We use climate ensembles with crop models, from seasonal to multi-decadal timescales, including perturbed-parameter and multi-model crop yield ensembles. The work shows that treatment of uncertainty can enable, rather than prevent, high confidence in scientific statements. We have also applied successfully a “time of emergency” approach to determine the timing of expected impacts. Coupled with analysis of current agricultural practices, this approach can directly inform adaptation (Challinor et al., 2016).
Other expanding areas of interest include i. the international dimensions of climate change– namely the transmission of risk across international borders and across sectors (including beyond food systems). ii. combining different perspectives and data on adaptation options, including household surveys and crop models.