Delimitation of ecological corridors between conservation units in the Brazilian Cerrado using a GIS and AHP approach

[Display omitted] •Land use, slope terrain, permanent preservation areas are essential to set ecological corridors.•AHP, LCP, and landscape metrics offer methodological support to the study.•AHP, based on mathematical weights, enhance empiric human perception in ecological functions.•The method to define ecological corridors contributes to policies of ecosystems protection.•This study is a scientific advance to connect habitats and can adapt to other ecoregions. The Brazilian Cerrado is a biodiversity hotspot of global importance; however, it is fragmented. The objective of this study is to identify suitable areas for ecological corridors (ECs) between two conservation units (CUs) located in the Espinhaço Range Biosphere Reserve (ERBR) in the Cerrado biome, Brazil, based on analytic hierarchy process (AHP), least-cost path (LCP) methods and landscape metrics. The methodology was based on the use of the AHP as a decision-making tool. Using the AHP, relative weights were attributed to the criteria used in the construction of land use, occupation, terrain slope and permanent preservation area (PPA) maps, which were then used for the delimitation of ECs in a GIS environment based on the LCP approach and landscape metrics. The results included a proposal of three ECs, listed as A, B and C, with areas of 34.28, 27.04 and 28.80 km2, respectively. The use of the AHP minimizes the subjectivity of the criteria used. The analyses identified that the ECs had similar habitat quality to that of the conservation units. EC-A presented the largest area of natural vegetation. EC-B presents comparable vegetation to that of EC-A and the shortest connection. EC-C contains the largest area of monoculture and is the least suitable option, given the choice of ECs. Additional studies using other spatial algorithms should be used to find paths to optimize distances and costs for establishing ECs in Brazil and/or other world ecoregions.