Geographia Technica, No. 1/2010, pp. 25 - 35

EMPIRICAL MODELLING OF WINDTHROW RISK USING GIS AND LOGISTIC REGRESSION

 L. Krejci

ABSTRACT: This paper is attempting to shed light on the basic approaches and methods (observational, empirical and mechanistic approach) of assessing wind damage hazard using widely available GIS software. The paper describes data acquisition and building dataset, determination of dependant and independent (explanatory) variables, creation of sample units and application of basic statistical methods to assess windthrow risk. The statistical method of logistic regression is used to predict the probability of this hazard. Logistic regression is one of the most commonly used tools for applied statistics and data mining and has proven to be a useful tool to estimate the probability of windthrow. Variables and coefficients of logistic model are calculated using statistical software SAS 9.1. Stepwise method is applied to aid in the formulation of model and finding the most important explanatory variables. Logistic regression formulas are incorporated into GIS and a windthrow hazard map is then derived from the model using raster calculator in ArcGIS Spatial Analyst. The potential for spatial prediction of wind damage using logistic regression, and it’s results are discussed at the end of the paper.
Keywords: windthrow, GIS, natural hazard, logistic regression, probability

Full article: krejci_empirical_modelling_of_windthrow_risk_using_gis_and.pdf