Abstract of the talk// Drylands, covering 40% of Earth’s land, are fragile ecosystems where water scarcity drives the formation of regular vegetation patterns extending over hundreds of kilometers. Because manipulative experiments are hard to perform at these large scales and still datasets tracking pattern dynamics are lacking, most of what we know about vegetation self-organization comes from theoretical models based on Turing instabilities. These models suggest that self-organized patterns are key for drylands, making them more resistant to environmental change. In this talk, Ricardo will first show that patterns could, instead, be a signature of ecosystem fragility under more realistic modeling assumptions, such as spatial heterogeneities leading to non-reciprocal plant interactions. Motivated by this finding, he will discuss the recent progress in developing a multiscale framework to study dryland resilience, using a combination of individual-level models and coarse-grained theories validated with greenhouse experiments and remote-sensing data.