Coverage-Methodology-Glossary

Coverage

The list of cities corresponds to the cities featured in the Urban Audit dataset of Eurostat at the time of extraction. The Urban Audit features 870 cities from 30 countries (EU-27, with the addition of Norway, Switzerland and the UK), corresponding to cities for which the Local Authority Unit (LAU) exceeds 50,000 inhabitants. In the analysis, we discarded overseas cities from France, Spain and Portugal, as well as Reykjavik (Iceland) given the very different climates compared to mainland Europe. This leaves a sample of 854 cities.

Results are reported for 5-year periods from 2015 to 2099 according to three Shared Socioeconomic Pathways (SSP) and Representative Concentration Pathways (RCP): SSP1-RCP2.6, SSP2-RCP4.5 and SSP3-RCP7.0. We also report results for three levels of global warming: 1.5°C, 2°C, 3°C above pre-industrial levels. These warming levels are reported in the context of scenario SSP3-RCP7.0, being the one of the three for which climate models reach each of these warming levels.

Methodology

Projected excess deaths due to climate change rely on exposure-response functions previously estimated for all cities described above (openly shared on Zenodo). These functions indicate, for a specific temperature, what is the expected increase in mortality.

To estimate future excess deaths we retrieved projections of daily temperature for all cities from 19 General Circulation Models from the NASA Earth Exchange Global Daily Downscaled Projections database, and projections of population and death rates for the 30 countries from the Wittgenstein Centre Human Capital Data Explorer. These projections were extracted for the period and scenarios described.

For each SSP-RCP scenario, we projected annual excess deaths using a standard attribution methodology, then computed rates from projected population figures. Two sub-scenarios were considered: one in which we consider changes of both temperature and demographics (Historical + Demographic or demo), and one in which the distribution of temperature is kept constant across the century (Full). The difference between these two scenarios (Climate change or full-demo) provides an estimation of heat and cold-related deaths that can be attributed to climate change.

Glossary

Impact measures

Attributable Number / Excess deaths: Represents the estimated annual number of deaths related to temperature in a given period.

Attributable fraction: Represents the annual proportion of deaths attributed to temperature in a given period.

Excess death rates: Represents the annual proportion of inhabitants that could die because of temperature in a given period. Expressed as a number of deaths attributed to temperature per 100,000 persons.

Scenarios

SSP: Shared Socioeconomic Pathway
RCP: Representative Concentration Pathway

SSP1-RCP2.6: In the SSP1 scenario, the world shifts towards a more sustainable path, reducing inequalities and investing more in health and education. Coupled with RCP 2.6, this represents a low greenhouse gases emission scenario that would reach net zero around 2075. In this scenario, warming would be expected to remain below 1.8°C. Population growth in Europe is higher in SSP1 than in SSP3 and baseline mortality rates are lower.

SSP2-RCP4.5: The SSP2 scenario represents a world in which the socioeconomic and technological trend remain consistent with the historical ones. This is coupled with RCP4.5, representing a moderate greenhouse gases emission scenario in which emissions start decreasing the second half of the century without ever reaching net zero. Under this scenario, warming would likely be above 2°C and could even exceed 3°C.

SSP3-RCP7.0: The SSP3 scenario assumes an increase in regional rivalries, a resurgence of nationalism and a low priority given to environmental concerns. Coupled with RCP7.0, this corresponds to a high greenhouse gas emission scenario in which emissions would double by the end of the century. Under this scenario, global warming would likely exceed 3.5°C at the end of the century.

Sub-scenarios

Please see the last paragraph of Methodology.