基于元胞自动机的月亮山景区景观格局动态模拟

Dynamic Simulation of Landscape Pattern in Yueliang Mountain Scenic Spot by Cellular Automata

研究以中国云南省"三江并流"世界自然遗产地月亮山核心景区作为实例,依据1994年和2004年Landsat TM多光谱遥感影像,基于元胞自动机(CA)模型进行景观格局的时空动态模拟,用灰色局势决策方法确定邻域转换规则,同时用Monte Carlo法考虑了模拟时转换的随机性。对4种情景进行了趋势预测:按原有模式演变;按聚集模式演变;按生态模式演变;按廊道生态模式演变。结果表明:依据研究区1994年的景观格局现状对2004年未来演变情景总趋势预测结果与实际相符;在信息不完全的情况下,邻域转换规则可以根据景观特征的动态变化随地理空间位置和时间而变化;该方法用于景观格局动态变化的预测是可行的。实现了对景观格局的动态模拟及规划方案的评估,为景观生态规划方案的确定提供了支持。

With the Yueliang Mountain Scenic Spot of the ＂Three Parallel Rivers＂ World Heritage in Yunnan Province for example,the Cellular Automata（CA）-based dynamic simulation models were applied to analyze the spatial and temporal changes of its eco-patterns according to the Landsat TM multi-spectral remote sensing image of 1994 and 2004.The method of grey situation decision was used to determine the transition rules of cellular neighborhood.Randomness of transition was handled by the Monte Carlo method.Four scenarios are predicted： following the current natural pattern;focusing on the aggregation effect;focusing on the ecological conservation;focusing on the ecological corridors.Simulation results showed that the future trends of landscape eco-pattern changes were highly in line with actual practices;when there was only incomplete information available,transition rules of cellular neighborhood might change along with dynamic landscape changes;therefore,it was advisable to apply grey situation decision-based CA to the prediction of dynamic landscape eco-pattern changes,which could not only facilitate assessment of dynamic simulation and designing,but also provide scientific grounds for decision-making in landscape eco-planning.