A remote sensing-based method for high-resolution crop water footprint quantification in an irrigation district with complex planting structure

Abstract

The crop water footprint (WF) refers to the water consumption per unit grain production, which can reflect the quantity and types of water consumption over the entire growth period, providing a comprehensive view that is useful for agricultural water resources management. Generally, crop water consumption and crop yield are the main component of WF, but there are inconsistencies in their spatial resolution, thus it is difficult to effectively evaluate agricultural water usage under complex and variable planting structures. This study quantified crop yield and crop water consumption at fine resolution (250 m) using remote sensing methods and evaluated the spatiotemporal variability of the green water footprint (WFg) and blue water footprint (WFb) of Chinese Baojixia Irrigation District (BID). Specifically, we firstly mapped the spatial distribution of the wheat and maize field using a remote sensing-based support vector machine (SVM) method, then the partial least squares regression (PLSR) method was used to construct the yield estimation model based on multi-temporal remote sensing data, and a remote-sensing-based water balance assessment tool (RWBAT) model were constructed to quantify water consumption. Finally, the spatiotemporal variability of WFg and WFb at 250 m resolution was analyzed. The results show that the average annual WFg of wheat and maize in BID is 0.525 and 0.469 m3/kg, accounting for 81 % and 93 % of the total WF, respectively. From the perspective of spatial distribution, the low values of the total WF are concentrated in the central BID, while the high values are distributed in northwest part. The remote sensing-based method can effectively reflect the WF at high spatial and temporal resolution under complex planting structures, which can provide a basis for guaranteeing the efficient use and regulation of agricultural water resources.