---
license: cc
---

# Cluster-wise Regression Reveals Drivers of Present-Day Glacier Changes

TODO: add link paper

This dataset contains glacier-level data used for cluster-wise glacier mass-balance regression.

## CSV data

`df_mb_with_morpho_and_clim_GSWP3_W5E5.csv` contains one row per glacier, indexed by RGI ID. It includes selected morphological predictors, climate anomalies, location/region metadata, projected coordinates, observed mass balance (`mb`), and mass-balance uncertainty (`mb_err`). The main predictors used here are:

- `slope_20prc`
- `z_mean`
- `area_debris_prc`
- `norm_contrib_area`
- `temp_anom_2000_to_2019`
- `prcp_anom_2000_to_2019`

## Cluster GeoPackages

The files in `clusters_geocoded/` contain geocoded glacier points for each analysis region. Each file includes the glacier attributes above, a selected `cluster` label, and the full set of available clustering labels (`clabels_*`) for different numbers of clusters.

Cluster labels were generated with the `SpatialRegWard` method. The GeoPackage geometries are point locations reprojected to the local CRS used for each region.