Predicting Wildlife Refugia in Hyperfragmented Tropical Forests Using Microclimate Mosaics and Fine-Scale Movement Patterns

Abstract

The fragmentation of habitats in tropical forests alters microclimatic conditions and connectivity for movement, making it very challenging for wildlife. The key to successful conservation planning is thus in identifying climate-buffered refugia in hyperfragmented landscapes. This paper forecasts wildlife havens by applying coarse-scale animal movement patterns and high-resolution microclimate mosaics on a hyper-fragmented tropical forest. The data included in situ microclimate sensor data (temperature and relative humidity), remotely sensed canopy structural data, and GPS telemetry from several forest-related vertebrate species. Refugia detection was performed using a machine-learning framework that linked microclimate stability indices to step-selection and movement-resistance models. Areas with low thermal variance, maintained humidity, and seasonal, multi-species use were described as refugial patches. Findings show that refugia covered 12.4% of total forest cover, yet supported 46.7% of all locations of all animals recorded. The average daytime temperatures in refugia were 2.8 °C colder (p < 0.001) than those in non-refugial patches, and the persistence of relative humidity across seasons was 19.3% greater. The movement analyses showed that the likelihood of animals choosing microclimate-stable corridors rose 1.9-fold (0.64, SD = 0.11), although the travel distance in microclimate-stable corridors also increased. Those who inhabited refugia had 31% lower thermal exposure variance and a higher 24% site fidelity than those who used more open patches. These results indicate that heterogeneity in microclimates, rather than patch size, is the sole determinant of refugia formation in hyperfragmented tropical forests. Combining microclimate mosaics and animal movement behaviour provides a robust mechanism for delineating functionally significant refugia and ranking conservation measures to achieve climate resilience in swiftly shifting tropical landscapes.