Linking spectral recovery from remote sensing to ground observations at the Elephant Hill wildfire

The 2017 Elephant Hill wildfire is considered as one of the most destructive fires in Canada. Wildfires are a major ecosystem disturbance which also causes residential displacement and financial loss. Monitoring vegetation recovery following wildfires becomes crucial to rebuilding the local community and ecological system. Since recovery rates may have discrepancies between ground data and remote sensing data, this project aimed to investigate if the post-fire spectral recovery based on Landsat time series analysis can accurately capture vegetation recovery as ground plots indicate. A total of 20 Landsat scenes with the 30-meter resolution from a year before to three years after the fire event were used to analyze the spectral recovery at the Elephant Hill region. Two spectral indicators were calculated in the analysis including normalized difference vegetation index (NDVI) and normalized burn ratio (NBR). Clear spectral recovery patterns were observed from resulting maps and time series scatterplots while there was a significant decline of NDVI and NBR in the year of 2017. NDVI and NBR values have been gradually increasing after the fire and reached pre-fire pixel values by the year 2020. However, the results did not find any significant relationships between the NDVI / NBR values and vegetation cover percentage. Since this research focused on understory recovery with almost half of the plots having greater than 30% of crown closure, it is challenging to evaluate ground vegetation types and coverage in detail with 30-meter-spatial-resolution data. Therefore, the future study can consider sampling ground data with low canopy cover every year.2017年象山野火被认为是加拿大破坏性最强的野火之一。野火作为一类主要的生态系统干扰因子，同时会造成居民流离失所与经济损失。野火后的植被恢复监测，对于当地社区重建与生态系统修复而言至关重要。鉴于地面实测数据与遥感数据的植被恢复速率可能存在差异，本项目旨在探究基于Landsat时间序列分析的火后光谱恢复，能否如地面样地所反映的那样，精准捕捉植被恢复情况。本研究共使用20景30米分辨率的Landsat影像，覆盖野火发生前1年至野火后3年的时段，用于分析象山地区的光谱恢复特征。分析过程中计算了两类光谱指标，分别为归一化差分植被指数（Normalized Difference Vegetation Index，NDVI）与归一化燃烧比（Normalized Burn Ratio，NBR）。从生成的结果影像与时间序列散点图中可观测到清晰的光谱恢复模式，2017年的NDVI与NBR值则出现了显著下降。野火过后，NDVI与NBR值逐步回升，并在2020年达到火前像素水平。然而，研究结果并未发现NDVI/NBR值与植被覆盖百分比之间存在显著相关性。由于本研究聚焦于林下植被恢复，且近半数样地的冠层郁闭度超过30%，使用30米空间分辨率的数据难以详细评估地面植被类型与覆盖度。因此，未来研究可考虑每年对低冠层覆盖区域开展地面数据采样。