Supplementary Information for Herreid and Pellicciotti, Nature Geoscience, 2020

Data S1:  

For each RGI region (all files are in ArcGIS format .shp): 

1.    Updated RGI v6.0 glacier area (>1 km2)
2.    Debris map (glacier area >1 km2) where debris shapes are cut by source satellite image edges and assigned the satellite image acquisition date.   
3.    Updated RGI v6.0 glacier area (>2 km2) with per glacier metrics and constituent quantities in the attribute table
4.    Debris map (glacier area >2 km2)
5.    Ablation zone area
6.    Equilibrium line
7.    Debris expansion line
8.    RGI v6.0 false positive area
9.    RGI v6.0 false negative area

Data S2:

1.    Landsat image footprints with NASA IDs

Data S3:  

For each RGI region (all files are in ArcGIS format .shp): 

1.    This study’s results compared to [1] false positive area
2.    This study’s results compared to [1] false negative area
3.    This study’s results compared to [1] true positive area

For High Mountain Asia only (RGI regions 13, 14 and 15 merged):

4.    This study’s results compared to [2] false positive area
5.    This study’s results compared to [2] false negative area
6.    This study’s results compared to [2] true positive area

References

1.    Scherler, D.,Wulf, H. & Gorelick, N. Global assessment of supraglacial debris‐cover extents. Geophys. Res. Lett. 45, 11798-11805 (2018).
2.    Kraaijenbrink, P. D. A., Bierkens, M.F.P., Lutz, A.F. & Immerzeel, W.W. Impact of a global temperature rise of 1.5 degrees Celsius on Asia’s glaciers. Nature 549, 257-260 (2017).

Data S1. 
S1 consists of a file structure where each RGI region (excluding Antarctica) has a subdirectory containing 9 files (05Greenland contains 7 files) in the format .shp. The naming convention of each file begins with the RGI region number and name, e.g. “01Alaska”, followed by a shorthand name as defined below:  

1.    “_minGl1km2.shp” An amended RGI v6.0 where glaciers with a surface area less than 1 km2 are removed, area identified as false positive is removed and areas identified as false negative is added. These errors were manually identified by one individual (Sam Herreid) for all regions on Earth. 

The attribute table of this file include unaltered RGI v6.0 fields [3, 4]: 
[RGIId], [GLIMSId], [RGIFlag], [01Region], [02Region], [Zmin], [Zmax], [Zmed], [Slope], [Aspect], [Lmax], [GlacType] and [Name].

New fields calculated with the amended geometries are:
[totGl_km2] = Total glacier area within this region (km2)
[totDeb_km2] = Total debris-covered area within this region (km2)
[spatDo_km2] = Total area of the spatial domain considered for this region in this study (km2)
[perc_deb] = Region-wide percent debris-covered area: [totDeb_km2]/[spatDo_km2]
[FP_km2] = Total RGI v6.0 false positive area within this region (km2)
[FN_km2] = Total RGI v6.0 false negative area within this region (km2)
[FP_perc] = Region-wide percent false positive area: [FP_km2]/[spatDo_km2]
[FN_perc] = Region-wide percent false negative area: [FP_km2]/[spatDo_km2]
[regionFrac] = Fraction of altered regional total area considered, i.e. 1-[regionFrac] is glacier area meeting the criteria: glacier area >1 km2, but falls outside of the satellite image composite (Fig. S1): [spatDo_km2]/[totGl_km2]
[RGI6consid] = Fraction of the total, unaltered RGI v6.0 area (including glaciers smaller than 1 km2) considered in this study: [spatDo_km2]/raw RGI v6.0 area (calculated using the projections listed in Table S1) 

2.    “_minGl1km2_debrisCover.shp” Debris-cover map accompanying the above glacier area (1.)  where glaciers smaller than 1 km2 are removed. The attribute field [img_time] gives the decimal year, time of satellite image acquisition, that was the source data to derive each debris cover shape. Due to the source imagery being a composite dataset, there are image seams present within some debris shapes to preserve accurate map dating.    
 
3.    “_minGl2km2.shp” An amended RGI v6.0, identical to 1. above, but where glaciers with a surface area less than 2 km2 are removed. 

The attribute table of this file include unaltered RGI v6.0 fields [3, 4]: 
[RGIId], [GLIMSId], [RGIFlag], [01Region], [02Region], [Area], [Zmin], [Zmax], [Zmed], [Slope], [Aspect], [Lmax], [GlacType] and [Name].

New fields calculated with the amended geometries for each glacier are:

[FloDivErro] = Flow divide error flags a glacier by assigning a value of 1 if the calculation [deb_km2]/[abl_km2] is greater than 1. This can occur, for example, if a glacier shape has two ablation zones defined (per the definition used here, one glacier can only have one ablation zone). Because the code anticipates only one value for ablation zone area it does not sum the area of several shapes and thus the total debris area can be greater than one ablation zone area alone. To identify remaining glaciers with two or more ablation zones (indicative of a flow divide error), the centroid of each ablation zone area was found. Where two or more centroids fell within one glacier shape (accounting for centroids that might fall within a nunatak void), [FloDivErro] was assigned a value of 2. For both [FloDivErro] = 1 and [FloDivErro] = 2, the remaining metrices were set to the null value of ‘-9999’ and the potential flow divide error was not assessed or corrected.

[CenLon] = Centroid longitude, recalculated with appended glacier geometry. 

[CenLat] = Centroid latitude, recalculated with appended glacier geometry.

[gl_km2] = Glacier area (km2) calculated with appended glacier geometry.

[gl_km2_nan] = Glacier area (km2) calculated with appended glacier geometry. Glaciers where the criteria was not met for the metrics with the most strict criteria (stage, moraine abundance, future debris expansion and AAR) are assigned a NaN (not a number) value of ‘-9999’. 

[deb_km2] = Debris-covered area (km2) calculated with appended glacier geometry.

[abl_km2] = Ablation zone area (km2) calculated with appended glacier geometry.

[abl_prim_m] = Raw ablation zone perimeter (m).

[abl_primSm] = Simplified ablation zone perimeter (m), filtered to reduce variability in glacier outline complexity due to the variable methods used to generate outlines within the RGI 6.0. 

[deb_prim_m] = Debris cover perimeter (m).

[eq_line_m] = Equilibrium line length (m).

[exPotLen_m] = Length of the debris cover perimeter that is composed of the boundary between bare glacier ice and supraglacial debris cover. 

[aar] = the AAR (accumulation zone area ratio).

[stage] = Debris cover stage metric. Glaciers that did not meet the criteria for this metric (see above) are assigned a NaN value of ‘-9999’.

[debExPot] = Debris expansion potential metric. Glaciers where no debris cover was mapped are assigned a NaN value of ‘-9999’.

[futExPot] = Future expansion potential metric. Glaciers that did not meet the criteria for this metric (see above) are assigned a NaN value of ‘-9999’.

[percDeb] = Percent debris-covered.

[percDebNaN] = Percent debris-covered. Glaciers where the criteria was not met for the metrics with the most strict criteria (stage, moraine abundance, future debris expansion and AAR) are assigned a NaN (not a number) value of ‘-9999’. 

[moraineAbu] = Moraine abundance metric. Glaciers that did not meet the criteria for this metric (see above) are assigned a NaN value of ‘-9999’.

4.    “_minGl2km2_debrisCover.shp” Debris-cover map accompanying the above glacier area (2.) where glaciers smaller than 2 km2 are removed. 

5.    “_ablationZone.shp” Estimated ablation zone area after smoothing and simplifying filters. Attribute table field [abl_perim] is the ablation zone perimeter length (m) for each glacier meeting the debris cover percentage and debris-covered area criteria described above.
 
6.    “_equilibriumLine.shp” Estimated equilibrium line for each glacier that met the criteria for defining an ablation zone. 

7.    “_debrisExpansionLine.shp” The boundary between supraglacial debris and bare glacier ice. 

8.    “_RGI60_fp.shp” Area within the Randolph Glacier Inventory version 6.0 that was manually identified as false positive relative to the Landsat imagery used in this study. 

9.    “_RGI60_fn.shp” Glacier area that was visually determined to be neglected from the Randolph Glacier Inventory version 6.0 and manually digitized from the Landsat imagery used in this study.

Data S2.
S2 consists of 15 shapefiles, one for each RGI region where: RGI regions 01 and 02 are combined; RGI regions 13, 14 and 15 are combined; and RGI region 19 (Antarctic and subantarctic) is excluded. Each shapefile contains the Landsat image footprints that have been pieced together to form an overlap-free image composite from which debris cover was mapped in this study. Attribute field [img_time] gives the decimal year of image acquisition for each satellite image and [Landsat_ID] gives the NASA assigned ID.

Data S3.
S3 contains the results from a debris cover map comparison analysis with [1] on a global scale and [2] for High Mountain Asia. For the comparison with [1], each file begins with the RGI region number and name, e.g. “01Alaska”, followed by a shorthand name as defined below:  

1.    “_Scherler2018_fp” False positive area in the results from [1] (Landsat ratio product: LS_DC_2013_2017_RATIO) over only the spatial domain where there was data coverage in this study and where "DC_CTSmean >= 10" from [1]. This term is imperfectly named because instances where [1] is correct and this study is wrong should not be classified as a ‘false positive’ difference. 

2.    “_Scherler2018_fn” False negative area in the results from [1] (Landsat ratio product: LS_DC_2013_2017_RATIO) over only the spatial domain where there was data coverage in this study and where "DC_CTSmean >= 10" from [1]. This term is imperfectly named because instances where [1] is correct and this study is wrong should not be classified as a ‘false negative’ difference. 

3.    “_Scherler2018_tp” Area where the debris cover map from this study and [1] agree. 

For the comparison with [2], each file begins with “13_14_15HMA” followed by a shorthand name as defined below:  

4.    “_Kraaijenbrink2017_fp” False positive area in the debris maps from [2] over the spatial domain where there was data coverage in this study. This term is imperfectly named because instances where [2] is correct and this study is wrong should not be classified as a ‘false positive’ difference. 

5.    “_Kraaijenbrink2017_fn” False negative area in the results from [2] over the spatial domain where there was data coverage in this study. This term is imperfectly named because instances where [2] is correct and this study is wrong should not be classified as a ‘false negative’ difference. 

6.    “_Kraaijenbrink2017_tp” Area where the debris cover map from this study and [2] agree.

References
1.    Scherler, D.,Wulf, H. & Gorelick, N. Global assessment of supraglacial debris‐cover extents. Geophys. Res. Lett. 45, 11798-11805 (2018).
2.    Kraaijenbrink, P. D. A., Bierkens, M.F.P., Lutz, A.F. & Immerzeel, W.W. Impact of a global temperature rise of 1.5 degrees Celsius on Asia’s glaciers. Nature 549, 257-260 (2017).
3.    W. T. Pfeffer, A. A. Arendt, A. Bliss, T. Bolch, J. G. Cogley, A. S. Gardner, J. O. Hagen, R. Hock, G. Kaser, C. Kienholz, E. S. Miles, G. Moholdt, N. Mölg, F. Paul, V. Radić, P. Rastner, B. H. Raup, J. Rich, M. J. Sharp, The Randolph Consortium, The Randolph Glacier Inventory: a globally complete inventory of glaciers. J. glaciol. 60, 537-552 (2014).
4.    RGI Consortium Randolph Glacier Inventory-A dataset of global glacier outlines: version 6.0 (Global Land Ice Measurements from Space (GLIMS), 2017).