Country-level energy demand for cooling using CMIP6 and world population projections

Cooling Degree Days (CDD) are commonly used to quantify energy demand for cooling and recent works highlighted the importance of population weighting to better represent energy load distribution. This study builds on the work of Scoccimarro et al. (2023), who assessed country-level cooling demand from 2000 to 2020 using both standard dry CDDs and humid CDDs (CDDhum), corrected with population weighting (CDD values are averaged at the national level, weighted by population). The humidity correction uses perceived temperature, which combines both temperature and humidity effects, rather than relying on temperature only. This adjustment offers a more accurate representation of cooling needs, as humidity plays a significant role in human stress and the demand for cooling.

This study aims to assess future cooling demand by utilising a selection of CMIP6 global climate models (GCMs), combined with country-level population projections from the United Nations World Population Prospects 2024. We analyse future trends (2015–2100) for the two mentioned metrics—standard cooling degree days (CDD) and humidity-adjusted cooling degree days (CDDhum) — both weighted by country-level population projections. Temporal evolution of these two metrics is assessed according SSP1-2.6 and SSP5-8.5 societal/emission scenarios, applying a consistent population weighting for both. GCM biases affecting population-weighted CDD and CDDhum are also assessed by considering ERA5 as reference product.

Preliminary results —calculated over Europe during the reference period 1971-2000 and without the application of humidity correction or population weighting — show that, despite some biases in the trend magnitude, the CMIP6 GCMs generally capture the spatial pattern of ERA5 CDD showing a general increasing trend in the energy required for cooling buildings during summer season. In particular, the Mediterranean Basin is projected to experience the most significant increase in CDDs, with considerable inter-model variability. In contrast, some northern European regions, such as the Scandinavian Peninsula and Iceland, show no trend in CDDs.

This work is based on ERA5 and CMIP6 data, collected and tailored as part of the H2020 BlueAdapt project (Grant agreement action Number 101057764), and on analysis codes developed under the Copernicus-funded contract (C3S2_520).