An uncertainty-based multivariate statistical approach to predict crop water footprint under climate change: a case study of Lake Dianchi Basin, China

Agricultural water sustainability in a basin environment experiencing climate change has become a critical issue in the past few decades. This study used the DPSIR (Driver–Pressure–State–Impact–Response) framework as a conceptual basis to explore the relationship between water footprint (WF) trends and climate change and agricultural-economic variation. With the aim of mitigating water crisis and ensuring robust responses to the uncertainty of the future, an uncertainty-based multivariate statistical approach was proposed for WF prediction by using various scenarios combined with multiple linear regression and Monte Carlo simulation. Lake Dianchi in China was used as the case study area. The results indicate that (1) the total WF had an increasing trend of 394.39 m³ ton⁻¹ year⁻¹; the WF_green (the precipitation used in the crop production process) had a decreasing trend, while the WF_blue (the irrigation water withdrawn from the ground or surface water) and WF_grey (the water used to dilute the load of pollutants, based on existing ambient water quality standards) exhibited an increasing trend; (2) the total WF showed a distinct increasing trend under climate change and agricultural-economic variation due to the increase of the WF_grey during 1981–2011; and (3) the impact of agricultural-economic factors on WF trends, especially on the WF_blue and WF_grey, far outweighed the impact of climatic factors under the alternative scenarios. Our results suggest that adaptive management of anthropogenic activities should be prioritized to mitigate water stress under climate change.