Identification of key meteorological factors influencing crop evapotranspiration using time-frequency domain analysis

Abstract

Crop evapotranspiration (ET_c) is a critical factor for understanding water demand in agricultural systems, influencing irrigation scheduling and water resource management. Identifying the meteorological factors influencing ET_c is crucial for predicting variations in water needs and optimizing irrigation plans. Traditional correlation analysis methods, such as Pearson correlation, often fail to capture the time-frequency variations in ET_c, which limits their ability to effectively identify the primary influencing factors. This study integrates the Penman–Monteith model, Pearson correlation analysis, wavelet analysis, and vector projection length calculation method to propose a comprehensive approach for identifying primary and secondary meteorological influences on ET_c from a time-frequency perspective. Using rice (Oryza sativa) in the Gaoyou Irrigation District of Jiangsu Province, China, as a case study, the research examines seven meteorological factors—including air temperature, relative humidity, rainfall, and sunshine duration—along with four circulation indices, such as the East Asian Summer Monsoon index and ENSO index, from 1980 to 2021. The results indicate that sunshine duration and relative humidity are significant factors affecting the high-frequency and low-frequency signal components of local rice ET_c, respectively. Additionally, other factors, such as minimum temperature, show strong correlations with ET_c signals within specific frequency bands, positioning them as secondary influencing factors. This research presents a versatile framework that can be extended to other areas of hydrometeorology and beyond.