Lagged crop responses to drought across China: contrasting multivariate environmental pathways in rainfed and...

Under climate warming, crop responses to drought are increasingly shaped by the interactions among climate, soil, topography, and human regulation. However, widely used drought indices may not adequately represent crop-perceived water deficits, and the ways in which lag effects interact with multiple environmental drivers to regulate crop responses remain insufficiently understood. In this study, we evaluated lagged correlations between three drought indices-the standardized evapotranspiration deficit index (SEDI), the standardized precipitation evapotranspiration index (SPEI), and the standardized soil moisture index (SSMI)-and multiple vegetation-related indicators across irrigated and rainfed croplands in China. We further characterized crop drought-response lags and used partial least squares structural equation model (PLS-SEM) to disentangle the pathways through which environmental drivers regulate crop dynamics. The results showed that SEDI outperformed SPEI and SSMI in capturing crop-relevant drought stress, with consistently stronger and more stable relationships in both irrigated and rainfed croplands. Crops were most sensitive to drought at one-month timescale, with irrigated croplands exhibiting a clearer pattern and more immediate responses. The PLS-SEM further revealed distinct pathways: in rainfed croplands, crop growth was mainly driven by precipitation but constrained by soil properties and human activities, whereas in irrigated croplands, anthropogenic water inputs reshaped the linkages among energy, water, topography, soil, and crop processes, weakening direct climate constraints and amplifying the regulatory effects of human activities. These findings clarify how environmental drivers interact with drought lag effects to regulate crop growth, providing a process-based foundation for drought-resilient agricultural management and climate adaptation strategies.