Influencia de la topografía en la severidad de un incendio forestal de bosque seco tropical en la Cuenca Alta del Río Magdalena.

Abstract

In the last fifty decades, multiple fires have been registered in the Colombian tropical dry forest, this ecosystem is one of the most vulnerable to these continuous disturbances. However, the studies of these fires are based on the climatic characteristics and the types of coverage, ignoring the topographic which affects the occurrence and behavior of the fires. Therefore, this study evaluated the severity of a fire that occurred in a tropical dry forest in the municipality of Honda - Tolima and the order of importance of the topographic variables (Slope, Aspect, TPI and TWI) was determined. Decision Trees and Random Forests machine learning models were used to model the probability of severity prediction based on topography. The results of the present study concluded that the influence of the topography largely defines the different severities of the fire, showing that the slope, the aspect and the topographic position are variables with the greatest influence on the classification of fire severities, while the Topographic humidity was not relevant in this determination. Therefore, it is important to take into account topographic variables when evaluating severity in order to generate mitigation and management actions before fire in highly vulnerable areas.