The challenges of ending hunger and adapting to the impacts of climate change are high on the global development agenda and they are highly interdependent. Managing these challenges requires understanding of physical, social, political and environmental systems, their emergent properties and their interactions across scales - from individual perspectives to global processes. The Central American Dry Corridor (CADC) is a climatologically and politically defined region running through Central America, which is considered one of the most vulnerable regions in the world to the impacts of climate change. This research presented an attempt to apply cross-disciplinary approaches to the issue of food insecurity and climate impacts in the CADC, focusing in Guatemala. Applied methods range from ethnographically informed interviews and household case studies, to quantitative analyses with regional or globally standardized indices and data including household surveys, climate observation and climate impact modelling. A complex picture of the climate-food system is constructed through this integration of methods. For example, statistical analyses at the regional or national level evidenced: the risks of severe food insecurity during household transitions towards market participation; the association between agricultural labour income and severe food insecurity experiences; the association between the severity of the mid summer drought and prevalence of stunting; the association between changes in food insecurity classification used for early warning systems and drought indices (including the duration and magnitude of the mid summer drought); and the role assets and livestock (and the capabilities they represent) play in mediating climate impact and adaptation. Of the included variables across analyses, climatic and non-climate drivers differed in their relative importance regionally within Guatemala. However, this is not a complete picture, as some key findings from the individual and household level interviews were unable to be represented and tested in analyses at this scale. The main barrier to integrating participants narratives and knowledges into protocol for broad scale assessment, monitoring and decision-making - in this research - was the way in which top-down understanding and problem framings still shape what data is systematically collected and available. Recommendations identify how transdisciplinary approaches that engage iteratively with stakeholders at all levels of design, data collection, and analysis can be applied to contribute to overcome this limitation.