 4) the dataset meets: Green: "goal", orange: "threshold", yellow: "break-through".

Example visualization 

Here we illustrate with selected examples what the vector and raster datasets look like and how they can be displayed. Both types of data can be visualized using the open software QGIS or special software for netCDF files (e.g. ncview or Panoply). Vector and GeoTIF files can also be displayed and manipulated with the widely used but commercial software ArcGIS from ESRI. The vector data are stored in the open shape file format which stores projection information, coordinates of the vector data (in the case of glaciers polygons) and the attribute information related to each polygon in separate files with the ending .prj, .shp and .dbf, respectively. The .dbf file can also be opened with other software and exported to other formats for further analysis (e.g. .csv or .xls). The examples below are screenshots from QGIS (Table 3), ArcGIS (Figure 4a and b) and Xmgrace (Figure 5a and b).

Table 3 illustrates a subset of the information stored for each glacier in the attribute table of the shape file. The vector outlines of the glaciers can be shown on top of other geocoded datasets (e.g. satellite images, DEM hill shades) but one can also colour-code the range of values for a specific attribute within the glacier outlines (Figure 4a) or as filled circles (Figure 4b). The values can also be arranged in histograms, for example to show the size-class distribution of glaciers in a region (Figure 5a) or in scatterplots (Figure 5b), revealing possible dependencies among selected attributes. The topographic information can be used to derive several further glacier characteristics (e.g. Haeberli and Hoelzle 1995) or climatic information (Sakai et al. 2015).

Table 3. Screenshot of the information provided in the attribute table for each glacier. The lower panel is the right continuation of the upper panel.

Figure 4. a) Colour-coded mean glacier aspect for glaciers in the north of Baffin Island (Canada). b) Median elevation of glaciers in the Alps shown as colour-coded circles (from Paul et al. 2020).

Figure 5. a) Histogram showing the size class distribution of glaciers in the Alps. b) Glacier aspect vs. mean elevation for glaciers in the Alps (both panels from Paul et al. 2020).

The raster version of RGI 7.0 (available in netCDF format), showing the percentage of glacier coverage per grid cell, is displayed in Figure 6 for the 0.5° resolution version. This dataset is also available at 1° and 0.1° resolution. A further variable included in each dataset (but not shown here) is the glacier area per grid cell in km². As mentioned above, the glacier extents in RGI 7.0 refer as close as possible to the year 2000, but there is still a wide spread in particular for regions that have not been updated since RGI 6.0. These partly much older (back to 1943) or also newer (up to 2021) outlines are usually only referring to a few selected glaciers (see RGI user guide by Maussion et al. 2023) for full details for each of the 19 regions). Hence, for the regionally averaged raster product the impact of the inconsistent timing on the total area per grid cell for the year 2000 is small. As a further note, the glacier are in km² per grid cell has been calculated with an approximation, assuming that the Earth is a sphere. As the general shape of the Earth is an ellipsoid with regional variability (geoid), the calculated area per grid cell deviates from the values provided by the RGI (that was calculated in the local UTM projection of each glacier).

Both, the vector and the raster datasets do not have any uncertainties or quality flags associated with them and are provided as is. However, for the individual datasets contributing to any version of the RGI, related publications usually provide this information (e.g. Paul et al. 2020). For the entire RGI, the overview paper by Pfeffer et al. (2014) and the PQAD (Paul et al. 2023) provides this information.

Figure 6. Colour-coded version of the rasterized RGI 7.0 presenting the percentage of glacier cover per grid cell at a 0.5° resolution. The image in the background is the ESRI Basemap.

Data usage information

Data format and file naming 

From the datasets available for RGI 7.0 we have brokered for the Copernicus CDS the 'Glacier product', i.e. the individual glacier polygons separated by ice divides. From this product we have obtained the shape files (glacier outlines) and the csv files describing the hypsometry (area distribution with elevation) of each glacier. The glacier outlines for all 19 regions were merged by the C3S Glacier Service into one large shape file, while the hypsometry files are kept separated per region but were merged into one zip file. The raster versions providing the percentage of coverage and glacier area per grid cell in three different resolutions (0.1°, 0.5° and 1.0°) have been derived from the RGI 7.0 shape files by the C3S Glacier Service and are available in netCDF format. Both the vector and raster version of RGI 7.0 are provided in geographic projection using the WGS84 datum. For further details on file naming conventions the reader is referred to Chapter 3 of the RGI 7.0 user guide (Maussion et al. 2023).

Quality flags and data masks 

As mentioned above, both datasets do not have any quality flags and are provided as is. Data masks are not used.

File contents

The shape file format consists of several sub-files with the same name but different endings. The .shp file contains the coordinates of the glacier polygons, the .dbf file stores the attribute information for each polygon and the .prj file the projection information (EPSG:4328). An example (screen shot) of the .dbf file contents is shown in Table 3, the shape file itself (the outlines) are shown in Figure 3a, colour-coded according to the entries of one item in the .dbf file. A detailed description of all attributes for the 'Glacier product' of RGI 7.0 is available in Section 3.5 of the user guide (Maussion et al. 2023).

The raster file in netCDF format and EPSG:4328 projection contains for each of the three resolutions the percentage of glacier coverage and the related area (in km²) per grid cell. The variable is named 'Glacier_area' and the units of the variables are % and km². The latter has been derived by multiplication of the percentage with the area of each grid cell that has been derived by assuming a spherical Earth. As the Earth does not have the shape of a sphere but is an ellipsoid with larger and random regional scale deviations, this simplification introduces a systematic and random bias. In consequence, the glacier areas are different from those listed in RGI (which were calculated in the local UTM zone with WGS84 datum).

Examples of known climate applications and best practices

Climate applications of the RGI (version 6.0) are numerous. Although it was recommended to only use the RGI for global to continental scale studies, it has also been widely used at regional and even local scales. Accordingly, the first publication about the RGI (Pfeffer et al. 2014) is now the most cited ever published in the Journal of Glaciology. Apart from the fact that glacier outlines are required for all kinds of glacier-specific applications (e.g. elevation changes and mass balance, flow velocities, ice thickness distribution and glacier volume) they also serve as a base for modelling future glacier evolution, run-off and sea level contribution under climate change or change assessment in general. They further help to identify stable terrain off glaciers which is required for the co-registration of DEMs (that has to be performed before calculating geodetic glacier mass balance from DEM differencing) or the calculation of uncertainties off-glaciers, e.g. for glacier flow velocities.

The raster version of the RGI was created for application in climate models that can consider the percentage of glacier coverage in a specific region when calculating energy and mass fluxes with land surface schemes adopted by climate models (e.g. water, soil, forest, urban, ice). So far, we are not aware of specific applications of the dataset for this purpose.

Known Issues and Limitations

The vector dataset of RGI 7.0 is currently the best dataset of global glacier outlines. However, it still has a number of issues that were partly already mentioned above. The most relevant are:

1. The target year 2000 (± a few years) was not achieved in many regions, the time span is more 1990 to 2010, partly even older or more recent. In part this is due to missing datasets from this period or due to missing resources to update further inventories.
2. Several glaciers and drainage divides are missing or could be at the wrong location.
3. The quality of the outlines and what was considered to be a glacier is variable and not consistent through all regions. This concerns a variable minimum size threshold, the interpretation of debris-covered glacier parts, ice-cored moraines and ice patches of unclear nature, the inclusion of seasonal snow and missing rock outcrops.
4. In polar regions the separation of glaciers from connected ice shelves could be wrong.

Accordingly, there are some limitations when working with this dataset. Most relevant are:

• The RGI has been created for continental to global scale applications and might have quality issues at the regional or local scale.
• In the same sense, it is not recommended to use the RGI as a base for change assessment at the regional scale without checking its quality and the rule-set applied for the mapping. Otherwise glacier area changes might be due to the different interpretation rather than climate change.
• For applications such as the calculation of ice thickness distribution it is recommended to use the dataset without ice divides, i.e. the glacier complexes.

The raster version of the RGI is provided to facilitate applications that can consider glaciers as a separate surface type in climate models. The values provided give the glacier cover per grid cell in per cent and square kilometres. The latter values have systematic and random errors as the size calculation is based on the wrong assumption that the Earth is a sphere.

Data access information

The RGI 7.0 (doi.org/10.5067/f6jmovy5navz) provided to the CDS has been downloaded from https://nsidc.org/data/nsidc-0770/versions/7 (last access: 24.11.2025) and is described in a related documentation (Maussion et al. 2023). Shape files as well as the raster dataset can be visualised by a range of commercial (e.g. ESRI ArcGIS) and freely available (e.g. QGIS) software packages. Data are made available through the Copernicus Climate Data Store (CDS). Registration is free but a login is required to access the CDS and its toolbox software suite. Data can be downloaded from the website and used under the License CC-BY 4.0 (included in download page), i.e. when using the RGI v7.0 dataset, the publication RGI 7.0 Consortium (2023) has to be cited.

User Support

The front-end of the user support is provided by ECMWF, please contact the ECMWF Support Portal if there is a question related to the documents or datasets. Any support on specific technical aspects (Level 2 support) concerning the provided products will be forwarded to the team responsible for the Glacier Service.

References

Bolch, T., B. Menounos, and R. Wheate (2010): Landsat-based glacier inventory of western Canada, 1985–2005. Remote Sensing of Environment, 114, 127-137, doi:10.1016/j.rse.2009.08.015

Frey, H. and Paul, F. (2012): On the suitability of the SRTM DEM and ASTER GDEM for the compilation of topographic parameters in glacier inventories. International Journal of Applied Earth Observation and Geoinformation, 18, 480-490; doi.org/10.1016/j.jag.2011.09.020

GCOS (2022): The 2022 GCOS ECVs Requirements. GCOS-245. WMO, Geneva, 244 pp. Online: https://library.wmo.int/index.php?lvl=notice_display&id=22135

Haeberli, W. and Hoelzle, M. (1995): Application of inventory data for estimating characteristics of and regional climate-change effects on mountain glaciers: A pilot study with the European Alps. Annals of Glaciology, 21, 206–212; doi.org/10.3189/S0260305500015834

Kienholz, C., R. Hock and A.A. Arendt (2013): A new semi-automatic approach for dividing glacier complexes into individual glaciers. Journal of Glaciology, 59 (217), 925-937; doi.org/10.3189/2013JoG12J138

Maussion, F., Hock, R., Paul, F., Raup, B., Rastner, P., Zemp, M, Andreassen, L., Barr, I., Bolch, T., Kochtitzky, W., McNabb, R. and Tielidze, L. (2023): The Randolph Glacier Inventory version 7.0 User guide v1.0. doi.org/10.5281/zenodo.8362857.

Paul, F., Kääb, A., Maisch, M., Kellenberger, T. W. and Haeberli, W. (2002): The new remote-sensing-derived Swiss glacier inventory: I. Methods. Annals of Glaciology, 34, 355-361; doi.org/10.3189/172756402781817941

Paul, F., R. Barry, J.G. Cogley, H. Frey, W. Haeberli, A. Ohmura, S. Ommanney, B.Raup, A. Rivera, M. Zemp (2009): Recommendations for the compilation of glacier inventory data from digital sources. Annals of Glaciology, 50 (53), 119-126; doi.org/10.3189/172756410790595778

Paul, F., T. Bolch, K. Briggs, A. Kääb, M. McMillan, R. McNabb, T. Nagler, C. Nuth, P. Rastner, T. Strozzi, J. Wuite (2017): Error sources and guidelines for quality assessment of glacier area, elevation change, and velocity products derived from satellite data in the Glaciers_cci project. Remote Sensing of Environment, 203, 256-275; doi.org/10.1016/j.rse.2017.08.038

Paul, F., Rastner, P., Azzoni, R.S., Diolaiuti, G., Fugazza, D., Le Bris, R., Nemec, J., Rabatel, A., Ramusovic, M., Schwaizer, G., and Smiraglia, C. (2020): Glacier shrinkage in the Alps continues unabated as revealed by a new glacier inventory from Sentinel-2. Earth Systems Science Data, 12(3), 1805-1821; doi.org/10.5194/essd-12-1805-2020

Paul, F. et al. (2023) C3S Glacier Distribution Service version 6.0: Product Quality Assurance Document. Copernicus Climate Change Service. Document ref.: C3S2_312a_Lot4.WP2-FDDP-GL-v1_202212_A_PQAD-v4 (Online: https://confluence.ecmwf.int/pages/viewpage.action?pageId=369463069)

Paul, F. et al. (2025a) C3S Glacier Service version 6.0: Product Quality Assessment Report (PQAR). Copernicus Climate Change Service. Document ref.: C3S2_313d_WP1-DDP-GL-v1_202511_A_PQAR-v6_i0.1 (Online: https://confluence.ecmwf.int/pages/viewpage.action?pageId=589474326)

Paul, F. et al. (2025b) C3S Glacier Service version 6.0: Algorithm Theoretical Baseline Document (ATBD). Copernicus Climate Change Service. Document ref.: C3S2_313d_WP1-DDP-GL-v1_202511_A_ATBD-v6_i0.1 (Online: https://confluence.ecmwf.int/pages/viewpage.action?pageId=589474321)

Pfeffer, W.T., 18 others and the Randolph Consortium (2014): The Randolph Glacier Inventory: a globally complete inventory of glaciers. Journal of Glaciology, 60 (221), 537-552; doi.org/10.3189/2014JoG13J176

RGI 7.0 Consortium (2023): Randolph Glacier Inventory - A Dataset of Global Glacier Outlines, Version 7.0. Boulder, Colorado USA. NSIDC: National Snow and Ice Data Center. doi.org/10.5067/f6jmovy5navz (last access: 23 Nov 2025).

Sakai, A., Nuimura, T., Fujita, K., Takenaka, S., Nagai, H., and Lamsal, D. (2015): Climate regime of Asian glaciers revealed by GAMDAM glacier inventory. The Cryosphere, 9, 865–880; doi.org/10.5194/tc-13-2043-2019

Vaughan, D.G., J. C. Comiso, I. Allison, J. Carrasco, G. Kaser, R. Kwok, P. Mote, T. Murray, F. Paul, J. Ren, E. Rignot, O. Solomina, K. Steffen and T. Zhang (2013): Observations: Cryosphere. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the IPCC. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, pp. 317-382. (Online: ipcc.ch/report/ar5/wg1)

Related articles

• Page:
  Glacier Area data version 7.0: Product Quality Assessment Report (PQAR)
• Page:
  Glacier Area data version 7.0: Algorithm Theoretical Basis Document (ATBD)
• Page:
  Glacier Area data version 7.0: Product User Guide and Specification (PUGS)
• Page:
  SEC Antarctica Version 5.0 and Greenland Version 6.0: Product User Guide and Specification (PUGS)
• Page:
  SEC Antarctica Version 5.0 and Greenland Version 6.0: Product Quality Assessment Report (PQAR)