Accurate land-cover mapping in tropical regions remains challenging because of high environmental heterogeneity, complex vegetation structure, and persistent cloud cover. Recent geospatial foundation models provide pre-computed Geo-embeddings that offer a new representation of Earth Observation data (EO), yet their potential for detailed tropical land-cover mapping and their performance relative to, and in combination with, conventional Sentinel-1 and Sentinel-2 satellite image time series remains largely unexplored. This study evaluated AlphaEarth Foundation Geo-embeddings for detailed vegetation mapping in northern Madagascar and investigated whether combining them with satellite image time series processed using the Satellite Image Time Series (SITS) R package could improve classification performance. Geo-embeddings were classified using a Random Forest model (GEO), while four supervised classifiers were trained on Sentinel-1 and Sentinel-2 time series. GEO achieved the highest standalone performance (OA = 0.74), outperforming all classifiers trained on the satellite image time series. Probability-level ensemble models were then used to assess whether both data representations could be beneficially combined. The best-performing ensemble, combining GEO (75%) with Temporal Convolutional Neural Network (TempCNN; 25%), increased overall accuracy from 0.74 to 0.82. These results suggest that Geo-embeddings provide an effective standalone representation for detailed tropical land-cover mapping and indicate that combining them with conventional satellite image time series may exploit complementary information. As geospatial foundation models continue to evolve, understanding how they can be integrated with established Earth observation workflows may support future operational land-cover mapping.

Biodiversity is declining globally principally because of land degradation and more because of climate change. Its effective conservation is vital for species and habitats, but also to maintain the related ecosystem services they provide for human well-being. In this context, evaluating the ability of Protected Areas (PAs) to cover species distribution under current and future environmental conditions is highly valuable. Considering the distributions of 1692 species of plants in the cross-border region of Grand Genève, located between France and Switzerland, the effectiveness of existing PAs in preserving plant diversity through local hotspots and priority areas for rare and vulnerable species was evaluated. The results show that PAs are moderately effective in conserving plant diversity, but are not expected to lose effectiveness in future conditions because important areas for plant diversity conservation will remain at similar locations. To address this gap, a spatial conservation network combining hotspots and priority areas was identified to cover 30% of the study area. It captures a significantly higher proportion of species distributions under both current and future conditions, and covers a greater representation of rare and ecologically important habitats, such as subalpine meadows and wetlands. The proposed solution aims to inform local stakeholders about areas of high ecological value that could be used to identify the Blue-Green Infrastructure, supporting the expansion of PAs and the improvement of conservation strategies in the face of environmental change.