The performance of wildfire danger indices: A Swedish case study

Wildfire danger indices, or fire danger rating systems, are widely used to inform decision-making related to wildfire risk management. Several studies have compared different indices’ performance; however, results have shown that this varies in different parts of the world. Further, many of these indices have not yet been assessed for Swedish conditions. In this study, four different weather-based wildfire danger indices have been investigated by comparing their performance in predicting wildfire activity when applied to Swedish conditions. Daily index values were calculated for seven Swedish geographical regions during 2018, a year with high wildfire activity. The aim of the study has been to rank the ability of the methods to predict wildfire danger in Sweden which is why a single year with high fire frequency has been chosen as the “test year”. The daily index values were compared to wildfire incident data, i.e.: fire and rescue service personnel hours spent on a fire, total burned area, and number of daily fire ignitions. Kendall's Tau_b correlation coefficient was calculated for each index and wildfire data in the seven regions. The indices were then ranked based on the strength of the correlation. It was found that three of the indices (those based on cumulative weather data) exhibited a significantly higher correlation with wildfire activity than the fourth index. Further, the Fire Weather Index, developed in Canada and currently used in Sweden, was identified as a good choice for Swedish conditions when compared to the three other indices.