Analyzing Fire Severity and Post-Fire Vegetation Recovery in the Temperate Andes Using Earth Observation Data

Abstract

In wildfire areas, earth observation data is used for the development of fire-severity maps or vegetation recovery to select post-fire measures for erosion control and revegetation. Appropriate vegetation indices for post-fire monitoring vary with vegetation type and climate zone. This study aimed to select the best vegetation indices for post-fire vegetation monitoring using remote sensing and classification methods for the temperate zone in southern Ecuador, as well as to analyze the vegetation’s development in different fire severity classes after a wildfire in September 2019. Random forest classification models were calculated using the fire severity classes (from the Relativized Burn Ratio—RBR) as a dependent variable and 23 multitemporal vegetation indices from 10 Sentinel-2 scenes as descriptive variables. The best vegetation indices to monitor post-fire vegetation recovery in the temperate Andes were found to be the Leaf Chlorophyll Content Index (LCCI) and the Normalized Difference Red-Edge and SWIR2 (NDRESWIR). In the first post-fire year, the vegetation had already recovered to a great extent due to vegetation types with a short life cycle (seasonal grass-species). Increasing index values correlated strongly with increasing fire severity class (fire severity class vs. median LCCI: 0.9997; fire severity class vs. median NDRESWIR: 0.9874). After one year, the vegetations’ vitality in low severity and moderate high severity appeared to be at pre-fire level.

Novel Approaches and Techniques for Understanding Vegetation Fires in South America

“…The authors concluded that the most reliable vegetation indices for monitoring post-fire vegetation recovery were the Leaf Chlorophyll Content Index (LCCI) and Normalized Difference Red-Edge and SWIR2 (NDRESWIR). They also observed that vegetation recovered to a great extent within the In a similar study also conducted in southern Ecuador, Maxwald et al [9] identified the most reliable vegetation indices for post-fire vegetation monitoring, and analyzed vegetation recovery across different classes of fire severity. The authors concluded that the most reliable vegetation indices for monitoring post-fire vegetation recovery were the Leaf Chlorophyll Content Index (LCCI) and Normalized Difference Red-Edge and SWIR2 (NDRESWIR).…”

“…In a similar study also conducted in southern Ecuador, Maxwald et al [9] identified the most reliable vegetation indices for post-fire vegetation monitoring, and analyzed vegetation recovery across different classes of fire severity. The authors concluded that the most reliable vegetation indices for monitoring post-fire vegetation recovery were the Leaf Chlorophyll Content Index (LCCI) and Normalized Difference Red-Edge and SWIR2 (NDRESWIR).…”

Assessment of Burn Severity and Monitoring of the Wildfire Recovery Process in Mongolia

“…This study of post-fire recovery concepts is very important to understand this phenomenon. There are some different approaches to define recovery processes, including utilizing the strong performance and suitability of the post-fire stability index [20]; random forest classification models that use the fire severity classes (from the Relativized Burn Ratio (RBR)) as a dependent variable and 23 multitemporal vegetation indices [21]; post-fire stream water responses observed in watersheds [22]; multiple factors of a forest's recovery rate post-wildfire such as fire severity, tree species characteristics, topography, hydrology, soil properties, and climate [23]; and a composing study based on a Tasseled Cap linear regression trend in a post-wildfire study site [24].…”