Mapping and synthesizing evidence on global adaptation progress: Achievements of GAMI and lessons for the future

As climate change adaptation is unfolding across the world, a scientific evaluation of progress in adaptation is ever more important. Adaptation projects are increasingly reported and assessed in the scientific literature, covering an growing breadth of case studies from across different world regions and sectors such as agriculture, critical infrastructure or settlements. Learning from the systematic mapping and synthesis of adaptation evidence documented in such scientific studies was the main objective of the Global Adaptation Mapping Initiative (GAMI). Launched as a research community effort during the IPCC’s Sixth Assessment Cycle (AR6) it has by now generated an important global knowledge base for the evaluation of adaptation. Next to global overview publications (synthesizing close to 3000 articles), a considerable number of regional or sector-specific in-depth studies as well as thematic and methodological extensions to the original data set have been published by now. With a view towards AR7 it is therefore time to stake stock current achievements and debate requirements for the future.
The session aims to collect and discuss lessons learned within GAMI and reflect on the implications for future adaptation mapping and synthesis. Contributions with a methodological, conceptual, empirical or modeling focus are all welcome. Presentations might concentrate on different sectors and regions or address the overall global picture. We invite not only contributions from colleagues who have contributed to GAMI but explicitly also from users who have not been involved but see a need to discuss potential shortcomings and improvements. Questions of the session include but are not limited to: Which methodological challenges exist in the GAMI approach and how can they be overcome? How can GAMI results be validated and triangulated with other data sets? How can the coding methodology of GAMI be combined with other methods e.g. in the field of large language models?