Aspectos biogeográficos e ecológicos de palmeiras de babaçu

Abstract

The Cocais Forest, an emblematic vegetation formation in the Brazilian Mid-North, constitutes a unique ecological mosaic in which palm trees represent structural elements of the landscape, extending across several states and harbouring remarkable sociobiological diversity. Despite its ecological and cultural relevance, its spatial delimitation remains marked by conceptual and cartographic uncertainties, a biogeographical challenge, as it directly affects how the landscape is understood, planned and conserved. Among its most representative species are the palms of the Babassu Complex (Attalea spp.), recognised for their wide distribution and ecological dominance, as well as their social, economic and symbolic importance to the traditional communities that depend on them. However, both the babassu palm forests and the Cocais Forest itself are under increasing anthropogenic pressure due to land use conversion, habitat fragmentation and the effects of climate change. Thus, understanding these systems requires integrating spatial, ecological and social approaches capable of translating uncertainties into instruments applicable to conservation. In this context, this study sought to fill gaps in knowledge about the biogeographical uncertainties that permeate Cocais Forest and the Babassu Complex. Initially, a systematic review of existing spatial delimitations attributed to the Cocais Forest was carried out. Among the main results, seven sources were discovered that presented significant divergences in terms of their territorial extension and definition criteria. The delineation by Nascimento and Lima (2016) covered the largest area, and the overlap of the different maps allowed us to identify a zone of consensus, which was named the ‘core region,’ comprising 14 municipalities in the State of Maranhão. This cartographic synthesis resulted in a potential area of occurrence covering 392 municipalities distributed across five Brazilian states. Next, we employed Species Distribution Modelling (SDMs) for seven species of the Babassu Complex and for the group overall, using the Maximum Entropy (MaxEnt), Random Forest (RF), Boosted Regression Trees (BRT) and Generalised Linear Models (GLM) algorithms. Occurrence data from GBIF were combined with bioclimatic variables from CHELSA 2.1 under the current scenario (2011-2040) and future high emission scenarios (SSP3-7.0 and SSP5-8.5) for 2041–2070 and 2071–2100. The results indicated that RF and BRT generated more conservative projections, while MaxEnt and GLM presented broader distributions. Temperature seasonality was the main bioclimatic predictor. Attalea maripa, A. phalerata, and A. speciosa exhibited broader ranges of suitability, while A. funifera and A. vitrivir presented more restricted distributions. Future scenarios projected an increase in high suitability, particularly in the Amazon and Cerrado regions. Only A. funifera and A. vitrivir showed declines. By revealing how cartographic and predictive uncertainties limit territorial planning and ecosystem conservation, this study reinforces the importance of conservation biogeography as a strategic field for understanding and intervening in complex socio-ecological landscapes, especially in tropical regions subject to multiple environmental pressures.