论文部分内容阅读
地质灾害成因复杂,其中以气象因素、地质地貌因素引发的地质灾害最为常见。以金华地区为例,通过对金华市地质地貌条件及其对地质灾害点的调查,将全区划分为4个地质灾害隐患风险等级的网格区域。在此基础上利用金华中尺度气象资料,采用BP神经网络模型,建立地质灾害细网格预报模型,对该模型进行模拟和预报试验。结果表明,合理的隐患风险等级分区能使预报模型更符合科学规律,而采用分布较细的中尺度资料作为预报因子能进一步提高预报精度。模型的预报结果达到一定的可信度,为防灾减灾工作提供了科学依据。
The causes of geological disasters are complex, of which geological disasters caused by meteorological factors and geological and geomorphological factors are most common. Taking Jinhua area as an example, this paper divides the whole area into four grid areas of hazard risk level of geological disasters according to the geological and geomorphic conditions of Jinhua City and its investigation of geological disasters. Based on this, the mesoscale meteorological data of Jinhua is used to build a fine grid forecast model of geological hazards using BP neural network model. The model is simulated and forecasted. The results show that a reasonable classification of risk level can make the prediction model more scientific, and using the mesoscale data with fine distribution as a predictor can further improve the prediction accuracy. The forecast result of the model reaches a certain degree of credibility and provides a scientific basis for disaster prevention and mitigation.