Data from: Optimizing restoration outcomes in the Brazilian Cerrado: A spatial planning framework integrating endemic lizard biodiversity, connectivity, and costs
Description
Given the extensive loss and fragmentation of natural areas in the Brazilian Cerrado, ecological restoration of degraded lands is essential to halt biodiversity decline in this globally significant hotspot. Ambitious restoration targets, limited financial resources, and data-deficient scenarios demand strategies that address these challenges while ensuring cost-effectiveness. We developed an innovative spatial prioritization framework using the prioritizr package in R to identify priority areas for restoration in the Cerrado, integrating endemic lizard biodiversity, functional connectivity, and restoration costs. For species with few occurrence records, we built Species Distribution Models (SDMs) based on phylogenetic inference, applying the novel ENphylo method recently proposed in the literature, while conventional SDMs were used for species with sufficient records. By combining total species richness and the richness of threatened species as proxies for biodiversity value, we accounted for both conservation urgency and representativeness. Functional connectivity was assessed by quantifying the contribution of each restorable area to landscape connectivity, whereas restoration cost was estimated using natural regeneration potential as a proxy. The optimization produced an efficient and replicable solution that identified four focal regions for restoration—central, southeastern, western, and northern Cerrado—balancing ecological value with implementation feasibility. These results highlight the need for strategic restoration in regions under heavy agricultural pressure and habitat fragmentation. Our approach advances restoration planning in data-deficient contexts by integrating cutting-edge biodiversity modeling and optimization tools, offering a robust framework to inform national and global restoration policies and biodiversity planning efforts across tropical ecosystems facing similar conservation and socio-economic challenges.
Files
Steps to reproduce
We developed an innovative spatial prioritization framework using the prioritizr package in R to identify priority areas for restoration in the Cerrado, integrating endemic lizard biodiversity, functional connectivity, and restoration costs. We used the 2023 land use and land cover map from MapBiomas Collection 9.0 for the Cerrado to identify the candidate areas. Due to computational limitations, the map was resampled to a 15 km resolution and reclassified into four categories. For the prioritization analysis, we used only the pixels classified as non-native/non-vegetated. As proxies for biodiversity value, we considered (1) the total species richness and (2) the richness of threatened species, weighted by their threat level. We fit correlative Species Distribution Models (SDMs) and used the binary predictions to characterize the presence area and calculate richness values for the spatial planning framework. We used expert-validated occurrence records for 43 endemic lizard species from the Cerrado and downloaded historical climate and elevation data from the WorldClim 2.1 database to fit the models. We used the Random Forest (RF) machine learning algorithm to fit the SDMs for 18 species with more than 20 records. For the 25 species with 20 or fewer records, we used the ENphylo algorithm. ENphylo is an innovative approach designed to model the distribution of rare species through phylogenetic imputation and outperform other algorithms’ predictive performance under these conditions. We used the Integral Index of Connectivity (IIC) to evaluate each pixel's contribution to functional connectivity. Two connectivity layers were developed: one using native vegetation within strictly protected areas (PAs) as nodes, and another including all native vegetation. The PAs shapefiles were obtained from the Brazilian National Registry of Protected Areas (CNUC). We used a continuous surface of natural regeneration potential of Cerrado pastures, produced by Silva et al. (2023), as a proxy for restoration costs, interpolating missing values with universal kriging. The surface was then inverted so that areas with higher regeneration potential represented lower restoration costs.
Institutions
- Universidade de Brasilia