Study On The Occurrence Of Forest Fire In Heihe Regions

Heihe region is one of forest fire-prone areas in China. Forest fires have caused temendous damage of forest resources, so the prediction of forest fires has been the focus study on forest fires. Knowing the relevant rules of bringing out and developing of forest fires is an important project to carrying out fighting and protecting fires scientifically and effectively. The study on the prediction of forest burnt areas is of great significance in prediction of forest fres of Heihe region.This paper analyzed the incidence of forest fires in Heihe region over the past 37 years aid concluded the rules and characteristics of forest fires; based on analyzing and processing historical data of forest fires and corresponding meteorological data, the paper established the forecasting models of forest burnt area and forest fires frequency using time series analysis, further tested and verified the two models.The data from in Heihe region stations show, from 1959 to 2008, there are significantly higher trends in annual average temperature and downward trend in precipitation, but the latter is not very significant; the maximum monthly temperature is in July and minimum temperature is in January; Maximum precipitation is in July, the lowest value is in February; Range of seasonal temperature changes exits some differences, but does the same trend, in which the largest increase in average temperature is in spring; the smallest increase in average temperature is in autumn.In Heihe region, forest fires occurred most frequently between 1971 and 2008,39 times average one year. Fires of unknown cause is majority, most of fires of known causes are caused by man. Forest fires in the region is mainly in March, April, May, June, September, October, of which in April, May, June, the difference of the frequent and burnt areas of forest fires is no significant, and in March, September, October, the difference is more significant.Using time series forest burnt area model and time series forest fires frequency model to forecast separately, the results show that time series forest fires frequency model is better, and illustrate the time series analysis has better adaptability to historical fire occurrence data samples, real-time weather data samples can be brought into the model to predict the number and the damage area of fires which can achieve good results, but the prediction accuracy is slightly worse than the time series fire occurrence model.