Empleo de regresión logística para la obtención de modelos de riesgo humano de incendios forestales

  1. Vilar Del Hoyo, L. 1
  1. 1 Consejo Superior de Investigaciones Científicas
    info

    Consejo Superior de Investigaciones Científicas

    Madrid, España

    ROR https://ror.org/02gfc7t72

Book:
El acceso a la información espacial y las nuevas tecnologías geográficas
  1. M.T. Camacho Olmedo (ed. lit.)
  2. J.A. Cañete Pérez (ed. lit.)
  3. J.J. Lara Valle (ed. lit.)

Publisher: Universidad de Granada

ISBN: 84-338-3944-6

Year of publication: 2006

Pages: 531-543

Congress: Congreso Nacional de Tecnologías de la Información Geográfica (12. 2006. Granada)

Type: Conference paper

Abstract

The simulation and making models of geospatial phenomena allow to understand the observed phenomenon, to verify hypothesis and theories, to thus predict their behaviour in the space and the time under different conditions and scenes like discovering new operations and behaviours. This communication presents the use of Logistic Regression to generate a model of ignition wildfire risk. In this model contains the factors relative to the risk produced by the human activities and their relation with the fire occurrence due to this cause in the Community of Madrid, in period 1990-2004 (resolution 1km2). The tie factors to the human activity have special relevance although in the quantification of the risk a relative difficulty in valuing them exists and to spatialise them facing obtaining models that allow to predict the beginning and the propagation of the fire. However, the identification of the different types of factors from tie risk from the human activity is possible to generate variables of risk, related to the uses of the territory (protected routes of communication, garbage dumps, natural protected areas, etc.), socioeconomic aspects (population occupied in agriculture, level of rent, rate of unemployment, etc.) and others as zones of contact with the forest use (interface) or means of forest fire monitoring, that allows to obtain sufficiently trustworthy predictive models. These models can be of great interest for the planning and establishment of prevention strategies and management of the forest spaces.