A spatial directivity-based sensitivity analysis to farmland quality evaluation in arid areas

Multi-criteria decision-making (MCDM) is an important means for evaluating resources and environment, and sensitivity analysis can enhance understand the robustness of evaluation results. Spatial visualization has been used in sensitivity analysis of MCDM, but the sensitivity results are still generally summarized by presenting traditional statistical measurements that omit the spatial information. To address this issue, this paper proposed a novel spatially measurement approach of sensitivity analysis by introducing the spatial barycenter model (SBM), which overcame the limitations of existing statistical methods and provided the spatial directivity of uncertainty for the MCDM results. According to our proposed method and its application in farmland quality evaluation (FQE) in an arid area of China, the mean of the absolute average change rate (MACR) and the SBM were applied to test the sensitivity of farmland quality to different evaluation factors from both numerical and spatial perspectives. From the numerical perspective, the soil organic matter and irrigation capacity were the most sensitive factors determined by the MACR. From the spatial perspective, the ≥10 °C accumulated temperature (AT) and precipitation were the most sensitive factors measured by the SBM. Based on the SBM, the spatial configuration of farmland quality index was most sensitive to increase of AT in a northwesterly direction. Calculating the SBM is computationally inexpensive and provides a straightforward indication of spatial direction for the changes of FQE results with changes of parameters. This means it can provide improved understandings and new insights into the comprehensive measurement of sensitivity analysis and agricultural production layout.