Modelagem da vulnerabilidade à ocorrência e propagação de incêndios florestais

Abstract

The systematic reduction of natural forests, from successive fires, has stimulated the development of mechanisms for prevention, control and firefighting. This resource has the objective to determine the areas of risk of forest fires (RIF) in Vale Nature Reserve, located in city the Linhares, ES. For this, used information about use and occupation of land, land slope, orientation of relief and proximity to roads and homes. With the techniques of support of Geographic Informations Systems determined the influence of each variable to RIF by Fuzzy modeling in computational application ArcGIS/ArcINFO 10.0. Each Fuzzy set represented by the matrix of the image input variable, a function was defined the pertinence which presents each set a degree of certainty, variation between 0 and 1, the greater is indicated RIF was set when the actual value of the variable assumes 1 and null when the actual value of the variable is 0 . For the variable the use and occupation of the land was used the function pertinence Fuzzy Gaussian. The roads variable was modeled by pertinence function Fuzzy Small. For the slope variable was used the function pertinence Fuzzy Large. The orientation of the relief was programmed in Python in ArcGIS/ArcInfo 10.0 by function Fuzzy Generalized Bell and the variable the proximity to residential was modeled by Fuzzy Linear function. Posteriorly, the variables in their influence on the onset and spread of fire were combined through Fuzzy Gamma to represent RIF in the study area. The biggest risk the fire was considered the physiognomy native fields with value in Fuzzy set 1. The concentration of risk values for the distance to roads corresponds mainly to areas of urban-forest interface with 41.93% of the pixels in the range of 0.95 – 1. For the variable the slope of highest concentration s of the frequency of pixels was observed in lower values in the set of Fuzzy (0 – 0.2). The orientation of the relief, although the study area presents low risk to fire, faces associated with lower solar radiation, the results show that the highest values are in the Fuzzy set in the class of highest risk, mainly associated with the North face. The proximity variable is the main factor influencing the risk of fire by distance from the urban-forest interface with 70.59% of the pixels in the range of 0.75 – 1. The study area has variant of low the medium risk. The highest risk areas represent especially in forest and urban interface. Socioeconomic factors play an important paper in fire risk in the study area and provide a useful new insight into the spatial distribution of human caused the fire. The model was adequate to evaluate the impact of different variables on the risk of fire.