Dharpure et al. (2026) Future projections of glacier mass change in High Mountain Asia using GRACE and climatemodel data

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-02-13
• Authors: Jaydeo K. Dharpure, Ian M. Howat, Akansha Patel
• DOI: 10.1038/s41598-026-39404-8

Research Groups

• Byrd Polar and Climate Research Center, The Ohio State University, Columbus, USA
• School of Earth Sciences, The Ohio State University, Columbus, USA
• Texas A&M AgriLife, Blackland Research and Extension Center, Temple, TX, USA

Short Summary

This study quantifies past (2002-2023) and projects future (2024-2100) glacier mass change in High Mountain Asia using GRACE/GRACE-FO data, machine learning for gap-filling, and a Generalized Additive Model driven by ISIMIP3b climate scenarios, revealing significant continued mass loss, especially under high-emission pathways.

Objective

• To examine glacier mass change in High Mountain Asia (HMA) from 2002/03 to 2022/23 using GRACE/GRACE-FO data.
• To project future glacier mass change in HMA from 2023/24 to 2099/2100 under low- (SSP126) and high-emission (SSP585) scenarios.
• To quantify the independent and combined influence of climatic and radiative flux variables on glacier mass change.

Study Configuration

• Spatial Scale: High Mountain Asia (HMA) and its 15 sub-regions, covering a total glacier area of 99,468 square kilometers. Data resolutions vary from 0.05° to 0.5° for satellite products and 9 km for reanalysis.
• Temporal Scale:
  • Past/Reference Period: 2002/03 to 2022/23 (21 hydrological years).
  • Future Projections: 2023/24 to 2099/2100, analyzed in early (2024-2049), mid (2050-2075), and late (2076-2100) century periods. Monthly timescale for most data.

Methodology and Data

• Models used:
  • MissForest algorithm: A non-parametric machine learning algorithm used for gap-filling the 33-month data gap in GRACE and GRACE Follow-On terrestrial water storage anomaly (TWSA) observations.
  • Generalized Additive Model (GAM): A semi-parametric model employed to capture nonlinear relationships between glacier mass change (GMC) and its climatic-radiative drivers, used for both historical analysis and future projections.
• Data sources:
  • Satellite Gravimetry: GRACE (April 2002 - May 2017) and GRACE Follow-On (GRACE-FO) (July 2018 onward) terrestrial water storage anomaly (TWSA) mascon solutions from Center for Space Research (CSR) RL06.2 (0.25°), Jet Propulsion Laboratory (JPL) RL06.3Mv04 (0.5°), and German Research Center for Geosciences (GFZ) RL06v2.0 (0.5°). An ensemble mean of these solutions was used.
  • Land Surface Model: GLDAS Noah Land Surface Model L4 (for soil moisture, snow water equivalent, and canopy water).
  • Reanalysis Data: ERA5-Land (for air temperature, incoming shortwave radiation, incoming longwave radiation, and specific humidity) at approximately 9 km spatial resolution, monthly timescale.
  • Precipitation: Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG) Version 07 Level 3 “Final Run” daily product (0.1° x 0.1° spatial resolution).
  • Land Surface Temperature: Moderate Resolution Imaging Spectroradiometer (MODIS) Terra (MOD11C3) and Aqua (MYD11C3) mean monthly products (0.05° resolution).
  • Glacier Inventory: Randolph Glacier Inventory (RGI) Version 7.0 (RGI7) for glacier boundaries.
  • Future Climate Scenarios: Inter-Sectoral Impact Model Intercomparison Project Phase 3b (ISIMIP3b) dataset, providing bias-corrected outputs from five CMIP6 General Circulation Models (GCMs): MPI-ESM1-2-HR, UKESM1-0-LL, GFDL-ESM4, IPSL-CM6A-LR, and MRI-ESM2-0. Data at daily and 0.5° grid resolution for Shared Socioeconomic Pathway 126 (SSP126, low-emission) and SSP585 (high-emission) scenarios.

Main Results

• Past Glacier Mass Change (2002/03-2022/23):
  • High Mountain Asia experienced an overall glacier mass loss of -13.9 ± 3.6 gigatonnes per year (Gt/yr), equivalent to an average loss of -9.0 ± 2.4 millimeters per year (mm/yr) over the region.
  • Subregional variability was pronounced: Eastern Kunlun showed a mass gain of 1.1 ± 0.2 Gt/yr, while the West Tien Shan and Western Himalaya experienced the most rapid mass losses, exceeding -1.8 Gt/yr.
• Hydroclimatic Trends (2002/03-2022/23):
  • HMA-wide, precipitation changes were negligible (+1.06 ± 1.78 mm/yr).
  • Consistent warming was observed: air temperature increased by +0.04 ± 0.03 °C/yr and land surface temperature by +0.04 ± 0.04 °C/yr.
  • Incoming longwave radiation increased by +0.17 ± 0.11 W/m²/yr, while incoming shortwave radiation declined by -0.15 ± 0.10 W/m²/yr.
• Future Glacier Mass Change (2023/24-2099/2100):
  • Under the low-emission SSP126 scenario, the rate of mass loss slows considerably, averaging -2.3 ± 0.3 Gt/yr, with a slightly positive mass balance (0.8 ± 1.2 Gt/yr) projected by the late 21st century.
  • Under the high-emission SSP585 scenario, accelerated glacier mass loss is projected, averaging -19.5 ± 11.3 Gt/yr. Rates are expected to increase from -14.0 ± 6.1 Gt/yr in the early century to -24.5 ± 15.3 Gt/yr by the late century. By 2049/50, SSP585 projects nearly three times higher loss than SSP126.

Contributions

• Enhanced temporal continuity and robustness of glacier mass change monitoring by employing the MissForest machine learning algorithm to bridge data gaps between GRACE and GRACE Follow-On missions.
• Integrated a comprehensive set of climatic (temperature, precipitation, specific humidity) and radiative flux variables (incoming shortwave and longwave radiation) to provide a more physically grounded assessment of glacier-climate interactions.
• Utilized a non-parametric Generalized Additive Model (GAM) to effectively capture and interpret nonlinear relationships between glacier mass change and its drivers.
• Provided scenario-based future projections of glacier mass change in HMA using bias-corrected ISIMIP3b climate data under SSP126 and SSP585, extending GRACE-based retrospective analyses to forward-looking risk evaluation for water resource management and climate adaptation.

Funding

• Byrd Postdoctoral Fellowship awarded to JKD from the Byrd Polar and Climate Research Center, The Ohio State University.

Citation

@article{Dharpure2026Future,
  author = {Dharpure, Jaydeo K. and Howat, Ian M. and Patel, Akansha},
  title = {Future projections of glacier mass change in High Mountain Asia using GRACE and climatemodel data},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-39404-8},
  url = {https://doi.org/10.1038/s41598-026-39404-8}
}