The Study Of The Prediction Of The Harm's Degree Of Forest Fire In Guangzhou

The harms degree of the forest-fire is refers to the forest-fire to occur time, the area of the forest fire, suffers injury to the forest area, loss of direct and indirect economic, loss of forest resources ecology value, The forest-fire disaster area accounts for this local forest total area the percentage and the forest-fire occurs frequency degree and so on each aspect synthesis loss. This article first establishes the forest-fire to harm degree appraisal index system, and obtains each index through the AHP method the weight, to the historical data which collects carries on the forest-fire to harms degree the division;harms degree index value correspondence through the forest-fire each influence factor, through the gradually regression analysis method, establishes the forest-fire to harms degree forecast model. The primary coverage and the conclusion are as follows:(1) Actual harm creates which through the analysis forest-fire, here uses in dividing the forest-fire to harm degree the multiple coefficient of correlation for altogether to have six, respectively is:the forest-fire goes too far the area and suffers injury the wooded area, the forest-fire creates in the unit area the direct economic efficiency loss quantity,the forest-fire creates in the unit area the indirect economic efficiency loss quantity,the forest-fire creates in the unit area the ecology value loss quantity, the forest-fire disaster area accounts for the frequency abundant degree which this local forest total area the percentage and the forest-fire occurs,Harms degree appraisal index system through the establishment forest-fire, and gives each kind of concrete index the concrete computational method.(2) Extracts each index using the AHP method the weight. First first obtains the judgment matrix through the expert investigation method, then then extracts each index the weight. Through expert's investigation method, obtains the forest-fire to harm degree appraisal system various indexs the weight respectively is:The forest-fire goes too far the area and suffers injury the wooded area w1= 0.3254; The forest-fire creates in the unit area the direct economic efficiency loss quantity w2=0.1523; The forest-fire creates in the unit area the indirect economic efficiency loss quantity w3=0.0442; The forest-fire creates in the unit area the ecology value loss quantity w4= 0.2548; The forest-fire disaster area accounts for this local forest total area the percentage w5=0.1615; The forest-fire occurs frequency abundant degree w6=0.0618.(3) According to historical data division rank.Extracts the forest-fire which its corresponds to harm degree index value. Carries on the graduation to it, I.e.:light disaster (0～0.2),disaster (0.2～0.4),big disaster (0.4～0.6),heavy disaster (0.6-0.8),great disaster (0.8～1)。May know through the computation; Guangzhou's forest-fire mostly is the light disaster, accounts for the total 93.3333%; the disaster has 3 examples, accounts for the total 3.33%; the big disaster, the heavy disaster, the great disaster have 1 example separately, accounts for the total separately 1.11%. May see from the division result, forest fires of Guangzhou are mostly light calamity, this is mainly because the economy of Guangzhou is comparatively developed, and Guangzhou city build up forest fire prevention the management information system, can find the condition of a fire and can be put out in time in time already, thus make the extent of injury of the forest fire reduce greatly.(4) The extent of injury of the forest fire predicts the setting-up of the mode.This part needs to set up concrete analysis model, according to influencing the leading factor of the extent of injury of the forest fire, combine the corresponding extent of injury of forest fire to classify, by returning to the analytical method progressively, obtain the extent of injury of the forest fire and predict models: Use the experimental data and theory data to carry on the test of precision and reply the coefficient correlation R=0.998428, the precision is relatively high. And uses has not participated in the modelling the data to carry on the compatible analysis, obtains its harm degree index value to have certain deviation, but forecast obtains the harm degree accuracy is 100%. Explained this model harms the degree to the forest-fire to have the good forecast function, may use in the actual forecast, and needs to adjust gradually in the actual use process.(5) Carries on when the forest-fire harms the degree forecast research, uses the historical material fitting law. When uses this method, obtains the result often appears its computed result and the historical data has the high accuracy, to future forecast in relative accuracy not too high. This kind of question appears the reason possibly is because in the historical data record process appeared the big deviation, but actually does this kind of influence have how many, but also needs further to conduct the research.When the harms degree forecast research to the forest-fire, divides each appraisal index during some indexs has certain relation, but its relevance is not specially strong. Therefore, whether during this kind of index relation to does divide the forest-fire to harm between the degree influential, but also waits for further studies.