Site-scale drivers of post-fire vegetation regrowth in gullies: A case study in Mediterranean Europe

Abstract

Mediterranean forests are very degraded, mainly due to the intensification of wildfires in recent decades, which, boosted by human activity, have contributed to the acceleration of erosion processes and soil degradation. Under certain conditions, this also contributes to the formation of gullies. The aim of this study is to identify and characterise gullies considering their morphological and topographical aspects and determine the factors that control vegetation regrowth in gullies in a Mediterranean environment after a wildfire. The gullies were identified based on the 2018 orthophotograph, after the large wildfire of October 2017 that affected the entire study area. To analyse the vegetation regrowth, we used the normalised difference vegetation index (NDVI) derived from seven Landsat 8 OLI/TIRS images (2017–2022). The Spearman's rho correlation coefficient was selected to estimate the correlation between gully characteristics and vegetation regrowth. Before running the model, a multicollinearity test was conducted (VIF ≤ 10 and tolerance ≥ 0.1). Stepwise multiple regression was conducted in order to identify the independent variable that has a strong relationship with vegetation regrowth. A marginal effects plot was drawn up. The 38 gullies identified are in forest areas, mainly composed of pine (Pinus pinaster) trees (17 gullies) or a combination of pine trees and broadleaf (Eucalyptus globulus) trees (eight gullies). In all, invasive species are present in 11 gullies, alone (one gully), together with pine trees (four gullies) or with other species (six). The other gully has broadleaf trees. The vegetation in the gully channel recovered well in the year after the wildfire. In the following years there was growth at a slower rate until it reached similar values of NDVI in 2022, 5 years after the wildfire. Stepwise multiple regression (SMR) produced a solution with three models. The three derived dimensions covered 66.8% of the variance, considering the mean width, altitude and flow accumulation. The results can help to devise more effective management strategies for areas where the recurrence and intensity of wildfires have contributed very effectively to soil loss and degradation from gully erosion, with a view to a more resilient and sustainable territory.