Forecasting Lives Lost to Climate Change: A Precautionary Empirical Framework for Climate-Related Mortality

Nigel Peter Howard , Peter Newman

Climate-related mortality remains underdeveloped in climate-risk assessment, despite extensive evidence that climate change affects food security, extreme events, conflict risk, and several health outcomes. This study develops a simplified, precautionary empirical framework to estimate climate-related mortality across selected pathways using historical relationships among population, greenhouse-gas emissions, atmospheric CO2 concentration, and global mean surface temperature, and introduces a cumulative mortality metric expressed as climate-related deaths per cumulative Mt CO2-e over 1900--2300. Publicly available datasets were assembled for undernourishment, conflict, disasters, selected health categories, population, emissions, atmospheric CO2 concentration, and temperature, and mortality was estimated under four stylised emissions pathways: business-as-usual, net zero by 2070, 2060, and 2050 (Paris aligned). Undernourishment and associated conflict emerged as the dominant components of the estimated burden, which also included storm, flood, landslide, and wildfire disaster categories. Air pollution was included although this is arguably an associated rather than climate-implicated cause of mortality. Other pathways were reviewed but excluded from the quantitative totals where globally consistent relationships with climate indicators were presently too weak or conflicting. Across scenarios, business-as-usual emissions are associated with substantially larger climate-related mortality burdens, with harms concentrated disproportionately in lower-income populations. Regional economic capacity, the framework's primary proxy for resilience and adaptive capacity, is the main driver of this unevenness; institutional capacity and population awareness were not modelled separately and instead contribute to the width of the reported uncertainty ranges. In the central estimates, cumulative climate-related deaths over 1900–2300 are roughly 5.8 billion under business-as-usual, compared with around 1.6–2.1 billion under net-zero pathways. Earlier progress toward net zero materially reduces projected mortality within the modelled ranges. Climate risk assessment can be improved by use of the 710 deaths per Mt CO2-e metric (assuming zero emissions by 2050) in projects that are causing emissions as well as those that save lives by reducing or drawing down emissions. Uncertainty remains substantial, particularly for undernourishment and conflict and for excluded health pathways, so the deaths-per-megatonne estimates are intended as indicative, approximate, and explicitly precautionary rather than as precise forecasts of climate-related mortality. Excluded potential causes and the prospect of accelerating climate change due to compounding feedback loops suggest that these estimates may still be underestimates of true climate mortality.