A neural network architecture for disaggregating age-specific population projections to the sub-national level

Tamburini, A., Bosco, C., & Striessnig, S. (2025). A neural network architecture for disaggregating age-specific population projections to the sub-national level. IIASA Working Paper. Laxenburg, Austria: WP-25-003

Preview

Text WP-25-003.pdf - Published Version Available under License Creative Commons Attribution Non-commercial. Download (2MB) | Preview

Abstract

Improving our understanding of future risk from climate change requires realistic projections of future
populations, both in their size and distribution. Distribution refers not only to geographic breakdowns but also to the breakdown by important characteristics, such as age. While the location where people will live may determine future exposure to hazards, population characteristics also co-determine the degree of vulnerability and the capacity to adapt to changing environmental conditions. Despite the importance of these factors, there remains a paucity of population projections (or disaggregations thereof) at the sub-national level. We develop a machine learning-based model to disaggregate age-specific population projections based on the Shared Socioeconomic Pathways (SSPs) to the sub-national NUTS-2 level for 34 European countries. Our focus on Europe is driven by its high degree of spatial variability, both in terms of climatic conditions and population structure, as well as the rapid pace of climate change and population aging there.

Actions (login required)

View Item