Contributors: University of Zurich: Jacqueline Bannwart, Inés Dussailant, Frank Paul, Michael Zemp
Issued by: UZH/Frank Paul
Date: 28/05/2024
Ref: C3S2_312a_Lot4.WP2-FDDP-GL-v2_202312_A_PUGS-v5_i1.2
Official reference number service contract: 2021/C3S2_312a_Lot4_EODC/SC1
History of modifications
List of datasets covered by this document
Related documents
Acronyms
General definitions
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 distribution service, the Randolph Glacier Inventory (RGI) is brokered for the CDS from https://nsidc.org/data/nsidc-0770/versions/7 under a CC-BY 4.0 license.
Crevasse: A crevasse is a wedge-shaped crack in the surface of a glacier caused by differential movement (faster towards the centre, slower towards the boundary) of the ice. A crevasse can be several metres wide and more than 30 m deep.
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 the image pixel size.
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.
Nominal glaciers: These are based on tabular data (longitude / latitude coordinates and size) from the World Glacier Inventory (WGI) and are shown in RGI 6.0 (and earlier versions) at the coordinates of the WGI as circles with an area that represents their size.
Scope of the document
This document is the Product User Guide and Specification (PUGS) for the glacier area (outlines) product from the glacier distribution service. It describes the Climate Data Record (CDR) brokered to the CDS from the Randolph Glacier Inventory (RGI) version 7.0 in its vector (shape file) and raster (netCDF) formats. For the detailed product description and usage, we refer to the freely available reference documents [RD1] and [RD2].
Executive summary
This document provides a description of the Climate Data Record (CDR) dataset provided by the Copernicus Climate Change Service (C3S) Glacier Distribution Service to the Climate Data Store (CDS): the 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.
In Section 1, we first provide a description of the datasets and methods used to create the RGI, before we give an overview on its spatio-temporal coverage and usage limitations. Details of the technical specifications and data fields are described in the following Section 2, including some illustrations visualizing the information provided. Information on data access is provided in Section 3.
1. The Randolph Glacier Inventory (RGI)
1.1. Product description
The RGI is a globally almost complete collection of digital glacier outlines 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 Intergovernmental Panel on Climate Change Fifth 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 (glims.org) and additional outlines derived from mostly Landsat but also Advanced Spaceborne Thermal Emission and reflection Radiometer (ASTER) satellite images. In a few regions the outlines are also based on digitised topographic maps, or nominal data from the World Glacier Inventory (WGI) are used and 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 from scratch. Figure 1 is providing an overview of the reworked outlines in RGI 7.0. Additionally, internal drainage divides have been adjusted in many regions using more recent digital elevation models (DEMs). The fresh 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 (see [RD5] for details). Hence, all glacier outlines are quality checked against a reference dataset but the applied corrections are based on the experience of the analyst and interpretation differences exist (see Product Quality Assurance Document (PQAD), [RD3]). 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) calculate topographic information (e.g. min, max, mean, 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), (ALOS) World 3D - 30m (AW3D30) and the Copernicus DEM ( COP 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 (EO) data processing.
In the present document there will be no details regarding the individual datasets used for the RGI. Instead, related information can be found in the corresponding description for each first-order region in the RGI User Guide that is available online [RD1]. 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 and its contents can be found in the freely available study by Pfeffer et al. (2014, see [RD2] for the DOI). This overview paper relates to RGI v3.2 but is also valid for more recent versions (up to RGI v6.0), as it also provides details on the RGI history, the merging of the contributing datasets, the applied measures for quality control, and known limitations of the dataset.
For a thorough description of RGI 7.0 and its new way of creation a publication is currently in preparation.
1.2. Target requirements
The RGI covers all regions with glaciers globally. To ease data handling, the entire dataset has been subdivided into 19 first-order regions (Figure 1). The temporal coverage is highly variable but clusters now much better around the years 2000 (77% of the datasets are now from the year 2000 ±5 years, in RGI 6.0 it was 65%). About 60,000 glaciers smaller than 1 km2 have been added to RGI v7.0 resulting in a more realistic size-class distribution (Figure 2). Random and systematic errors have not been determined regularly for the provided dataset, but can be found in the Product Quality Assessment Report (PQAR) [RD4] for the datasets provided by C3S. It has also to be noted that systematic errors (omission and commission, see Section 1.1) are removed by the analyst during manual editing. 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. A condensed overview of the error terms and a generalised assessment for the datasets is presented in [RD2] and will be described in the forthcoming publication for RGI 7.0.
In the following, 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’ (Table 1) that is taken 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.
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 [RD1].
Figure 2: Size-class distribution in RGI 6.0 (blue line) and 7.0 (red line). The image is taken from [RD1].
Table 1: GCOS requirements for the ECV Glaciers.
Methods to obtain the uncertainty of the digitized glacier outlines are described by Paul et al. (2013 and 2017).
1.3. Data usage information
The technical details about the RGI (e.g. data format, file names, product content, attributes, quality indicators, projection etc.) are summarized in Section 2. Usage restrictions can be found in [RD2]. Key points to consider are:
- The RGI is a vector dataset in shape file format consisting of about 275,000 glaciers with related attribute information and comes in a geographic projection with WGS1984 datum (EPSG code 4326).
- The RGI has been created for continental to global-scale applications and might, thus, have regional or local quality issues. The C3S glacier distribution service has considerably helped to to improve on the latter.
- Whereas the GLIMS database (from which the RGI is derived) is multi-temporal and includes all datasets provided by the community, the RGI is a frozen snapshot including only one outline from each glacier.
- The new RGI 7.0 also includes a dataset where all glacier divides were removed resulting in so-called glacier complexes. This dataset has been used to create the raster product.
- It is not recommended to use the RGI for change assessment without carefully checking its quality in the region of interest.
The raster version of the RGI is provided to facilitate applications which can consider glaciers as a separate surface type, but cannot work with the vector data (shape files) provided so far. Details of the related vector-raster conversion are described in the Algorithm Theoretical Basis Document (ATBD) [RD5]. The raster version of the RGI is provided in netCDF format. The values give the relative glacier cover per grid cell in square kilometres and percent. The latter is the most common way of integrating this information, e.g. in a land surface scheme assimilated by climate models.
2. Technical details of the provided product
2.1. Technical Specification of the dataset
The Copernicus Glacier Distribution Service is providing two types of datasets: (A) glacier outlines with attribute information in a vector format (shape files with polygon topology) and (B) a gridded version of the RGI in raster format (netCDF) containing the percentage of glacier coverage per grid cell. Both datasets are provided in geographic coordinates (longitude and latitude, in degrees), which are referenced to the WGS84 datum. For a description of the product files and their contents for the vector product (A) we refer to Section 3.5 in [RD1] or the online document available here: http://www.glims.org/rgi_user_guide/products/glacier_product.html
.
Each glacier outline from (A) conforms to the data-model conventions of ESRI ArcGIS shape files. That is, each glacier consists of an outline encompassing the glacier, followed immediately by outlines representing all of its nunataks (ice-free areas enclosed by the glacier). For each outline, successive vertices are ordered such that glacier ice is on the right, i.e. the outline has a clockwise rotation.
The information provided with the raster version of the RGI is given in Table 2. The value given is the percentage of glacier cover per grid cell.
Table 2: Overview of the attribute data provided with the raster version of the RGI.
| Item (short) | Item (full) | Unit | Format |
|---|---|---|---|
| Value | Relative glacier coverage of grid cell area | % | Float |
2.2. Product examples
Below, we illustrate with selected examples how the vector datasets look like and how they can be displayed. Table 3 illustrates a subset of the information stored for each glacier in the attribute table of the shape file. Apart from just overlaying glacier outlines on a satellite image, one can also colour-code the range of values for a specific attribute within the glacier outlines (Figure 3a) or as filled circles (Figure 3b). The values can also be arranged in histograms, for example to reveal the size-class distribution of glaciers in a region (Figure 4a) or in scatterplots (Figure 4b), 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. Click on the table for full size image.
Figure 3: a) Colour-coded mean glacier aspect for glaciers of the Jostedalsbreen Ice Cap in southern Norway. The abbreviations in the legend stand for the terrain aspect orientation: N for North, NE for North-East, E for East, SE for Sout-East, S for South, SW for South-West, W for West and NW for Noth-West. b) Median elevation of glaciers in the Alps represented as colour-coded circles (from Paul et al. 2020). Click on the figure for full size image.
Figure 4: 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).
2.5. Data sources
All glacier outlines in the RGI are extracted from the GLIMS glacier database to which they are provided by a global community (e.g. Kargel et al. 2005) who derived them from a variety of sources (see Pfeffer et al. 2014). Whereas a main input data source are satellite data (Landsat, ASTER, Satellite Pour l’Observation de la Terre (SPOT)), analysts have also digitized outlines from topographic maps, aerial photography and other sources. Whereas GLIMS is a multi-temporal database containing all outlines being made available, the RGI is a snap-shot in time referring only to one dataset.
As a general rule, outlines are derived from satellite scenes acquired around the year 2000 (from Landsat Enhanced Thematic Mapper Plus (ETM+)) to have a good temporal match with the SRTM DEM (that was acquired in February 2000). A common date of all outlines would also be desirable from a modelling perspective, but in reality the outlines in the RGI span a time period of 50 years (1960 to 2010) centred around the years 2000 and 2010. For the new RGI 7.0 the 'closer to the year 2000' requirement was achieved by having 12% more outlines acquired in 2000 ±5 years.
For RGI 7.0, extensive quality control of existing or newly submitted datasets was performed. Glacier outlines from RGI 6.0 have been replaced when datasets with a better quality and/or closer to the year 2000 have been made available by the community (i.e. they were already in GLIMS or provided to GLIMS after a request) after RGI 6.0 was published. This allowed consideration of new national inventories for the Caucasus, Chile and Argentina as well as new regional datasets for Arctic Canada North, Kamchatka, the Pyrenees and High Mountain Asia. For other regions (Alaska, Canadian Arctic South, Greenland, Kerguelen, New Zealand) existing datasets were modified to improve their quality and/or bring them closer to the year 2000. Fresh datasets were created for Peru/Bolivia and Jan Mayen within the framework of C3S.
Glacier outlines are not separated into dynamically or hydrologically separate units when derived from satellite images by automated mapping, but the resulting glacier maps are glacier complexes, i.e. collections of contiguous glaciers. For the datasets provided by C3S to RGI 7.0 we used semi-automated algorithms (Bolch et al. 2010, Kienholz et al. 2013) to separate the glacier complexes into individual glaciers using a watershed algorithm applied to the most appropriate DEM available. The algorithm output requires intense manual checking and corrections of the generated flow divides before they can be digitally intersected with the glacier complexes. We created these divides freshly from a range of DEMs (e.g. ArcticDEM, COP DEM, and Advanced Land Observing Satellite (ALOS) World 3D - 30m (AW3D30)) for the regions where we provided new outlines (e.g. Peru/Bolivia, Jan Mayen, Kerguelen, New Zealand) and edited them in several of the other new regions (e.g. Alaska, Canadian Arctic North and South, Greenland, Kamchatka, Southern Andes).
2.6. Data Citation requirement
The following reference must be cited when using the RGI v7.0 dataset: RGI 7.0 Consortium (2023) [RD1].
3.Data access information
The RGI v7.0 provided to the CDS has been downloaded from https://nsidc.org/data/nsidc-0770/versions/7 and is described in a related documentation [RD1]. 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 will be made available through the Copernicus Climate Data Store (CDS). Registration (free) is required to access the CDS and its toolbox software suite. Data can be downloaded from the website and used under the License (included on download page). All requests for information or further data should be channelled through the CDS Knowledge Base.
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.
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.
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 (last access: 9 Feb 2024)
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.
Kargel, J. S., Abrams, M. J., Bishop, M. P., Bush, A., Hamilton, G., Jiskoot, H., ... & Wessels, R. (2005). Multispectral imaging contributions to global land ice measurements from space. Remote Sensing of Environment, 99 (1-2), 187-219
Kienholz, C., R. Hock and A.A. Arendt (2013): A new semiautomatic approach for dividing glacier complexes into individual glaciers. Journal of Glaciology, 59 (217), 925-937.
Kienholz, C., J.L. Rich, A.A. Arendt and R. Hock, 2014, A new method for deriving glacier centerlines applied to glaciers in Alaska and northwest Canada. The Cryosphere, 8, 503–519.
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.
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.
Paul, F., N. Barrand, E. Berthier, T. Bolch, K. Casey, H. Frey, S.P. Joshi, V. Konovalov, R. Le Bris, N. Mölg, G. Nosenko, C. Nuth, A. Pope, A. Racoviteanu, P. Rastner, B. Raup, K. Scharrer, S. Steffen and S. Winsvold (2013): On the accuracy of glacier outlines derived from remote sensing data. Annals of Glaciology, 54 (63), 171-182.
Paul, F., S.H. Winsvold, A. Kääb, T. Nagler and G. Schwaizer (2016): Glacier Remote Sensing Using Sentinel-2. Part II: Mapping Glacier Extents and Surface Facies, and Comparison to Landsat 8. Remote Sensing , 8(7), 575; doi:10.3390/rs8070575.
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 (last access: 9 Feb 2024)
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.
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.
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: 9 Feb 2024).
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.
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.
Zemp, M., Thibert, E., Huss, M., Stumm, D., Rolstad Denby, C., Nuth, C., Nussbaumer, S.U., Moholdt, G., Mercer, A., Mayer, C., Joerg, P.C., Jansson, P., Hynek, B., Fischer, A., Escher-Vetter, H., Elvehøy, H., and Andreassen, L.M. (2013): Reanalysing glacier mass balance measurement series. The Cryosphere, 7, 1227-1245.
