Anthropogenic landscape change and amphibian diversity in tropical montane biodiversity hotspots: insights from satellite remote sensing in the Madagascar highlands

The magnitude of anthropogenic landscape change in tropical montane biodiversity hotspots and its relationship with biodiversity is a global issue that remains 'locked-in' in the broad narrative of tropical change in Africa. Over a montane biodiversity hotspot of Madagascar highlands (Ankaratra Massif), we conducted analysis on land cover change with Landsat satellite sensor data to identify the magnitude of change (1995-2016) and on the habitat change-amphibian diversity relationship to understand links with biodiversity. The results evidenced that 17.8% of the biodiversity hotspot experienced change in land cover in only 20 years. That pressured the already threatened forests, particularly since 2005. Of the total forest area in 1995 (2062.7 ha), 21.5% was cleared by 2016 (1618.3 ha). Changes in forest cover followed a bidirectional pattern. While in the period 1995-2005, forests expanded at a rate of 2.0% year-1 (from 2062.7 to 2524.8 ha), the area declined between 2005 and 2016 at a rate of - 4.1% year-1, fourfold the rate reported nationally for Madagascar (- 1.1% year-1). Forest-to-shrubland transitions emerged as being of increasing concern to forest integrity. We identified a significant link between habitat change and amphibian diversity, but only for species richness. Counter to expectations, no significant relationship was found between species richness and deforestation rates, and between microendemism rates and any of the habitat change variables. Species richness responded to the spatiotemporal variability in vegetation dynamics represented by the standard deviation of the Normalized Differenced Vegetation Index (NDVI_std). Species richness was strongly negatively related to NDVI_std in the short-term (R2 = 0.91, p = 0.003) and long-term (R2 = 0.69, p = 0.03), increasing where the spatiotemporal variability in NDVI was lower. The magnitude of changes in this biodiversity hotspot suggests that region-specific assessments are necessary in the context of the tropical change narrative in Africa and should consider conservation policies tailored for local conditions. Reducing deforestation and land conversion rates through a management plan codesigned with local communities is urgent. Habitat change appears to impact on amphibian diversity by altering the functional attributes of the habitat and not just by reducing habitat extent. NDVI_std seems a relevant indirect metric for monitoring such change although other biophysical attributes obtained from satellite sensor data should be integrated and explored.