Shahbazi et al. (2025) The compounding effects of agricultural expansion and snow drought on lake urmia’s drying crisis

Identification

• Journal: Scientific Reports
• Year: 2025
• Date: 2025-10-31
• Authors: Afsaneh Shahbazi, Yusuf Alizade Govarchin Ghale, Gülüzar Duygu Semiz, Elifnaz Torun, Babak Vaheddoost, Neda Beirami, Alper Ünal, Kaveh Madani, Amir AghaKouchak
• DOI: 10.1038/s41598-025-21735-7

Research Groups

• Graduate School of Natural and Applied Sciences, Ankara University, Ankara, Turkey
• Department of Climate and Marine Sciences, Eurasia Institute of Earth Sciences, Istanbul Technical University, Istanbul, Turkey
• Department of Farm Structures and Irrigation, Ankara University, Agricultural Engineering Faculty, Ankara, Turkey
• Department of Civil Engineering, Bursa Technical University, Bursa, Turkey
• Department of Civil and Environmental Engineering, Amirkabir University of Technology, Tehran, Turkey
• Department of Civil and Environmental Engineering, University of Washington, Seattle, WA, USA
• United Nations University Institute for Water, Environment, and Health (UNU-INWEH), Richmond Hill, Ontario, Canada
• The City College of New York, The City University of New York Remote Sensing Earth Systems (CUNY-CREST) Institute, New York, USA
• Department of Civil and Environmental Engineering, University of California, Irvine, CA, USA
• Department of Earth System Science, University of California, Irvine, CA, USA

Short Summary

This study investigated the combined impacts of agricultural expansion and climate variability, including snow drought, on river inflows to Lake Urmia from 1985 to 2020. Findings reveal that agricultural water use was the dominant factor, accounting for approximately 66% of the total impact on reduced river inflows, amplified by a persistent snow drought.

Objective

• To examine the compounding effects of anthropogenic (agricultural expansion) and climatic factors (precipitation, evaporation, and snow dynamics) on river inflows and water level fluctuations of Lake Urmia.
• To develop an efficient multi-model approach, incorporating statistical and deep learning methods, for accurately capturing the complex hydroclimatic dynamics of the Lake Urmia Basin and estimating river inflows.

Study Configuration

• Spatial Scale: Lake Urmia Basin (LUB), Iran, covering approximately 52,000 km². Lake Urmia has an average surface area of 5,000 km². Elevations within the basin range from 1,222 meters to 3,875 meters above sea level.
• Temporal Scale: 1985 to 2020 for the main analysis of river inflows and drivers. Snow Water Equivalent (SWE) data from ERA5 extended to 2023. In-situ snow measurements were available from 2007 to 2013.

Methodology and Data

• Models used:
  • Statistical models: Multiple Linear Regression (MLR), Response Surface Model (RSM), Mann-Kendall trend analysis, Welch’s T-Test, Bivariate simple linear regression.
  • Deep Learning: Convolutional Neural Networks (CNNs).
  • Interpolation methods: Co-Kriging, Inverse Distance Weighted (IDW).
• Data sources:
  • In-situ data: Hydro-climatic and water storage data (precipitation, rain, temperature, evaporation, river inflows, Agricultural Water Use (AWU), and lake water level fluctuations) from Iranian government organizations (1985-2020). Monthly snowfall data (snow density, snow depth, SWE) from 81 ground-observed stations (2007-2013) from the Iran Meteorological Organization.
  • Satellite data:
    • AMSR-E (Aqua satellite): Daily Snow Water Equivalent (SWE) (2002-2011), 25 km spatial resolution.
    • MOD10A1 (Terra/Aqua MODIS): Daily Snow Cover (SC) (2000-present), 500 m spatial resolution.
    • PERSIANN-CDR: Daily precipitation (1983-present), 0.25° spatial resolution.
    • IMERG-GPM: Half-hourly precipitation (1998-present), 0.1° spatial resolution.
    • ASTER-GDEM: Digital Elevation Model (DEM) (30 m spatial resolution).
  • Reanalysis and Land Data Assimilation System data:
    • ERA5-Land (ECMWF): Hourly/monthly SWE, Precipitation, Temperature (1950-present), approximately 9 km spatial resolution.
    • GLDAS (Noah) (NASA): 3-hourly SWE (1985-2022), 0.25° spatial resolution.

Main Results

• A persistent snow drought began in the late 1990s, coinciding with a more than fourfold increase in irrigated lands within the basin.
• Mean monthly average SWE decreased significantly from 42.3 mm (1985–1998) to 23.9 mm (1999–2023), representing a 43.49% decline in annual SWE. The peak snow season shifted from March to February after 1999.
• The mean temperature in the basin increased by approximately 1 °C after 1999, and the rain-to-SWE ratio showed a noticeable increase, indicating a shift towards more rain and less snow.
• Spatially, the largest decreases in SWE (20–30 mm) were observed in the mountainous western region of the basin.
• Mann-Kendall trend analysis revealed statistically significant decreasing trends in river inflows, SWE, and lake water level, and a significant increasing trend in Agricultural Water Use (AWU) from 1985 to 2020. No significant trends were found for overall precipitation or evaporation.
• Correlation analysis showed a strong positive correlation between river inflows and snow dynamics (0.81), and a negative correlation with AWU (-0.60). Lake water level fluctuations were strongly dependent on AWU (-0.95), river inflows (0.63), and SWE (0.42).
• The Convolutional Neural Network (CNN) model consistently outperformed Multiple Linear Regression (MLR) and Response Surface Model (RSM) in predicting river inflows. In the 80% training scenario, the CNN model achieved R² and KGE values of approximately 0.9 in both calibration and verification phases, with significantly lower RMSE values (745 x 10^6 m³ for calibration and 347 x 10^6 m³ for verification).
• Scenario-based analysis, restoring key parameters to pre-1999 levels, quantified the relative contributions to changes in river inflows:
  • Reverting agricultural water use was the dominant factor, accounting for approximately 66% (95% CI: 56%–76%) of the total impact.
  • Restoring precipitation contributed 25% (95% CI: 18%–33%).
  • Restoring evaporation contributed 9% (95% CI: 7%–12%).
  • Restoring both precipitation and evaporation simultaneously explained 34% (95% CI: 26%–43%) of the change.
• These results underscore the primary role of agricultural water demand, amplified by declining snowpack and climatic shifts, in altering the basin's hydrology.

Contributions

• This study is the first to comprehensively investigate the compounding effects of agricultural expansion and snow drought on Lake Urmia's drying crisis, explicitly integrating snow dynamics alongside other climatic and anthropogenic factors.
• It developed and applied a novel hybrid framework combining statistical models (MLR, RSM) and Convolutional Neural Networks (CNNs) to accurately estimate river discharge and disentangle the complex effects of hydroclimatic and anthropogenic drivers, demonstrating the superior performance of CNNs.
• The research quantitatively attributed the relative contributions of agricultural water use, precipitation, and evaporation to river inflow changes through a robust scenario-based analysis, definitively highlighting the dominant role of anthropogenic factors.
• It provided a detailed spatio-temporal analysis of Snow Water Equivalent (SWE) and temperature changes, revealing a significant decline in SWE, a shift in the peak snow season, and an increased rain-to-SWE ratio, offering critical insights into the basin's changing hydrological regime.
• The findings emphasize the urgent need for integrated water resource management, focusing on climate adaptation, enhanced snowpack monitoring, and sustainable agricultural practices to ensure the long-term water security and ecological restoration of Lake Urmia.

Funding

The authors gratefully acknowledge the National Aeronautics and Space Administration (NASA), the European Centre for Medium-Range Weather Forecasts (ECMWF), the National Oceanic and Atmospheric Administration (NOAA) (for PERSIANN CDR system), and the Iran Meteorological Organization for providing access to the satellite, reanalysis, and in-situ data used in this study. No specific funding projects, programs, or reference codes were explicitly listed in the paper text.

Citation

@article{Shahbazi2025compounding,
  author = {Shahbazi, Afsaneh and Ghale, Yusuf Alizade Govarchin and Semiz, Gülüzar Duygu and Torun, Elifnaz and Vaheddoost, Babak and Beirami, Neda and Ünal, Alper and Madani, Kaveh and AghaKouchak, Amir},
  title = {The compounding effects of agricultural expansion and snow drought on lake urmia’s drying crisis},
  journal = {Scientific Reports},
  year = {2025},
  doi = {10.1038/s41598-025-21735-7},
  url = {https://doi.org/10.1038/s41598-025-21735-7}
}