林火行为预测和森林火险预报中气象场的插值方法

摘要

Methods for weather date interpolation currently used in forest fire danger rating and behavior modeling were reviewed. The methods can be classified into two groups: one is the direct spatial interpolation, the other is the process-based downscale model. The former includes two methods (Thiessen polygon method and Inverse distance weighted method). The latter consists of MM5 model, Regional spectral model (RSM), mesoscale spectral model (MSM), Wind Wizard and WindNinja. Based on the comparison of the two kinds of methods, the suggestions for further working were put forward as followings: (1) to establish MSM model with small scale (100 meters) and micro sacle (10 m) to satisfy the need for fire behavior modeling in complicated topography; (2) to fasten computation speed to realize real-time prediction; and (3) to conduct more validation of these models and to apply them in forest fire prevention.%介绍了目前在森林火险预报和火行为预测中使用的气象场插值方法和应用情况.这些方法可以分成直接空间插值和基于过程的降尺度模式法两类.直接空间插值方法在林火科学中应用的有泰森多边形法和地形修正的反距离权重法.基于过程的降尺度模式主要有MM5、RSM、MSM、Wind Wizard和WindNinja等.在对比两类方法的基础上,提出了对未来相关研究工作的建议:(1)建立小尺度(100米尺度)甚至微尺度(10米级)上的MSM模式,以满足复杂地形的林火行为模拟需要；(2)提高模式的运行速度,向实时化方向发展；(3)加强过程模式的验证和在森林防火工作中的应用.