Application of advanced clustering methods in a GIS based forest fire risk assessment tool

Summarization: In recent years the forest fire risk management practices tend to a convergence both in Europe and worldwide. The majority of countries in Europe and especially in the Mediterranean area adopt various indices for the fire risk assessment. The aim of the present paper is to describe a GIS based Informative System for the dynamic assessment of the forest fires risk through the use of forest fire risk assessment indicators where meteorological and climate parameters are incorporated. The fire risk index developed combines the Fire Hazard Index (FHI) which is a measure the fire initiation danger attributed to flammable fuels and to the forest topology with the Fire Danger Index (FDI) which indicates the possibility of forest fire ignition based on meteorological parameters. The aforementioned indices provide an indication of the fire risk on a five scale categorization ranging from 0 (no fire danger) to 5 (increased fire danger). These five classes are usually defined by experts in a heuristic way. In the present paper advanced clustering techniques for the fire risk index classification are developed and tested. The clustering techniques tested are Fuzzy, K means and Gaussian. The clustering techniques which are incorporated in a GIS based Information System which provides the necessary information on-lineΠαρουσιάστηκε στο: 3rd International Conference on Industrial and Hazardous Waste Managemen