Radar and multispectral remote sensing data accurately estimate vegetation vertical structure diversity as a fire resilience indicator

Abstract

The structural complexity of plant communities contributes to maintaining the ecosystem functioning in fire-prone landscapes and plays a crucial role in driving ecological resilience to fire. The objective of this study was to evaluate the resilience to fire off several plant communities with reference to the temporal evolution of their vertical structural diversity (VSD) estimated from the data fusion of C-band synthetic aperture radar (SAR) backscatter (Sentinel-1) and multispectral remote sensing reflectance (Sentinel-2) in a burned landscape of the western Mediterranean Basin. We estimated VSD in the field 1 and 2 years after fire using Shannon's index as a measure of vertical heterogeneity in vegetation structure from the vegetation cover in several strata, both in burned and unburned control plots. Random forest (RF) was used to model VSD in the control (analogous to prefire scenario) and burned plots (1 year after fire) using as predictors (i) Sentinel-1 VV and VH backscatter coefficients and (ii) surface reflectance of Sentinel-2 bands. The transferability of the RF model from 1 to 2 years after wildfire was also evaluated. We generated VSD prediction maps across the study site for the prefire scenario and 1 to 4 years postfire. RF models accurately explained VSD in unburned control plots (R 2 = 87.68; RMSE = 0.16) and burned plots 1 year after fire (R 2 = 80.48; RMSE = 0.13). RF model transferability only involved a reduction in the VSD predictive capacity from 0.13 to 0.20 in terms of RMSE. The VSD of each plant community 4 years after the fire disturbance was significantly lower than in the prefire scenario. Plant communities dominated by resprouter species featured significantly higher VSD recovery values than communities dominated by facultative or obligate seeders. Our results support the applicability of SAR and multispectral data fusion for monitoring VSD as a generalizable resilience indicator in fire-prone landscapes.

Using Pre-Fire High Point Cloud Density LiDAR Data to Predict Fire Severity in Central Portugal

“…Our results highlight the synergies between different LiDAR metrics for predicting fire severity on a continuous scale. This can be inferred from the relatively low, but homogeneous accuracies (R 2 = 0.17 ± 0.04) of the univariate linear model fits and the considerably higher accuracy (pseudo-R 2 = 0.57) of the RF regression model, capable of capturing complex interactions within the range of variation of different predictors, and complex relationships with the dependent variable [87]. The latter can also explain the ranking of the D 4-10m metric as the fifth most important variable in the RF model and its lack of significance in the univariate model.…”

“…RF is an ensemble machine learning algorithm based on the fitting of multiple classification and regression trees (CART) [84] through bootstrap aggregating techniques, which help to improve the stability and accuracy of the algorithm [85]. We chose the RF algorithm because it can properly handle potential spatial autocorrelation [85] and disclose complex, non-linear relationships between the dependent variable and predictors, as well as complex interactions among predictors [86,87]. We tuned the model parameter mtry, whereas ntree parameter was set to 2000 for ensuring stable model outcomes [88].…”

Using unmanned aerial vehicle multispectral data for monitoring the outcomes of ecological restoration in mining areas

“…Hyperspectral data offers a multitude of spectral bands that enhance diversity detection potential; however, this might result in a decrease in classification accuracy in high spectral dimensions (Gholizadeh et al, 2018). Despite its advantages in characterizing vegetation vertical structure (Fernandez-Guisuraga et al, 2022), the widespread application of LiDAR is significantly limited due to its high expenses. Therefore, practitioners need to weigh both the economic costs of LiDAR and the time costs associated with hyperspectral data analysis.…”

“…The applications of UAVs in vegetation assessments in mining areas are wide‐ranging, including monitoring vegetation growth conditions such as soil temperature (Ruan et al, 2022); mapping vegetation communities, vegetation structure and height (Banerjee & Raval, 2022; Tang et al, 2022); and assessing biomass (Ren et al, 2022). Multispectral sensors can be used to monitor ecological indicators such as vegetation coverage, aboveground biomass, and tree crown coverage (Fernandez‐Guisuraga et al, 2022; Villoslada et al, 2020), and can provide an alternative to high‐cost airborne hyperspectral and LiDAR approaches. However, few studies have used spectral information provided by UAV remote sensing integrated with site‐based data to identify plant species and provide estimates of diversity.…”

“…Multispectral sensors can be used to monitor ecological indicators such as vegetation coverage, aboveground biomass, and tree crown coverage (Fernandez-Guisuraga et al, 2022;Villoslada et al, 2020), and can provide an alternative to high-cost airborne hyperspectral and LiDAR approaches. However, few studies have used spectral information provided by UAV remote sensing integrated with site-based data to identify plant species and provide estimates of diversity.…”