Search

Nothofagus alessandrii, categorized as Endangered on the IUCN Red List, is an endemic, deciduous tree species of the coastal range of central Chile. We assessed the effects of fire severity, invasion by the exotic fire-prone Pinus radiata, and land-cover composition and configuration of the landscape on the resilience of fragments of N. alessandrii after a mega-fire in 2017. We used remote sensing data to estimate land-use classes and cover, fire severity and invasion cover of P. radiata. We monitored forest composition and structure and post-fire responses of N. alessandrii forests in situ for 2 years after the mega-fire. In the coastal Maule region wildfires have been favoured by intense drought and widespread exotic pine plantations, increasing the ability of fire-adapted invasive species to colonize native forest remnants. Over 85% of N. alessandrii forests were moderately or severely burnt. The propagation and severity of fire was probably amplified by the exotic pines located along the edges of, or inside, the N. alessandrii fragments and the highly flammable pine plantations surrounding these fragments (> 60% of land use is pine plantations). Pinus radiata, a fire-adapted pioneer species, showed strong post-fire recruitment within the N. alessandrii fragments, especially those severely burnt. Positive feedback between climate change (i.e. droughts and heat waves), wildfires and pine invasions is driving N. alessandrii forests into an undesirable and probably irreversible state (i.e. a landscape trap). A large-scale restoration programme to design a diverse and less flammable landscape is needed to avoid the loss of these highly threatened forest ecosystems.

Successful restoration measures need a good understanding of how the composition, structure, and functioning of ecosystems change with degradation and what the best indicators of these changes are. To answer these questions, we worked on four ecosystem types in the Mediterranean Andes mountains in central Chile (from sclerophyllous forest to Andean shrublands), which represent an elevational gradient from 700 to 3,250 m. We sampled three plots on each of the three degradation levels (low, medium, and high) for each ecosystem at increasing distances from goat corrals. We measured 35 indicators that describe vegetation (14), soil (15), and ecosystem processes (6) for one growing season. Degradation caused a decrease in shrub cover, shrub productivity, the Normalized Community Structure Integrity Index (CSIIn), litter depth, total soil nitrogen and C/N ratio, and an increase in clay content. Plant species indicating low degradation were consistently native woody species. When comparing ecosystems (i.e., at different elevations) against the type of variable, process-based indicators showed more statistically significant differences. Based on their consistency across ecosystems and ease of measurement, we recommend using shrub cover and litter depth as indicators of degradation. Finally, we concluded that ecosystems are highly degraded when vegetation- and process-based indicators change ∼ 60% or when soil indicators change ∼ 25%. These results could also be used to set goals for restoration projects in these mountain ecosystems.

Abstract Plant invasions can result in serious threats for biodiversity and ecosystem functioning. Reliable maps at very‐high spatial resolution are needed to assess invasions dynamics. Field sampling approaches could be replaced by unmanned aerial vehicles (UAVs) to derive such maps. However, pixel‐based species classification at high spatial resolution is highly affected by within‐canopy variation caused by shadows. Here, we studied the effect of shadows on mapping the occurrence of invasive species using UAV‐based data. MaxEnt one‐class classifications were applied to map Acacia dealbata, Ulex europaeus and Pinus radiata in central‐south Chile using combinations of UAV‐based spectral (RGB and hyperspectral), 2D textural and 3D structural variables including and excluding shaded canopy pixels during model calibration. The model accuracies in terms of area under the curve (AUC), Cohen's Kappa, sensitivity (true positive rate) and specificity (true negative rate) were examined in sunlit and shaded canopies separately. Bootstrapping was used for validation and to assess statistical differences between models. Our results show that shadows significantly affect the accuracies obtained with all types of variables. The predictions in shaded areas were generally inaccurate, leading to misclassification rates between 65% and 100% even when shadows were included during model calibration. The exclusion of shaded areas from model calibrations increased the predictive accuracies (especially in terms of sensitivity), decreasing false positives. Spectral and 2D textural information showed generally higher performances and improvements when excluding shadows from the analysis. Shadows significantly affected the model results obtained with any of the variables used, hence the exclusion of shadows is recommended prior to model calibration. This relatively easy pre‐processing step enhances models for classifying species occurrences using high‐resolution spectral imagery and derived products. Finally, a shadow simulation showed differences in the ideal acquisition window for each species, which is important to plan revisit campaigns.

Vegetation biomass is a globally important climate-relevant terrestrial carbon pool and also drives local hydrological systems via evapotranspiration. Vegetation biomass of individual vegetation types has been successfully estimated from active and passive remote sensing data. However, for many tasks, landscape-level biomass maps across several vegetation types are more suitable than biomass maps of individual vegetation types. For example, the validation of ecohydrological models and carbon budgeting typically requires spatially continuous biomass estimates, independent from vegetation type. Studies that derive biomass estimates across multiple vegetation or land-cover types to merge them into a single landscape-level biomass map are still scarce, and corresponding workflows must be developed. Here, we present a workflow to derive biomass estimates on landscape-level for a large watershed in central Chile. Our workflow has three steps: First, we combine field plot-based biomass estimates with spectral and structural information collected from Sentinel-2, TanDEM-X and airborne LiDAR data to map grassland, shrubland, native forests and pine plantation biomass using random forest regressions with an automatic feature selection. Second, we predict all models to the entire landscape. Third, we derive a land-cover map including the four considered vegetation types. We then use this land-cover map to assign the correct vegetation type-specific biomass estimate to each pixel according to one of the four considered vegetation types. Using a single repeatable workflow, we obtained biomass predictions comparable to earlier studies focusing on only one of the four vegetation types (Spearman correlation between 0.80 and 0.84; normalized-RMSE below 16 % for all vegetation types). For all woody vegetation types, height metrics were amongst the selected predictors, while for grasslands, only Sentinel-2 bands were selected. The land-cover was also mapped with high accuracy (OA = 83.1 %). The final landscape-level biomass map spatially agrees well with the known biomass distribution patterns in the watershed. Progressing from vegetation-type specific maps towards landscape-level biomass maps is an essential step towards integrating remote-sensing based biomass estimates into models for water and carbon management.

Peatlands are ecosystems of great relevance, because they have an important number of ecological functions that provide many services to mankind. However, studies focusing on plant diversity, addressed from the remote sensing perspective, are still scarce in these environments. In the present study, predictions of vascular plant richness and diversity were performed in three anthropogenic peatlands on Chiloé Island, Chile, using free satellite data from the sensors OLI, ASTER, and MSI. Also, we compared the suitability of these sensors using two modeling methods: random forest (RF) and the generalized linear model (GLM). As predictors for the empirical models, we used the spectral bands, vegetation indices and textural metrics. Variable importance was estimated using recursive feature elimination (RFE). Fourteen out of the 17 predictors chosen by RFE were textural metrics, demonstrating the importance of the spatial context to predict species richness and diversity. Non-significant differences were found between the algorithms; however, the GLM models often showed slightly better results than the RF. Predictions obtained by the different satellite sensors did not show significant differences; nevertheless, the best models were obtained with ASTER (richness: R2 = 0.62 and %RMSE = 17.2, diversity: R2 = 0.71 and %RMSE = 20.2, obtained with RF and GLM respectively), followed by OLI and MSI. Diversity obtained higher accuracies than richness; nonetheless, accurate predictions were achieved for both, demonstrating the potential of free satellite data for the prediction of relevant community characteristics in anthropogenic peatland ecosystems.