Constructing Political Region Agglomerations for Effective Science Communcation

Abstract:

Often political regions form a better scale for modeling than regulargrids, since they are used for collecting socioeconomic data and aremore relevant to stakeholders and policy-makers. However, differentpolitical regions can enclose very different areas, populations,ranges of climatic variability, and for many countries there are anundesireable number of regions available at any administrative level.We provide a solution to this problem, by providing a general systemfor agglomerating regions to a larger scale, to approximately optimizearbitrary objectives. These regions provide an intermediate scale formodeling and greater comparability than unagglomerated regions. Wefirst apply the approach to the US, producing region agglomerationsthat normalize region size, population, income, and climaticvariability. We then apply the technique globally, generating acononical collection of regions for studying the impacts of climatechange.