Spatio-temporal evolution of complex agricultural land use and its drivers in a super-large irrigation district (Hetao) of the upper Yellow River Basin (2000–2021)

Accurate spatio-temporal land cover information in agricultural irrigation districts is crucial for effective agricultural management and crop production. Therefore, a spectralphenological-based land cover classification(SPLC) method combined with a fusion model(flexible spatiotemporal data fusion, FSDAF)(abbreviated as SPLC-F) was proposed to map multi-year land cover and crop type(LC-CT) distribution in agricultural irrigated areas with complex landscapes and cropping system, using time series optical images(Landsat and MODIS). The SPLC-F method was well validated and applied in a super-large irrigated area(Hetao) of the upper Yellow River Basin(YRB). Results showed that the SPLC-F method had a satisfactory performance in producing long-term LC-CT maps in Hetao, without the requirement of field sampling. Then, the spatio-temporal variation and the driving factors of the cropping systems were further analyzed with the aid of detailed household surveys and statistics. We clarified that irrigation and salinity conditions were the main factors that had impacts on crop spatial distribution in the upper YRB. Investment costs, market demand, and crop price are the main driving factors in determining the temporal variations in cropping distribution. Overall, this study provided essential multi-year LC-CT maps for sustainable management of agriculture, eco-environments, and food security in the upper YRB.