Distributed snow modeling across contrasting climatic regimes: Forcing sensitivity and regime-dependent assimilation
Abstract. Mountain snowpacks regulate seasonal water availability in climatically diverse headwater basins, yet long-term distributed snow observations remain sparse in complex mountainous terrain. This study investigates snow regime characteristics, forcing sensitivity, multi-source snow mapping consistency, and data assimilation effectiveness across contrasting climatic conditions over the 2003–2024 period in two hydrologically important basins in Türkiye: the Mediterranean-continental transition Seyhan Basin and the continental Upper Euphrates (Karasu) Basin. The spatially distributed S3M model was driven by three alternative meteorological forcing configurations (station observations, ERA5-Land, and MSWX-Past) and evaluated against ground measurements together with independent reanalysis and satellite products, including CERRA-Land, MODIS, and C-SNOW. Results demonstrate that forcing uncertainty primarily affects snow magnitude (snow water equivalent and snow depth), whereas snow-covered area (SCA) remains comparatively robust across forcing configurations (NSE 0.82-0.91). MSWX-Past consistently outperformed ERA5-Land, particularly in representing precipitation, substantially reducing systematic snow overestimation and improving agreement with independent snow observations. Spatial and long-term analyses further reveal an increasing prevalence of negative snow anomalies after 2013, accompanied by episodic decoupling between snow magnitude and persistence. Data assimilation exhibited pronounced regime-dependent behavior. In the transient Seyhan Basin, assimilation substantially improved local snow dynamics by correcting timing errors and premature melt cycles, increasing station-scale performance from KGE 0.28 to 0.61. In contrast, assimilation in the continental Karasu Basin primarily reduced large-scale accumulation biases through spatially coherent adjustments, while point-scale improvements remained spatially variable. Overall, the results demonstrate that forcing uncertainty primarily affects snow magnitude, whereas the effectiveness of data assimilation is strongly governed by snow-regime characteristics. These findings provide a transferable framework for distributed snow monitoring and demonstrate the operational potential of S3M for hydrological forecasting in climatically heterogeneous, data-scarce mountain basins.
Authors
- Aynur Şensoy (ORCID: https://orcid.org/0000-0003-3004-4912)
- Simone Gabellani (ORCID: https://orcid.org/0000-0001-5117-7525)
- Francesco Avanzi (ORCID: https://orcid.org/0000-0003-4235-2373)
- Ali Arda Şorman (ORCID: https://orcid.org/0000-0003-3143-7793)
- Mevlüthan Sakallı (ORCID: https://orcid.org/0009-0001-5583-5422)
Institutions
- Eskisehir Technical University (TR)
- CIMA Research Foundation (IT)
Publication Details
- Published
- 2026-11-16
- DOI
- https://doi.org/10.5194/egusphere-2026-4921
- Primary Topic
- Cryospheric studies and observations
- Type
- preprint
Funders
- Eskişehir Teknik Üniversitesi
- European Organisation for the Exploitation of Meteorological Satellites