Padrões de ocupação de herbívoros ao longo do tempo em relação à heterogeneidade da cobertura vegetal no complexo de Marromeu, Moçambique

Abstract

This study evaluated the occupancy patterns of large herbivores in relation to vegetation cover heterogeneity in the Marromeu Complex, Mozambique, by integrating remote sensing and occupancy modeling. The work was conducted in two complementary parts. In the first, the landscape physiognomies were mapped using 10 m resolution Sentinel-2 imagery classified with the maximum likelihood algorithm. Fourteen thematic classes were identified, revealing a mosaic landscape dominated by the Zambezi Delta Floodplain Grassland, the Cheringoma Coastal Humid Miombo, the Cheringoma Plateau Humid Miombo, the Zambezi Delta Lowland Forest, and the Inhamitanga Sandy Forest. The mapping achieved an overall accuracy of 93.53%. In the second part, data from the 2021 aerial census were organized into 349 sampling sites of 5 × 5 km, subdivided into two spatial replicates, for fitting single-season occupancy models applied to buffalo (Syncerus caffer), reedbuck (Redunca arundinum), and waterbuck (Kobus ellipsiprymnus). The models allowed separate estimation of occupancy probability (ψ) and detection probability (p), considering environmental and anthropogenic covariates such as physiognomy, NDVI, river extent, human population, and fire hotspots. The results indicated a strong association between species occurrence, hydrological components, and the habitat mosaic. Reedbuck showed ψ = 0.576 ± 0.041 and p = 0.560 ± 0.040, with higher occupancy in areas with greater river extent. Buffalo also showed a positive relationship between occupancy and rivers, along with increased detectability along this gradient. For waterbuck, detectability was explained mainly by river extent, whereas occupancy varied according to physiognomies and fire. It is concluded that vegetation cover heterogeneity, water availability, and landscape disturbances are key factors shaping the spatial distribution of large herbivores in the Marromeu Complex, and that the products generated provide a robust basis for wildlife monitoring, conservation, and spatial management in the Zambezi Delta.