For experiment participation you must contribute a unique dataset of glacier mass change estimates for at least one of the 19 GTN-G glacier regions covering the observation period of GlaMBIE-2 based on a model run, which is driven by climate data and is aggregating individual glacier estimates. The runs must be based on an established glacier or mass-balance model with peer-reviewed method publication(s) in an academic journal and be submitted by the developer group.

In line with GlaMBIE guidelines, contributors of model runs are invited to join the author team of related publication(s). Note that all authors must consent to authorship and to the manuscript; we cannot organise or delay the publication process to accomplish institutional reviews or licensing outside the official journal process.

By submitting your data to GlaMBIE you agree that your data will be visually displayed to other participants and the wider community (e.g. in plots during the workshop or conferences). It is the intention of the GlaMBIE community that data submitted to the assessment may be made available with publication of the joint paper.

Data submission for GlaMBIE-2 model submissions opens in June 2026, and runs until the 1st of August 2026.

Purpose of experiment

Within GlaMBIE-2, we aim to use the regional model runs for (i) comparison to observation results over the GlaMBIE-2 observation period (1975-2025) outside the calibration period (2000-2019) and (ii) for the temporal correction of regional observations from hydrological to calendar years for global aggregation.

Standardisation of model runs

Glacier Model: Any model able to provide all or most of the output variables below is invited to participate, when provided with a peer-reviewed method publication(s) in an academic journal and be submitted by the developer group. For example, statistical models are welcome if they comply with the protocol outlined here.

Calibration data: The model can be calibrated with glaciological mass balances (WGMS 2026; https://doi.org/10.5904/wgms-fog-2026-02-10) and with geodetic elevation changes from Hugonnet et al. (2021; https://doi.org/10.1038/s41586-021-03436-z) based on RGI 6.0.

Reanalysis data: Model runs must be driven using ECMWF ERA5 reanalysis fields, which are available from the OGGM Shop for the period 1940-present, if required (https://docs.oggm.org/en/stable/shop-preprodirs.html#available-pre-processed-configurations).

Spatial domain: Model runs must cover one, multiple, or all of the first-order glacier regions 1 to 19 as used in RGI 6.0. The model results are to be submitted at regional resolution, aggregated based on individual glacier runs.

Temporal domain: Model runs must cover the period from January 1975 to March 2026 at monthly resolution (i.e., net balance from start to end date for each month). The aim is to cover the calendar years from 1975 to 2025 and to allow for temporal corrections of hydrological to calendar years for observations from both northern and southern hemispheres.

Unit domain: Model runs should provide total mass change in m water equivalent, ideally providing mass change above and below water level (which are contributing and not contributing to sea-level rise, respectively), as well as frontal ablation in additional fields.

Other model settings: Model runs should provide results for “fixed geometries”; optionally additional “dynamic runs” considering glacier area changes can be submitted. Model runs should report uncertainties at 68% or 95% confidence intervals (with choice to be indicated during submission).

To ensure consistency, a standard file format has been defined for all data submissions. Participants must provide their data in this format, including the required columns. Templates for the CSV data files are provided below. Submissions in other formats will not be accepted. Each dataset will be validated upon submission, and any issues will be communicated to the authors for correction.

• Expected Format and Attributes

The data file format is comma-separated values (CSV), following ASCII standards.

If you are submitting multiple regions, all regions should be submitted in a single file.

Note: Throughout the template and variable names, the term “observed” is retained for consistency with the broader GlaMBIE data structure and existing observational datasets. For model submissions, variables containing “observed” should be interpreted as referring to the corresponding modelled quantities.

The following attributes are expected to be present in each data table:

• region_id: GTN-G region ID (Table 1).
• start_date: start date of observation period in format DD/MM/YYYY.
• end_date: end date of observation period in format DD/MM/YYYY, each end date should ideally be the start date of the next observation (i.e. no temporal gaps in data). This means that the observed change is calculated from start_date to end date, excluding the day of end_date (for clarifications check the CSV example below).
• glacier_change_observed: Total absolute change between start_date and end_date. NaN values will not be accepted. For model runs, this field includes total mass change.
• glacier_change_uncertainty: (optional) uncertainty on the observed glacier change corresponding to 68% or 95% confidence interval. Will follow the same unit as “glacier_change_observed”.  In case of missing uncertainties, leave blank.
• glacier_change_observed_above_water_level: (optional) for model runs, this field includes the mass change relevant for sea-level rise and for comparison to observational results. If not calculated, leave blank.
• glacier_change_observed_above_water_level_uncertainty: (optional) in case of missing uncertainties, leave blank.
• glacier_change_observed_below_water_level: (optional) for model runs, this field includes the mass change NOT relevant for sea-level rise. Note that the sum of mass change above and below water level should correspond to the total mass change. If not calculated, leave blank.
• glacier_change_observed_below_water_level_uncertainty: (optional) in case of missing uncertainties, leave blank.
• glacier_frontal_ablation: (optional) for model runs, this field includes the frontal ablation as a positive value. If not calculated, leave blank.
• glacier_frontal_ablation_uncertainty: (optional) in case of missing uncertainties, leave blank.
• unit: the units of the “glacier_change_*” and “glacier_frontal_ablation” variables (as described above) and related uncertainties. For model runs, values should be “mwe”.
• glacier_area_reference_start: regional glacier area (km2) assumed at start_date. If a constant area is assumed this should be equal to the regional area from RGI. This information will help to standardise the spatial domain in case area changes were applied.
• glacier_area_reference_end: regional glacier area (km2) assumed at end_date. If a constant area is assumed this should be equal to the regional area from RGI. This information will help to standardise the spatial domain in case area changes were applied.
• observational_coverage_percentage: regional glacier area (%km2) covered by “glacier_change_observed”. For model runs, this value will often be close to 100%, but may be lower if some glaciers are excluded, fail during model execution, or otherwise do not contribute to the reported regional estimate.
• remarks: (optional) any remarks / comments.

Attributes marked as optional can be left blank (they still need to exist as a column in the CSV file).

Each submission must include an accompanying PDF document including a description of the method summarising the implementation of the model run and referring to the corresponding method publication(s) in an academic peer-reviewed journal. Model submissions do not need to use the method description template.