Assessment of Wildfire Risk at Recreational Sites in Sri Lanna National Park, Chiang Mai, Northern Thailand, using Remote Sensing and GIS Techniques

Abstract

Wildfires have had massive impacts on recreational areas in national parks in Thailand that have seen a decline in tourism activities due to fire damage. Therefore, prevention and control of wildfires can minimize fire damage at recreation sites for existence of national parks. This study integrates remote sensing and geographic information system (GIS) techniques for modeling and mapping wildfire risks to evaluate the potentials for fires at recreational sites. The factors comprehensively recognized as influencing wildfire occurrence, namely leaf fuel load, slope, aspect, elevation, distance from roads, and proximity to settlements were selected for establishing a wildfire risk model. A differenced normalized burn ratio (dNBR) was used to rate wildfire sensitivity for subclasses of each factor. All factors were then weighted using pairwise comparison, where the leaf fuel load was the most important factor and we show that soil moisture can be used as a new factor for modeling wildfire risk. Our model correctly classified 74.67% of real wildfire instances, confirming that the selected factors could be used for mapping wildfire-prone areas. The mapping showed three different categories of wildfire risk; 22.15% of the study area was predicted to be a high wildfire risk zone, and 42.25% and 35.60% were categorized as moderate and low risk, respectively. A map of wildfire risk zones was overlaid with recreation sites in Sri Lanna national parks, revealing that 6 of 22 recreation sites were at high risk from wildfires. Hence, this study contributes to reducing wildfire threats to recreational areas and can help develop appropriate method for accessing areas prone to wildfires.