Leopardi et al. (2026) A satellite-based approach for estimating runoff and river discharge in the Pan-Arctic region from 2003 to 2022

Identification

• Journal: Remote Sensing of Environment
• Year: 2026
• Date: 2026-03-10
• Authors: Francesco Leopardi, Jacopo Dari, Carla Saltalippi, Luca Brocca, Espen Volden, Diego Fernández-Prieto, Peyman Saemian, Nico Sneeuw, Mohammad J. Tourian, Stefania Camici
• DOI: 10.1016/j.rse.2026.115353

Research Groups

• Department of Civil and Environmental Engineering, University of Perugia, Perugia, Italy
• National Research Council, Research Institute for Geo-Hydrological Protection, Perugia, Italy
• European Space Agency, ESA-ESRIN, Frascati, Italy
• Institute of Geodesy, University of Stuttgart, Stuttgart, Germany

Short Summary

This study presents daily, long-term (2003–2022) satellite-based estimates of river discharge and gridded runoff at 0.25° × 0.25° spatial resolution across the continental Pan-Arctic region. Integrating various satellite observations into the adapted STREAM hydrological model, it demonstrates high performance (median Kling-Gupta Efficiency of 0.83) and quantifies freshwater fluxes to the Arctic Ocean at 4760 ± 619 km³ yr⁻¹.

Objective

• To estimate daily, long-term (2003–2022) satellite-based river discharge and gridded runoff at 0.25° × 0.25° spatial resolution across the continental Pan-Arctic region.
• To develop and apply a satellite-based approach (STREAM v1.0 model) integrating precipitation, soil moisture, snow cover fraction, and Terrestrial Water Storage Anomalies (TWSA) for this estimation.
• To regionalize STREAM model parameters to extend estimates to ungauged areas across the entire Pan-Arctic region.

Study Configuration

• Spatial Scale: Continental Pan-Arctic region (16.5 × 10⁶ km²), with a spatial resolution of 0.25° × 0.25°.
• Temporal Scale: Daily estimates over a long-term period from 2003 to 2022.

Methodology and Data

• Models used: STREAM (SaTellite based Runoff Evaluation And Mapping) model, specifically STREAM v1.0, a semi-distributed conceptual hydrological model adapted for Arctic conditions.
• Data sources:
  • Satellite:
    • Precipitation: Integrated Multi-satellite Retrievals (IMERG) for Global Precipitation Measurement (GPM) Final run product (IMERG-F v6).
    • Soil Moisture: ESA Climate Change Initiative (ESA CCI) combined soil moisture dataset (ESA CCI SM v8.1).
    • Terrestrial Water Storage Anomalies (TWSA): Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) Goddard Space Flight Center (GSFC) mascon solution.
    • Snow Cover Fraction (SCF): ESA Snow CCI project (v4.0) from Terra MODIS observations.
  • Ground-based (for calibration/validation):
    • Air Temperature: Climate Prediction Center (CPC) Global Unified Temperature dataset (NOAA PSL).
    • River Discharge: Global Runoff DataBase (GRDC), USGS National Water Information System, Arctic Great Rivers Observatory (ArcticGRO).
  • Ancillary: Digital Elevation Model (DEM) from HydroSHEDS, river basin boundaries and network from Major River Basins of the World (MRB) database, Integrated Pan-Arctic Catchments Summary Database (ARCADE).

Main Results

• The STREAM v1.0 model accurately reproduces daily river discharge observations in 15 donor Arctic basins, achieving a median Kling-Gupta Efficiency (KGE) of 0.83 with calibrated parameters.
• STREAM v1.0 significantly improves performance over the previous version (STREAM v0.0) for the 6 largest basins, showing a 71% median increase in KGE, a 71% median Pearson correlation (ρ) improvement, and a 49% median reduction in relative Root Mean Square Error (rRMSE).
• A regionalization approach, based on the aridity index and optimal 3 clusters, enabled parameter transfer to ungauged regions, maintaining acceptable accuracy with a median KGE of 0.51 across 26 independent gauging stations.
• The total average annual freshwater flux from the continental Pan-Arctic region to the Arctic Ocean is estimated at 4760 ± 619 km³ yr⁻¹ for the period 2003–2022.
• Over 90% of the total annual freshwater inflow to the Arctic Ocean occurs during the summer months (April–September), estimated at approximately 4406 ± 572 km³ yr⁻¹.
• The majority of freshwater runoff originates from the Eurasian region, contributing approximately 3880 ± 504 km³ yr⁻¹ (about 80% of the total continental runoff).
• Trend analysis for 2003–2022 indicates that 18% of the total Pan-Arctic area exhibits statistically significant runoff trends (p < 0.05). Among these, 43% show a positive trend (average annual increase of 6%), while 57% show a negative trend (average annual decrease of 4%).

Contributions

• Presents the first dataset providing daily river discharge and runoff estimates for the entire Pan-Arctic region based exclusively on multi-sensor satellite observations.
• Introduces STREAM v1.0, a novel semi-distributed conceptual hydrological model specifically adapted for Arctic conditions, demonstrating enhanced accuracy in simulating river discharge, particularly for near-zero winter flows and summer peak flows.
• Highlights the invaluable role of satellite observations, especially GRACE/GRACE-FO Terrestrial Water Storage Anomalies, in supporting large-scale hydrological modeling and reconstructing the Arctic freshwater cycle in data-scarce environments.
• Develops and validates a regionalization approach for STREAM parameters based on basin climatic attributes (aridity index), enabling reliable runoff and river discharge estimation in ungauged areas.
• Provides a spatially and temporally consistent runoff dataset that complements more complex process-based models, facilitating the quantification of freshwater fluxes to the Arctic Ocean and the assessment of associated runoff trends for climate impact assessments.

Funding

• European Space Agency (ESA) through the STREAM-NEXT Project (EO Science for Society Permanently Open Call 4000126745/19/I-NB-CCN2).

Citation

@article{Leopardi2026satellitebased,
  author = {Leopardi, Francesco and Dari, Jacopo and Saltalippi, Carla and Brocca, Luca and Volden, Espen and Fernández-Prieto, Diego and Saemian, Peyman and Sneeuw, Nico and Tourian, Mohammad J. and Camici, Stefania},
  title = {A satellite-based approach for estimating runoff and river discharge in the Pan-Arctic region from 2003 to 2022},
  journal = {Remote Sensing of Environment},
  year = {2026},
  doi = {10.1016/j.rse.2026.115353},
  url = {https://doi.org/10.1016/j.rse.2026.115353}
}