Evaluation of regional agricultural drought vulnerability based on unbiased generalized grey relational closeness degree

Purpose The purpose of this study is to propose an unbiased generalized grey relational closeness evaluation model to improve the accuracy of regional agricultural drought vulnerability decision-making results, as well as to provide theoretical support for reducing agricultural drought risk and losses. Design/methodology/approach The index weight is calculated using a rough set and deviation minimization criterion, and the relational degree between the research object and the double reference sequence is thoroughly investigated using the generalized grey relational closeness degree. Because different index rankings can correspond to different closeness degrees, the Monte Carlo method was used to calculate an unbiased estimate of the generalized grey relational closeness degree, which was used as a decision basis. Findings Agricultural drought vulnerability in Henan Province in 2019 was clearly spatially differentiated. The vulnerability to agricultural drought in the southern and eastern regions was generally higher than that in other regions. The evaluation results of this model are highly stable and reliable compared to those of the traditional generalized grey relational evaluation model. Practical implications This study proposes an evaluation model based on an unbiased generalized grey relational closeness degree, which is important to supplement the grey relational analysis method system and plays a positive role in promoting the quantitative evaluation of regional agricultural drought vulnerability. Originality/value The Monte Carlo method is used to calculate the unbiased estimation of the generalized grey relational closeness degree, which solves the problem of the replacement dependence of the traditional generalized grey relational degree and the one-sidedness of the evaluation results, and provides a new research idea for the evaluation of regional agricultural drought vulnerability under cross-sectional informatics.