Logo image
Back
Decoding drought-yield relationships in the U.S. Midwest: A multiscale analysis using climatic indicators and random forests
Journal article   Peer reviewed

Decoding drought-yield relationships in the U.S. Midwest: A multiscale analysis using climatic indicators and random forests

SM Samiul Islam, Jerry Mount and Ibrahim Demir
European journal of agronomy, Vol.177, 128088
06/2026
DOI: 10.1016/j.eja.2026.128088

View Online

Abstract

Agricultural drought develops when inadequate soil moisture, frequently caused by extended precipitation shortages and heightened evaporative demand, results in substantial declines in crop output. In Iowa, where maize and soybeans comprise almost 90% of total crop output, comprehending drought-crop dynamics is crucial for climate-resilient agriculture. This study evaluated the efficacy of many commonly utilized drought indicators in elucidating the variability of corn and soybean yields from 2000 to 2022. Both meteorological and satellite-derived indices were assessed, including the Standardized Precipitation Index (SPI), Standardized Precipitation-Evapotranspiration Index (SPEI), Palmer Drought Severity Index (PDSI), Evaporative Demand Drought Index (EDDI), Crop Moisture Index (CMI), and Normalized Difference Vegetation Index (NDVI), all utilized across various temporal scales. Spearman correlation study indicated a favorable association between soybean yields and long-term indicators, including SPI-6, SPI-12, SPEI-6, and SPEI-12, with PDSI exhibiting the most robust temporal link, underscoring soybeans' need on consistent moisture availability. Corn yields had more pronounced negative relationships with EDDI, highlighting their susceptibility to atmospheric dryness and thermal stress. A Random Forest regression model was utilized to assess the relative significance of each drought measure, therefore complementing these findings. The machine learning findings indicated that CMI is the most significant predictor for both crops, although EDDI and SPI-3 were predominant factors for corn yield, underscoring the necessity to observe both short-term moisture and atmospheric demand. These findings provide significant insights for agricultural planning, drought alleviation, and the creation of data-informed decision-support systems. The combined use of several drought indices and multivariate analysis underscores the potential for adaptive agricultural practices and resilient crop management in response to changing climate conditions.
 Agricultural drought indicators Corn and soybean yields SPI SPEI PDSI Crop Moisture Index (CMI) Random Forest regression

Details

Metrics

1 Record Views
Logo image