Forest Fragmentation Geometry Drives Sylvatic Vector-Borne Disease Incidence in the Southern Brazilian Amazon: A Geospatial Spatial-Autoregressive Assessment

Abstract

Background: Deforestation of the Brazilian Amazon has restructured the region's epidemiological landscape, yet the role of forest fragmentation geometry—as distinct from aggregate forest loss—in driving sylvatic vector-borne disease remains poorly quantified.

Objective: To assess geospatial relationships among fragmentation, edge microclimate and the combined incidence of malaria and cutaneous leishmaniasis across 150 quadrants (10 km × 10 km) in the Arc of Deforestation frontier of the southern Brazilian Amazon (Rondônia–southern Pará; Figure 1).

Methods: A composite Forest Fragmentation Index (FFI) was derived from Landsat-8 imagery (edge density, patch size, core isolation); incidence was obtained from national surveillance and standardised per 1000 residents; temperature and precipitation were extracted from WorldClim. Data were analysed with ANOVA and Tukey HSD, Pearson correlation, principal component analysis (PCA) and a Queen-contiguity spatial autoregressive (SAR) model.

Results: Mean FFI was 0.65 ± 0.12 and mean incidence 15.4 ± 4.2 per 1000. Incidence differed across fragmentation tertiles (F = 86.09, p < 0.001, η² = 0.539; Tukey a/b/c), rising from 11.7 ± 3.1 to 19.2 ± 3.1. FFI correlated strongly with incidence (r = 0.804, p < 0.001). The SAR model (Moran's I = 0.408; ρ = 0.457, p < 0.001) confirmed FFI as the dominant driver; a 0.1-unit increase raised incidence by 1.9 cases per 1000 (p < 0.001), independent of climate, whereas precipitation was non-significant. PCA attributed 61.5% of variance to a landscape component.

Conclusion: Landscape geometry should be treated as a public-health variable in Amazonian conservation and surveillance policy, supporting fragmentation-guided surveillance and stricter legal-reserve enforcement along active deforestation frontiers.