Model And Analysis On Forest Fires And Meteorological Factors In The Greater Xing’an Mountains Region

Forest fires to some extent constrained by weather conditions, forest fires in this paper based on historical data from 1990 to 2012 and the corresponding period of the Greater Xing’an Mountains of meteorological data, summarizes the variation of meteorological factors and forest fire by the method of trend line occurrence. Meteorological factors:the average annual temperature overall showing increasing trend; annual average wind speed overall downward trend, but the downward trend is not obvious; annual average relative humidity range of little change, more stable; the average annual rainfall more obvious changes, the overall downward trend. The number of fires:the forest fire frequency and interannual variability presents low high state; forest fire spread from March to October, five, June is the high incidence of forest fires; in 24 hours of forest fires mainly in 10:00 to 16:00, wherein during 13:00 to 14:00 peaked.Use the time series model and the linear regression model for the Greater Xing’an Mountains in fire frequency and the number of monthly fire modeling. Where the time series model ARIMA (2,1,2) model has the highest square of the stationary R that is 0.740, the fitted values in line with the actual value of the variation. May fire frequency and the monthly average temperature build linear equations, the model in the 0.05 significance level significantly, monthly fire frequency and the monthly average temperature showed a positive correlation; monthly fire frequency and average wind speed build linear equations, the model in 0.05 significance level under significant fire frequency and average monthly wind speed showed a positive correlation.Considering the impact of forest fires in selected large meteorological factors on thex, daily maximum temperature, x2 daily maximum wind speed, x3 daily minimum relative humidity, x4the average daily temperature, x5 daily average wind speed, x6 daily average relative humidity and x7 daily precipitation seven meteorological factors, the use of Logistic Regression establishment of forest fires to determine the model equations. Logistic regression values based on a level of 0.2, the forest fire danger rating is subdivided into five levels, so as to establish forest fire weather rating models. After modeling, through 2011-2012 fire back substitution test sample, found that most of the fire by the model sample is judged to be high fire danger levels, which fell four forest fire danger levels and five levels of the ratio reached 52.94 fire%, only 23.53 percent of the fires within a forest fire and a two-level, in line with the basic principles of division of forest fire danger rating. The logistic regression of forest fire weather classification is reasonable and can be used in the prevention of forest fires.