Abstract: Xishuangbanna is an important part of the ecological security barrier in southwest China and a key area for global biodiversity protection. We analyzed the spatiotemporal dynamics of MODIS fire points in Xishuangbanna from 2011 to 2020 using spatial correlation analysis and probability density estimation of fire points and natural and anthropogenic elements. The results showed that: (1) Global Moran’s index of the MODIS fire points in Xishuangbanna exceeded 0.94, with significant correlation in spatial distribution; (2) the temporal distribution of the MODIS fire points had an obvious seasonality, with fire points mainly occurring between February and April; (3) when the significance level was set at 0.01, the probability distributions of the fire points’ elevation, slope, distance from the water system, fractional vegetation coverage in winter and summer, precipitation, air temperature, wind speed, and humidity all followed Gamma distribution, indicating that these natural factors all contributed significantly to fire points; (4) the kernel density estimation results showed that MODIS fire points were mainly distributed in the forests within the distance of 5-8 km from the residential areas and within 2.5 km from the roads (accounting for 58.7%). Our results provided a scientific basis for understanding the pattern of fire occurrence, and a reference for fire prevention and control as well as ecological protection in tropical rainforest areas.

Key words: MODIS fire point, kernel density estimation, spatial correlation, Xishuangbanna.