Fine assessment of tropical cyclone disasters based on GIS and SVM in Zhejiang Province, China

Tropical cyclones represent major natural disasters in low- and mid-latitude coastal areas. Effective assessment of tropical cyclone disasters provides a scientific reference for the formulation of tropical cyclone prevention and disaster-relief measures. Tropical cyclone disasters in Zhejiang Province are mainly studied based on GIS technology, by considering disaster-causing factors, disaster-affected bodies, the disaster-formative environment, and spatial distribution of disaster prevention and relief capacity. In light of an uncertain nonlinear relationship between assessment factors and disaster factors, we used support vector machines to establish a fine, quantitative assessment model. This model evaluates the following disaster indices: Disaster-affected population, direct economic loss, affected crop area, and number of damaged houses resulting from a tropical cyclone disaster in Zhejiang, with the county as basic assessment unit. Assessment of tropical cyclone No. 0908 shows that the developed assessment model is able to accurately evaluate the geographical distribution of losses caused by a tropical cyclone.