Partitioning evapotranspiration in partially mulched interplanted croplands by improving the Shuttleworth-Wallace model

Abstract

The joint application of plastic mulch and interplanted patterns significantly increases the utilization efficiencies of water and energy in arid regions. Each individual crop transpiration directly related with agricultural production. Soil evaporation under transparent plastic mulch is decreased but can not be ignored especially under the condition with higher soil moisture and large film mulching fraction (fm). Accurate estimation of evapotranspiration (λET) and its components (transpiration, λTrc and its sub-components (individual plant transpiration, λTrci); soil evaporation, λEs and its sub-components (bare soil evaporation, λEsbs; and mulched soil evaporation, λEsms)) for partial mulched interplanted croplands is essential for water resources management. In this study, a multiple source evapotranspiration model (SWIM) was proposed to accurately estimate λET and its components in partially mulched interplanted croplands based on Shuttleworth-Wallace model (SW). In SWIM, the effect of partial plastic mulch was accounted for by introducing fm and mulched soil resistance (rsm), while the effect of intercepted energy distribution between the intercropped plants was considered by integrating a light-interception sub-model. The performance of the SWIM model was evaluated against measurements from two partially mulched intercropped seed maize field under border irrigation (BM) and drip irrigation (DM) during different growth periods. The results showed that: (1) the SWIM model accurately estimated λET and its components and outperformed the SW model during the entire growing season, especially significantly improved these items estimations during the sparse canopy growth period, with reduced NRMSE of 0.18 and 0.33, 0.25 and 0.20, 0.26 and 0.49 for λET, λTrc, and λEs of BM and DM sites, respectively; (2) better performance of the SWIM model in λTrci estimation during the dense canopy growth period and in λEsbs and λEsms estimation during the sparse period was demonstrated, with reduced NRMSE of 0.09–0.11 and 0.11–0.12, 0.04 and 0.02, 0.02 and 0.05 for λTrci, λEsbs, and λEsms of BM and DM sites, respectively. This study will improve our understanding of eco-hydrology processes in the soil-mulch-plant-atmosphere continuum and provide a scientific basis for water resource management in arid regions.