Diminished resilience toward bioclimatic limits of the world’s forests revealed by Earth observation

Accelerating climate change and biodiversity loss are driving rapid ecosystem transformations, making it increasingly important to track resilience – the capacity of ecosystems to resist and recover from disturbance. Traditional monitoring approaches often focus on gradual changes in biodiversity and ecosystem properties but overlook the dynamic processes that provide early warning signals of resilience loss and critical transitions (e.g., desertification). Advances in Earth observation (EO) now enable spatially comprehensive and temporally consistent assessments of ecosystem resilience. To effectively integrate EO data into monitoring frameworks, we need to reconcile existing approaches to measuring resilience, understand which aspects of resilience different EO-derived indicators capture, and evaluate how these indicators relate within and across ecosystem types. Here, we synthesize multiple EO-derived indicators that capture complementary aspects of resilience – resistance and recovery. Our findings highlight the multidimensional nature of resilience and show how the relationship between these indicators varies across biomes, underscoring differences in underlying ecological processes. Together, these metrics provide a more robust framework to predict emergent patterns in resilience across large spatial gradients. Using this approach, we find distinct differences in resilience across the bioclimatic envelopes of the world’s functional forest biomes. Notably, there are signs of lower resilience at the hot and dry edges of sparse woodlands in warm temperate, Mediterranean and dry tropical regions. In contrast, cold temperate and tropical moist forests generally show reduced resilience at both the hot and cold edges of their bioclimatic envelope. Supporting classic metabolic theory and emergent patterns from research on boreal vegetation trends, we find signals of increased resilience on the intermediate-to-cold edges of the boreal forests. These findings demonstrate the value of integrating multiple EO-based resilience indicators to capture the complexity of resilience and underscore the need for biome-specific interpretation. Our framework advances our understanding of global resilience patterns, complements biodiversity monitoring, and supports evidence-based strategies to anticipate and mitigate the risks of abrupt ecosystem change.