Merging New Data into GLIMS: Examples from the Himalaya and the RGI

The Global Land Ice Measurements from Space (GLIMS) glacier database was built at the National Snow and Ice Data Center (NSIDC) in 2005, and now contains outlines and metadata for 120,000 glaciers. GLIMS is one of the most popular data sets at NSIDC, and a web-based map interface and web map services allow users to obtain the data at no cost. As new glacier outlines are produced for glaciers which have already been mapped within GLIMS, we must ensure that each new outline is assigned the same GLIMS glacier ID as its previous outline. Otherwise, outlines that are supposed to pertain to the same glacier will appear to be different glaciers, causing errors in summary statistics of the database, such as glacier count or area. New sets of glacier outlines, including 12000 outlines from the Western Himalaya and 3500 outlines from New Zealand, have recently been merged into GLIMS. These outlines had significant overlap with existing outlines in the GLIMS database, necessitating new approaches to the merging process. The New Zealand outlines came from the Randolph Glacier Inventory (RGI), a data set created with the express purpose of filling the geographic gaps in GLIMS to produce a globally complete map of glaciers. The RGI lacks the rich metadata (e.g. data sources, time stamps) of GLIMS. More data from the RGI, such as from Arctic Canada and the periphery of Greenland, are expected to be merged into GLIMS as resources at NSIDC allow and as sufficient metadata can be obtained. This contribution discusses the status of the GLIMS Glacier Database and the merge of RGI data into GLIMS, showing how the merge is carried out. As of August, 2013, the GLIMS Glacier Database contains approximately 70% of the contents of the RGI, by both glacier count and area.