Closed ezwelty closed 7 months ago
Thanks for this @ezwelty , much appreciated. You are correct that computing the links regionally is wrong - there may be overlap with neighboring regions when two are close, regardless of changes in regions between versions.
The data is now corrected and linked here: https://github.com/GLIMS-RGI/rgi_issue_tracker/issues/41
Related statistics:
Updates the text and adds a table to describe the
_links.csv
files. See #83.p.s. @fmaussion I notice that you broke up the overlaps by region. If I recall correctly, glaciers from some regions in RGI 7.0 overlap those of other regions in RGI 6.0, and thus why I recommended the links be processed globally. Were the overlaps (and the corresponding
cluster_id
,n_rgi6
,n_rgi7
) computed globally and then split (how?) into region files? Iscluster_id
globally unique?