摘要
针对传统草原火蔓延模拟模型获取草原火蔓延初速度困难,模型中静态气象数据精度低的问题,依托高时间分辨率Himawari-8结合时序线性外推法代替实验获取草原火蔓延初速度,利用GRAPES_MESO气象数值预报产品作为高精度动态气象数据,地理元胞自动机结合王正非草原火蔓延模型,对一定时间内草原火场蔓延边界进行预测模拟。预测结果与Himawari-8下一周期实际影像提取的火情时空一致性较高,得到蔓延范围面积相对偏差Bias范围在1.49%-42.87%;蔓延边界均方根误差BRMSE范围在1.43-22.14个像元;像元平均绝对误差CMAPE为18.09%;模型正确模拟区域为84.68%。上述算法可用于监测草原火蔓延过程的重现及趋势预测方面的应用。
It is difficult to obtain the initial speed of grassland fire spread and the accuracy of static meteorological data in the traditional simulation model of grassland fire spread is low. Based on HIMAWARI-8 with high time resolution and time-series linear extrapolation method instead of experiments, the initial speed of grassland fire spread was obtained. By using the GRAPES_MESO meteorological numerical forecast product as high-precision dynamic meteorological data, the geographical cellular automata was combined with the Wang Zhengfei Grassland Fire Spread Model to predict and simulate the grassland fire spread boundary within a certain time. The spatio-temporal consistency between the simulation results and the actual HIMAWARI-8 image extraction is high, and the relative bias of the spread area is from 1.49% to 42.87%. The Boundary root-mean-square error(BRMSE) of the spreading boundary ranges from 1.43 to 22.14 pixels. The Cell mean absolute error(CMAPE) of pixel is 18.09%. The correct simulation area of the model is 84.68%. The algorithm can be used to reproduce and predict the spreading process of grassland fire.
作者
刘洋
玉山
都瓦拉
杨晓颖
LIU Yang;YU Shan;DU Wa-la;YANG Xiao-ying(College of Geography,Inner Mongolia Normal University,Hohhot Inner Mongolia 010022,China;Inner Mongolia Key Laboratory of Remote Sensing and Geographic Information Systems,Inner Mongolia Normal University,Hohhot Inner Mongolia 010022,China;Institute of Grassland Research,Chinese Academy of Agricultural Sciences,Hohhot Inner Mongolia 010022,China)
出处
《计算机仿真》
北大核心
2022年第9期201-207,共7页
Computer Simulation
基金
国家自然科学基金(41761101)。
关键词
草原火
火蔓延模拟
线性外推法
Grassland fire
Fire spreading simulation
Linear extrapolation method
