Blogpost about OGGM v1.6 gdirs and projections

[lilianschuster]

Here is a first draft for a new blogpost.

There are a few TODO mentions that may need to be rather explained by / discussed with @fmaussion (either here in the blogpost or in the documentation).

To discuss:

• a similar description of the gdirs will go into the OGGM documentation, should we keep the description, but I think it is still good to keep the gdir description here?
• where should the special RGI70 regional calibration procedure be explained (maybe in the OGGM documentation preprocessed gdir part)?
• I am still confused why the initial volume can be so much higher for 16,17,18 under ERA5 ... although the failing glacier area is not that different. Is it really just the different past climate from W5E5/ERA5 between RGI date and year 2000 that explains it?
• Regional dmdtda from Hugonnet et al. (2021) (this is estimated by theiy “time-varying“ RGI area (tarea from "linear interpolation") is different to the regional "constant RGI area-weighted" per-glacier dmdtda:
  • There is a clear model difference in the fixed-geometry dmdtda (https://cluster.klima.uni-bremen.de/~oggm/oggm-standard-projections/analysis_notebooks/1.6.3_2025_6_gdirs_proj_comparisons/figures/1_hist_reg_mass_balance_comparisons_dmdtda_fixed_geometry.png).
  • However this pattern vanishes when looking at the regional mass change (https://cluster.klima.uni-bremen.de/~oggm/oggm-standard-projections/analysis_notebooks/1.6.3_2025_6_gdirs_proj_comparisons/figures/1_hist_reg_mass_change_oggm_2025.6_comparisons.png)
  • Probably not so problematic ?! Is this clear enough from the text?

[lilianschuster]

Related to the per-glacier vs regional discussion:
--> Maybe the manual filtering of OGGM (https://docs.oggm.org/en/stable/reference-mass-balance-data.html#geodetic-mb-data) which removes outliers with very larger errors and non-existent glaciers and replaces their mass-balance with the regional average also explains some of the differences ?

[lilianschuster]

todo: mention the missing "spinup" in case of glaciers in New Zealand where the RGI date is often near to the year 1979 (i.e. the beginning of the spinup currently). In future, we may need to start the spinup earlier than 1979.

todo: link to the documentation where RGI 7.0 initialisation/calibration is described (once the documentation pull request is merged, it will be here: https://docs.oggm.org/en/stable/shop.html#pre-processed-directories). This documentation page also further describes the gdir differences (--> link to that in the blogpost description)

[fmaussion]

sorry for checking this only now. Reads very good. I think I addressed most of the comments - I don't have a place in the docs for detailed explanation on the gdirs yet, but I don't think its necessary to understand the post!

Very well done @lilianschuster - I'm tempted to merge soon, perhaps saying that 1.6.3 will be released shortly