Autocorrelation between Fuel Type Fragmentation and Fire Severity at the Elephant Hill wildfire in British Columbia

Geospatial analyses focused on quantifying fuel types fragmentation and its autocorrelation with megafire severity inform decision making in contexts such as forest management and human activities regulation. Fuel type fragmentation plays a crucial role in fire severity contribution. I evaluated the landscape and class level fragmentation of fuel types in two maps: FuelSat (derived from remote sensing) and a completely randomized map. Specifically, the C-2 (Boreal Spruce), C-3 (Mature Pine), slash, and standing grass were targeted for class level metrics calculation. Fire behavior could be described in two terms – burn probability and fire intensity. Flammability (burn probability) represent the likelihood of a given location on landscape burning, while the fire intensity is the rate of heat energy released by the fire. Burn probability and fire intensity of those four target classes were extracted from landscapes (FuelSat and Random). Boxplots were created to visualize the difference between burn probability and fire intensity of four classes from FuelSat and Random, respectively. Results indicated higher fragmentated fuel types would lower the fire intensity generally, but resulted in more extreme events. It was not evident that fuel type fragmentation has significant impacts on burn probability. Moran’s I was computed and did indicate there was autocorrelation between fuel fragmentation and fire intensity. It helps fill the gap in forest fire prediction by considering effects of fuel fragmentation.

地理空间分析（geospatial analyses）聚焦于量化燃料类型碎片化（fuel type fragmentation）及其与特大型火灾严重程度（megafire severity）的自相关性（autocorrelation），可为森林管理和人类活动监管等场景的决策提供依据。燃料类型碎片化在火灾严重程度的影响中扮演关键角色。本研究评估了两种地图中燃料类型的景观层面及类别层面碎片化特征：其一为FuelSat（源自遥感技术），其二为完全随机生成的地图。具体而言，针对C-2（北方云杉，Boreal Spruce）、C-3（成熟松树，Mature Pine）、采伐剩余物（slash）及直立草（standing grass）四类，计算了其类别层面的碎片化指标。火灾行为可通过两个维度描述——燃烧概率（burn probability）与火强度（fire intensity）。可燃性（燃烧概率）代表景观中特定位置发生燃烧的可能性，而火强度则指火灾释放热能的速率。研究从上述两类景观地图（FuelSat与随机地图）中提取了这四类目标燃料类型的燃烧概率与火强度数据，并通过箱线图（boxplots）可视化两者间的差异。结果表明，燃料类型越碎片化，通常会降低火强度，但会引发更多极端事件；燃料类型碎片化对燃烧概率的显著影响未被证实；通过计算Moran’s I指数，发现燃料类型碎片化与火强度之间存在自相关性，这一发现有助于填补森林火灾预测领域中考虑燃料碎片化影响的研究空白。