Cherie et al. (2025) Agricultural drought dynamics in East Gojjam: Insights on soil moisture, drought indices, and crop sustainability

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2025
• Date: 2025-11-20
• Authors: Walelgn Dilnesa Cherie, Bekalu Weretaw Asres
• DOI: 10.1016/j.ejrh.2025.102928

Research Groups

• Debre Markos University, Technology Institute, Hydraulic and Water Resources Engineering Department, Debre Markos, Ethiopia

Short Summary

This study investigates agricultural drought dynamics in East Gojjam, Ethiopia, from 2013 to 2024 by integrating remote sensing indices (NDVIMax, SWDI, SMA, SPEI) and rainfall data with machine learning models. It found significant interannual drought variability, identifying 2022 as a critical year, and highlighted Soil Water Deficit Index (SWDI) and Normalized Difference Vegetation Index (NDVIMax) as the most influential predictors of vegetation response.

Objective

• To quantify drought impacts by integrating historical and current satellite data (NDVI, SMA, SWDI, SPEI) with ground-based crop yield records.
• To analyze spatiotemporal drought patterns and their correlation with crop productivity, evaluating the predictive strength of each drought index.
• To identify the most reliable indicators for early warning and decision-making, offering woreda-specific insights into drought vulnerability and crop resilience for policy formulation.

Study Configuration

• Spatial Scale: Eight selected woredas within the East Gojjam Zone, Amhara Region, Ethiopia, in proximity to the Upper Blue Nile River Basin.
• Temporal Scale: 2013–2024 for satellite-based soil moisture and vegetation indices; 1996–2024 for precipitation and evaporation/drought index trends.

Methodology and Data

• Models used: Random Forest, Multiple Linear Regression (MLR), Penman-Monteith method (for ET0). Drought indices: Soil Moisture Anomaly (SMA), Soil Water Deficit Index (SWDI), Standardized Precipitation and Evapotranspiration Index (SPEI), Normalized Difference Vegetation Index (NDVI).
• Data sources:
  • Satellite-based soil moisture (NASA’s Climate Data Portal)
  • In-situ soil moisture (Field sampling across 8 woredas)
  • Precipitation (Ethiopian Meteorological Institute, CHIRPS)
  • Landsat 8 satellite imagery (for NDVI)
  • Crop productivity data (East Gojjam zone agricultural development office, in quintals per hectare)
  • Evaporation data (Ethiopian Metrological Institute)

Main Results

• Significant interannual variability in drought intensity was observed, with 2022 identified as a critical drought year marked by widespread hydrological stress and reduced crop vigor.
• SWDI and NDVI_Max emerged as the most influential predictors of vegetation response, while rainfall showed a nonlinear impact.
• Long-term rainfall trend analysis (30 years) revealed declining trends in Dejen, Enemay, and Enarg E woredas, suggesting increased drought susceptibility.
• SMA time-series showed pronounced negative anomalies around 2016 and 2022, indicating periods of moisture deficit.
• Aneded and Baso Liben consistently exhibited positive SWDI values (favorable soil moisture), contrasting with Dejen and Shebel Berenta which showed extreme deficits in certain years (e.g., -0.7 in Dejen, -0.8 in Shebel Berenta).
• Regression analysis indicated a positive linear correlation between SWDI and crop productivity (R² = 0.39), and weaker positive correlations for soil moisture (R² = 0.15) and NDVI (R² = 0.22) with crop productivity.
• Predictive models (MLR and Random Forest) for crop productivity yielded negative R² values (MLR: -0.034; RFR: -0.047), indicating limited explanatory power, but Mean Absolute Error (MAE) values were relatively low (approximately 4.77 quintals per hectare).
• Feature importance analysis from the Random Forest model identified Soil Moisture (0.37) and SWDI (0.27) as the most influential predictors of agricultural drought sensitivity.
• Bias correction analyses revealed slight overestimations in remote sensing data compared to in-situ measurements for soil moisture, field capacity, and wilting point, with Nash-Sutcliffe Efficiency (NSE) values of 0.89, 1.00, and 0.99, respectively.

Contributions

• Developed a multidimensional, data-integrated framework for drought monitoring tailored to East Gojjam’s woreda-level agroecological complexity, moving beyond routine national-scale assessments.
• Pioneered the fusion of satellite-derived drought indicators (NDVI, SPEI, SWDI, SMA) with ground-truth crop yield data (quintals per hectare) for precise quantification of drought impacts.
• Dissected the temporal dynamics of drought onset, duration, severity, and magnitude through the synergistic use of SPEI and SWDI, providing a time-sensitive lens.
• Proposed a resilience mapping approach across East Gojjam’s agroecological zones, integrating soil moisture trends to uncover spatial disparities in crop adaptability.
• Provided a replicable, data-driven framework for drought risk assessment and adaptive agricultural planning in moisture-stressed regions.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Citation

@article{Cherie2025Agricultural,
  author = {Cherie, Walelgn Dilnesa and Asres, Bekalu Weretaw},
  title = {Agricultural drought dynamics in East Gojjam: Insights on soil moisture, drought indices, and crop sustainability},
  journal = {Journal of Hydrology Regional Studies},
  year = {2025},
  doi = {10.1016/j.ejrh.2025.102928},
  url = {https://doi.org/10.1016/j.ejrh.2025.102928}
}

Generated by BiblioAssistant using gemini-2.5-flash (Google API)