基于ANN-CA模型的重庆市多级流域耕地景观格局演化及预测

为解决耕地演化的尺度效应问题，研究弥补现有不足，创新性地从多级流域尺度出发，基于ANN-CA模型对重庆市2000—2030年耕地景观格局演化进行模拟及预测，不仅为山区耕地的研究尺度提供新视角，且有助于解读重庆市耕地利用现状与趋势，为其耕地保护和规划提供参考。结果表明：(1) ANN-CA模型的模拟结果精度达89.33%,适用于模拟山区耕地演化研究；(2)在城市扩张、退耕还林和耕地撂荒多重影响下，2000—2020年重庆市耕地数量以先慢后快的速度由40 640.87 km²下降至38 590.33 km²,大部分转为城市建设用地，其次是林地、草地；(3)一级流域下，2000—2020年耕地规模均值由0.519缩减至0.491,耕地破碎度均值由0.217上升至0.233,耕地不规则度均值由0.311上升至0.336,耕地聚集度均值由0.657下降至0.620;(4)二级流域下，2000—2020年耕地规模均值由0.492下降至0.470;耕地破碎度均值由0.157上升至0.164;耕地不规则度均值由0.302上升至0.311;耕地聚集度均值由0....

Evolution and Prediction of Cultivated Land Landscape Pattern in Multi-level Watershed of Chongqing Based on ANN-CA model

To solve the problem of scale effect of cultivated land evolution, this study made up for the existing shortcomings and innovatively starts from the multi-level watershed scale. The evolution of cultivated land landscape pattern in Chongqing from 2000 to 2030 was simulated and predicted based on Ann-ca model, which could not only provide a new perspective for the research scale of cultivated land in mountainous areas, but also help to interpret the current situation and trend of cultivated land utilization in Chongqing, and provide a reference for its cultivated land protection and planning. The results show that:(1)the accuracy of Ann-ca model is 89.33%, which is suitable for simulating the evolution of cultivated land in mountainous areas;(2)under the multiple impacts of urban expansion, conversion of farmland to forests, and farmland abandonment, from 2000 to 2020, the amount of cultivated land in Chongqing dropped from 40 640.87 km² to 38 590.33 km² at a first slow and then fast speed; most of the cultivated land was converted to urban construction land, followed by woodland, grassland;(3)in the first-class watershed, the scale of cultivated land decreased from 0.519 to 0.491, the fragmentation of cultivated land increased from 0.217 to 0.233, the irregularity of cultivated land increased from 0.311 to 0.336, and the aggregation of cultivated land decreased from 0.657 to 0.620;(4)in the secondary watershed, the average scale of cultivated land decreased from 0.492 to 0.470; the mean value of cultivated land fragmentation increased from 0.157 to 0.164; the mean value of cultivated land irregularity increased from 0.302 to 0.311; the mean value of cultivated land aggregation decreased from 0.646 to 0.630; the cultivated land landscape pattern in the secondary watershed was roughly consistent with that in the primary watershed, but presented more refined characteristics of spatial polarization; moreover, the historical law and future trend of temporal and spatial evolution of cultivated land landscape pattern in ‘one area' and ‘two groups' were significantly different. To sum up, the multi-level watershed scale analysis of cultivated land is helpful to solve the scale effect of cultivated land evolution and deduce its evolution law across scales.