Associate Professor Seoul National University, Republic of Korea
Wildfire related to anthropogenic climate change is increasing globally. A severe fire that combusts the whole canopy collapses the vegetation structure and directly reduces forest productivity and biodiversity. To measure and monitor the damage of fire, developing remote sensing methods have been used. However, the two-dimensional spectral sensors are only able to observe surficial information of fire-induced forests. Therefore, a new sensing method is needed to monitor the structural information of vegetation, and the Global Ecosystem Dynamics Investigation (GEDI) that is NASA-launched spaceborne LiDAR is offering three-dimensional vegetation variables. The Plant Area Volume Density (PAVD), and Plant Area Index (PAI) were well used to detect the vegetation structure and vertical distribution of a canopy. GEDI offers those two plant indices and also canopy relative height values. Our goal was to compare pre-and post-fire vegetation structure change. The pine (pinus densiflora) forest of Andong, Korea was our study site. The crown fire burnt the whole tree canopy of 1,944 ha in April 2020 for 3 days. The GEDI dataset of the study site covers a range from May 4, 2019, to November 23, 2021, and we divided them into the pre-and post-fire periods. For the accurate fire perimeter extraction, the Sentinel-2 derived post-fire Normalized Burn Ratio (NBR) was used. The negative value of the NBR raster was polygonized to set the burnt area. The variables from GEDI such as PAVD and canopy relative height 95 and Sentinel-2 derived NBR, Normalized Difference Vegetation Index (NDVI) indices were compared. And a 500 m grid was set on the perimeter, and PAVD, PAI was averaged by the grid. Their gap was calculated by subtracting the pre-fire value from the post-fire value. Through the pre-and post-fire GEDI variable comparison result, the mean tree height was decreased by around 4 m (30 percent decreased). And the mean PAVD also decreased by about 30 percent. The relationship between gap values of GEDI and Sentinel-2 derived variables had shown a positive correlation with lower R-square values. In conclusion, this study suggests the effect of fire disturbance on vegetation structure and quantitatively measured the change in tree height and PAVD value from the pine forest in South Korea.
