滑坡时空动态的元胞自动机模拟研究

Study on Spatial-temporal Dynamic Simulation of Landslide with Cellular Automata

滑坡系统是开放的复杂巨系统,传统的数理方法无法对滑坡时空特性做出比较准确的模拟和预测。元胞自动机是时空离散的动力学模型,特别适合空间复杂系统的动态模拟和预测。将元胞自动机模型和地理信息系统集成,构建滑坡滑动时空预测模型,为滑坡时空动态模拟提供新的研究思路。运用逻辑回归方法获取元胞自动机的转换规则,结合地理信息系统对滑坡进行时空动态模拟,并以杉树槽滑坡进行实证研究。该模型相对简单,不需要多种数据,成本相对较低,逐点对比的模拟精度达到82.8%,与实际滑坡较为吻合,实用性较高。该方法能为滑坡的系统研究提供新的研究视角,为滑坡的综合治理提供可视化的辅助决策支持。

Landslide is an open complex system, and the traditional mathematical methods failed to make more accurate prediction and simulation on space-time performance of landslide. The cellular automata is a spatial-temporal discrete dynamics model,which is specially suitable for simulating and predicting the spatial dynamic system. The cellular automaton model and geographical information system are integrated to construct the space-time prediction model of landslide sliding, and put forward a new thought for landslide dynamic simulation. The logistic regression model was applied to obtain the transition rules of cellu- lar automata. Coupling with the geographic information system, landslide space-time dynamic simulation is made and verified in Shanshucao landslide. The model was relatively simple ,which didn＇t require a large dataset ,so its cost was relatively low. The simulated accuracy from point by point comparison could reach 82. 8%, which was considerably coincided with actual land- slide. It has a higher practicality. The methods can provide the new thoughts for the landslide research and the visualization of decision supports for comprehensive treatment of landslide.