Climate change impacts are emerging as key drivers of forced migration, particularly among hundreds of millions of agrarian households in the developing world. There are signs that temporary male labour migration appears to offer an alternative way out for “trapped” agrarian populations, but little is otherwise known about this escalating trend. Forecasting climate migration is crucial to prepare for its major societal impacts, but existing approaches are coarse and datasets on labour migration almost non-existent. The novelty of this project is that I address these dire data gaps, and will forecast climate migration at very fine spatial scales. India is used as a case study. I combine advanced satellite-based weather observations with a systematic tracking of agrarian workforces across 250,000 settlements using village-level microdata. I investigate spatial correlations between historical climate change and movement out of farming. This quantitative macro-analysis “from above” is combined with primary data collection on labour migration “from below”, creating a unique, custom-made dataset that can be used for fine-scale climate migration modelling. This approach can serve as a new paradigm in this field of studies, and results can inform policy-makers on the ‘when’, ‘where’, and scale of future climate migration flows under different climate scenarios.