Statistical Estimation of Forest Fire Precaution Periods

In order to maximize the efficiency of the limited resources for fire-fighting and prevention, it is important to set a forest fire precaution period by considering the temporal characteristics of forest fire occurrence and spread. In this study, the forest fire precaution periods were suggested based on statistical analysis of the national fire statistics data from 1991 to 2018. Multiple, ridge, and Lasso regression analyses were applied to identify the monthly (or 10 days) linear trend of fire occurrences, while k-means cluster analysis was used to select the precaution period. The fire occurrences in January, June, August and November have increased significantly over time, whilst those in March, April and May remained high without fluctuation over this period. The results suggest the extension of the fire precaution period in South Korea from 5 months (Spring: February 1 to May 15, Autumn: November 1 to December 15) to 8 months from mid-October to mid-June. This results from this study are expected to serve as statistical evidence for readjustment of forest fire precaution period in South Korea.