Spatio-temporal changes and its driving forces of irrigation water requirements for cotton in Xinjiang, China

Irrigation is an important means of ensuring sustainable agricultural production, and therefore is critical for realizing Sustainable Development Goals. However, the persistence drought caused by climate change can deplete water resources and affect the irrigated agricultural crops negatively. Understanding changes in crop irrigation water requirement (IWR) and its meteorological driving factors can thus provide key information for long-term water resources management and agricultural planning. This is particularly true for high water -consuming crops in arid and semi-arid regions. We investigated the trends of cotton IWR at several temporal scales (monthly, annual, and growth stages) for the period of 1961-2020 based on the Trend-free Pre-whitening Mann-Kendall method. The major meteorological factors including relative contributions of reference crop evapotranspiration (ETo) and precipitation to cotton IWR were further examined using multiple linear regression and Pearson correlation analysis. Results showed that the average annual cotton IWR exhibited a significant decreasing trend over the study period, but a shift occurred after 1992. An area of increased cotton IWR was primarily located in Southern Xinjiang. Moreover, the inter-annual fluctuation of IWR at the mid-season stage was larger than at the other growth stages, and that in Northern Xinjiang was larger than in Southern Xinjiang. A statistically significant decrease in monthly IWR occurred over the entire growing season, except during the planting stage. ETo was the dominant factor affecting cotton IWR, even though average annual precipitation showed a long-term increasing trend with a rate of change of 0.615 mm/year (p < 0.01). Relative humidity and wind speed were the two main factors affecting cotton IWR in Xinjiang due to the strong correlation with ETo. Demand for irrigation water was greatest during June to August when the relative contribution of ETo was more than 80%, followed by September. Our findings demonstrated that the temporal pattern of ETo should be given greater attention when planning for wise and efficient use of regional water resources and for designing irri-gation systems.