Forest Spectral Recovery and Regeneration Dynamics in Stand-Replacing Wildfires of Central Apennines Derived from Landsat Time Series

Donato Morresi,Alessandro Vitali,Matteo Garbarino,Carlo Urbinati

doi:10.3390/rs11030308

Donato Morresi, Alessandro Vitali + Show 2 more

Open Access

https://doi.org/10.3390/rs11030308

Copy DOI

Abstract

Understanding post-fire regeneration dynamics is an important task for assessing the resilience of forests and to adequately guide post-disturbance management. The main goal of this research was to compare the ability of different Landsat-derived spectral vegetation indices (SVIs) to track post-fire recovery occurring in burned forests of the central Apennines (Italy) at different development stages. Normalized Difference Vegetation Index (NDVI), Normalized Difference Moisture Index (NDMI), Normalized Burn Ratio (NBR), Normalized Burn Ratio 2 (NBR2) and a novel index called Forest Recovery Index 2 (FRI2) were used to compute post-fire recovery metrics throughout 11 years (2008–2018). FRI2 achieved the highest significant correlation (Pearson’s r = 0.72) with tree canopy cover estimated by field sampling (year 2017). The Theil–Sen slope estimator of linear regression was employed to assess the rate of change and the direction of SVIs recovery metrics over time (2010–2018) and the Mann–Kendall test was used to evaluate the significance of the spectral trends. NDVI displayed the highest amount of recovered pixels (38%) after 11 years since fire occurrence, whereas the mean value of NDMI, NBR, NBR2, and FRI2 was about 27%. NDVI was more suitable for tracking early stages of the secondary succession, suggesting greater sensitivity toward non-arboreal vegetation development. Predicted spectral recovery timespans based on pixels with a statistically significant monotonic trend did not highlight noticeable differences among normalized SVIs, suggesting similar suitability for monitoring early to mid-stages of post-fire forest succession. FRI2 achieved reliable results in mid- to long-term forest recovery as it produced up to 50% longer periods of spectral recovery compared to normalized SVIs. Further research is needed to understand this modeling approach at advanced stages of post-fire forest recovery.

Highlights

The interaction between climate and land-use changes is raising the frequency, surface area, and severity of wildfires in the Mediterranean Basin [1,2,3,4]
Post-fire variations were found to be higher in those spectral vegetation indices (SVI) contrasting one of the two SWIR bands and the NIR band (NDMI, Normalized Burn Ratio (NBR)) compared to those using the Red and NIR bands (NDVI) or both the SWIR bands (NBR2) (Figure 3)
The ability of normalized SVIs to track the rapid spectral recovery occurring soon after the fire seems related to their sensitivity toward non-arboreal vegetation dynamics rather than to tree canopy cover development, as observed in comparisons with Integrated Forest z-score (IFZ) [42] and Forest Recovery Index (FRI) [41]

Summary

Introduction

The interaction between climate and land-use changes is raising the frequency, surface area, and severity of wildfires in the Mediterranean Basin [1,2,3,4]. The development of appropriate management strategies is essential to prevent fire occurrence and to enhance forest recovery [6,7]. The latter is a critical ecological process after a stand-replacing disturbance, referring to the re-establishment or re-development of forest biomass and canopy structure [8,9,10]. This process affects regional and global carbon cycles [11,12] and promotes numerous ecosystem services [8]. Depending on the fire severity level, it might be a very diversified one [9], with great changes in forest structure and species composition [13]

Objectives

Methods

Results

Discussion

Conclusion