This paper explores the interplay between the biophysical and economic geographies of climate change impacts on agriculture. Towards this end, we employ a statistical meta-analysis which encompasses all studies available to the IPCC-AR5 report. This permits us to isolate specific elements of the biophysical geography of climate impacts, such as the role of initial temperature, and differential patterns of warming across the globe. We combine these climate impact estimates with the GTAP model of global trade in order to estimate the national welfare changes which are decomposed into three components: the direct (biophysical impact) contribution to welfare, the terms of trade effect, and the allocative efficiency effect. We find that the terms of trade interact in a significant way with the biophysical geography of climate impacts. Specifically, when we remove the biophysical geography, the terms of trade impacts are greatly diminished. And when we allow the biophysical impacts to vary across the empirically-estimated uncertainty range, taken from the meta-analysis, we find that the welfare consequences are highly asymmetric, with much larger losses at the low end of the yield distribution than gains at the high end. Acknowledgement :