土地利用模型时间尺度预测能力分析——以CLUE-S模型为例

Analysis of temporal predicting abilities for the CLUE-S land use model

模型模拟是生态学中的重要方法,特别是当实验不可进行时。在不同预案下基于模型的土地利用预测对于土地利用规划和政策制定具有十分重要意义。然而,很多研究没有对模型在研究区的时间尺度预测能力加以分析,从而可能导致模拟结果的不可靠。以岷江上游地区为例,采用Kappa指数系列对CLUE-S模型在研究区的时间尺度预测能力进行研究。结果表明CLUE-S模型在岷江上游地区时间尺度上的最大预测能力为22a,超过时间预测能力的预测结果不可靠。研究为土地利用模型模拟时间尺度确定提供了一种有效的方法。

Model simulation is an important approach in ecological studies when actual field experiments are unfeasible. Predicting future land use under different scenarios using land use models is necessary to assist land use planning and policy making. However, many studies failed to analyze the predicting ability in temporal scales, which may have incorrect simulation results. This study investigates the temporal predicting ability of a land use model （CLUE-S） based on kappa coefficient in the upper reach of Minjiang River. Our results showed that CLUE-S can be used to predict up to 22 years and beyond which the prediction results may become unreliable. Our study provides a usable approach in examining temporal variations in land use models.