Forest fires are a global issue due to their significant degradation of forest reserves and greenhouse gas emissions, as well as loss of human lives and livelihoods. Forest fires are mostly caused by nature (lightning-induced fires) and human activities (anthropogenic fires). Lightning-induced fires in China mostly occur in boreal forest, namely the Daxing'an Mountains of Heilongjiang province and Hulunbeier of Inner Mongolia. Lightning accounts for nearly a third of all forest fires in the Tahe area of the Daxingán Mountains. Most previous studies on lightning-induced fires have focused primarily on climatic factors, and studies of non-climatic factors such as forest fuel and terrain features are relatively rare, due to a lack of spatial data sets and spatial analysis technology. Thus, the aim of this study was to identify the key climatic and non-climatic factors driving lightning-induced fires in the Tahe area using fire occurrence and metrological data along with digital forest maps in conjunction with logistic regression models and spatial analysis.
Fire occurrence data included location, time, and area burned of lightning-induced fires in the Tahe region, Daxing'an Mountains, 1974-2009. Meteorological data were daily minimum temperature, daily maximum temperature, maximum wind speed, 24 hour precipitation, average air pressure, average wind speed, average relative humidity, sunshine hours, and minimum relative humidity. In addition, we calculated the Fine Fuel Moisture Code (FFMC), Duff Moisture Code (DMC) and Drought Code (DC) according to the Canadian forest Fire Weather Index (FWI). In this study, 1:100000 digital geographic and forest maps of the Tahe region were used to extract elevation, slope, aspect, depth of humus layer, litter cover, forest type, management regime, dominant tree, age class and canopy data in order to determine the factors driving lightning-induced fire occurrence in the study area.
A logistic regression model was developed to examine the relationship between lighting-induced fire, and climatic and non-climate factors. The spatial distribution of lighting-induce fires was analyzed using ArcGIS10.0. Three climate factors (Daily minimum temperature, maximum wind speed and minimum relative humidity) and two fuel indices (FFMC and DC) were significantly associated with lightning-induced fires (P < 0.05), and the goodness-of-fit of the model was R² = 0.326. Moreover, litter cover and tree age class were significantly related to the occurrence of lightning-induced fires, albeit with low R² (0.15). A map of fire likelihood was created using Kriging interpolation in ArcGIS, and the spatial coordinates of lightning-induced fires (1974-2005) along with the same number of random control points. This identified four high lightning-induced fire-risk regions in our study area, which are located in the middle and South of the Tahe area. In conclusion, the results from this study provide evidence that the consideration of not only climatic, but also fuel and non-climatic factors, is critical for understanding and predicting the occurrence of lightning-induced fires in the Tahe area.