Quantifying the human cost of global warming

Despite enhanced pledges, current policies put the world on a path toward roughly 2.7 °C of warming by 2100, far from the Paris Agreement’s 1.5 °C ambition. Monetary valuation of climate damages raises ethical concerns, as it tends to overweight losses of the rich and the present over those of the poor and future generations. The study reframes climate damages by quantifying the number of people who will live outside the historical human climate niche, defined by the conserved distribution of population density with respect to mean annual temperature (MAT) with peaks near ~12–13 °C and ~27 °C. The research question is: How many people have already been, and will be, pushed outside this niche under different warming and socioeconomic trajectories, and how do policy choices alter these human costs and their inequities? The purpose is to provide an ethically grounded, people-centered metric of climate risk that informs climate justice and policy.

Prior research links climate variability and change to morbidity and mortality, reduced labor productivity, cognitive performance declines, impaired learning, adverse pregnancy outcomes, crop yield losses, conflict, hate speech, migration, and disease spread. The human climate niche concept integrates physiological and ecological constraints, showing conserved peaks in human population density with MAT around ~13 °C and ~27 °C; similar distributions apply to domesticated crops and livestock, and even gross domestic product. Mortality increases at both high and low temperatures, supporting the niche concept. Past analyses identified strong coupling between human populations and freshwater access, subsistence agriculture, and livestock, with warming extending pests and pathogens and prompting crop migration. The historical stability of the niche suggests limited potential for technology to expand it. Other proposed thermal metrics include mean maximum temperature (MMT) and wet-bulb temperature (WBT), both highly correlated with MAT.

Reassessing the niche: The authors reconstructed the temperature niche using the 1980 population distribution (4.4 billion; HYDE 3.2) against 1960–1990 climate as the reference, applying a double-Gaussian fit to the running mean of population density versus MAT (1 °C step, 2 °C bin). This yielded a primary peak at ~12 °C and a secondary peak at ~27 °C. They also derived a temperature–precipitation niche (MAT and mean annual precipitation, MAP), noting strong declines below ~1,000 mm yr−1 and acknowledging conservative bias when using temperature alone. Climate data sources for the reference and recent climates included WorldClim v1.4, CRU TS v4.05/4.06, NASA GLDAS-2.1, ERA5-Land, FLDAS, and NCEP CFSv2. Exposure definitions: Three exposure measures were calculated: (1) Hot exposure, defined as the share of people living at MAT ≥29 °C (a threshold experienced by only 0.3% or ~12 million people in 1960–1990). (2) Exposure outside the niche due to temperature change alone, computed as the spatial difference between an ‘ideal distribution’ (applying the fitted niche to the changed climate) and the ‘reference distribution’ under 1960–1990 climate. (3) Exposure due to climate plus demographic change, computed as the difference between the projected ‘assumed distribution’ (future population with future climate) and the ‘ideal distribution’. Thermal metrics linkage: The study verified that MAT is highly correlated with annual MMT and mean annual WBT, and related MAT to extremes: days with Tmax >40 °C rise markedly for MAT >27 °C, reaching >75 days yr−1 on average at MAT 29 °C; days with WBT >28 °C begin increasing at MAT >22 °C and exceed ~10 days yr−1 at MAT ≥29 °C. Changes to present: Using the 2010 population (HYDE 3.2; 6.9 billion) with 2000–2020 climate (1.0 °C above pre-industrial), the authors quantified hot exposure and niche displacement to date. Future projections: Climate projections used downscaled CMIP6 outputs (up to eight models: BCC-CSM2-MR, CNRM-CM6-1, CNRM-ESM2-1, CanESM5, GFDL-ESM4, IPSL-CM6A-LR, MIROC-ES2L, MRI-ESM2-0) from WorldClim v2.0 at ~10 km resolution across SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, for 20-year periods centered on 2030, 2050, 2070, and 2090. Spatially explicit population projections (1 km; consistent with SSPs) were aggregated to match climate resolution. Multi-model ensemble means and 5th–95th percentiles were computed for exposure metrics. Controlling for demography: To isolate climate policy effects, the study fixed spatial population distributions at 6.9, 9.5, and 11.1 billion and combined them with climate states corresponding to global warming levels of 1.5, 1.8, 2.0, 2.1, 2.4, 2.7, 3.6, and 4.4 °C (20-year means from appropriate SSP/time combinations). Linear relationships between warming and exposure metrics were derived. Country-level analyses: For 2.7 °C and 1.5 °C warming with 9.5 or 11.1 billion people, exposed populations were summed within country boundaries where MAT ≥29 °C using GIS country polygons. Emissions-exposure linkage: Using 2018 country-level production-based CO2-equivalent emissions (EDGAR) and country exposure at 2.7 °C warming (9.5 and 11.1 billion), the authors estimated the number of future people exposed per unit of present emissions, and related exposure to national poverty rates (World Bank).

• To date: Under observed 2000–2020 climate (1.0 °C warming), hot exposure (MAT ≥29 °C) tripled to 0.9 ± 0.4% (62 ± 26 million). Due to temperature change alone, 9 ± 1% are outside the niche; including demographic change, 10 ± 1%—a total of 624 ± 70 million people pushed into less favorable temperatures since 1960–1990, with demography adding ~77 million.
• Reference future (SSP2-4.5; ~2.7 °C in 2081–2100, peak ~9.5 billion): By 2030 at 1.5 °C, hot exposure reaches 4 ± 2% (0.3 ± 0.1 billion). By 2090 at 2.7 °C, hot exposure is 23 ± 9% (2.1 ± 0.8 billion); outside niche due to temperature only is 29 ± 5% (2.7 ± 0.5 billion); combined climate + demography is 40 ± 4% (3.7 ± 0.4 billion).
• Across SSPs (2090): Hot exposure ranges 8–40% (0.6–4.7 billion). Outside niche due to temperature only: 18–47% (1.3–4.7 billion). Combined climate + demography: 29–53% (2.2–6.5 billion). Considering the temperature–precipitation niche raises exposure by ~20% relative to temperature-only niche.
• Demography-controlled linearity: For fixed populations, relationships between global warming and exposure are near-linear. Hot exposure increases by ~11.9% per °C (6.9 billion) to ~17.5% per °C (11.1 billion). Temperature-only displacement increases ~11.8% per °C. Combined displacement scales ~9.1–11.0% per °C depending on fixed population.
• Worst cases: If policies yield high end warming, 3.6 °C could expose 34 ± 10% (3.3 ± 0.9 billion) to MAT ≥29 °C; 39 ± 7% (3.7 ± 0.7 billion) outside niche (temperature only); 48 ± 7% (4.5 ± 0.6 billion) combined. At 4.4 °C, these rise to 45 ± 7% (4.2 ± 0.7 billion), 47 ± 8% (4.5 ± 0.7 billion), and 55 ± 7% (5.3 ± 0.6 billion), respectively.
• Policy gains: Reducing end-century warming from ~2.7 °C (current policies) to 1.5 °C cuts hot exposure from ~22% to ~5% (2.1 to 0.4 billion), temperature-only displacement from 29% to 14% (2.8 to 1.3 billion), and combined displacement from 39% to 28% (3.7 to 2.7 billion). Each 0.3 °C reduction avoids ~4.3% (410 million) hot exposure, ~3.7% (350 million) temperature-only displacement, and ~2.8% (270 million) combined displacement (for a 9.5 billion world).
• Country-level: In a 9.5 billion world at 2.7 °C, India (>600 million) and Nigeria (>300 million) have the largest exposed populations; at 1.5 °C, these drop to ~90 million and <40 million, respectively. Indonesia’s exposure falls >20-fold (~100 million to <5 million). Several Sahelian countries retain high land-area exposure at 1.5 °C, with near-total land exposure in some countries at 2.7 °C.
• Emissions-to-exposure mapping: One future person is exposed to MAT ≥29 °C per ~460 (330–760) tC emitted. With present global mean per-capita emissions (~1.8 tC eq cap−1 yr−1; life expectancy ~72.6 years), ~3.5 average global citizens—or ~1.2 average US citizens—emit enough over a lifetime to expose one future person to unprecedented heat. Exposed future individuals tend to come from countries with current per-capita emissions ~56% of the global average.

Quantifying human cost via displacement from the historical climate niche provides an ethically transparent, people-centered metric that aligns with observed increases in heat extremes, urban heat exposure, and associated labor losses. The results indicate that even at current warming levels, hundreds of millions have shifted into less favorable thermal environments. Without stronger policies, up to one-third or more of humanity could be outside the niche by late century under midrange scenarios, and roughly half under worst cases, with substantial exposure to potentially lethal heat extremes where MAT ≥29 °C often implies many days per year with Tmax >40 °C and WBT >28 °C. The burden falls disproportionately on countries with higher poverty rates and lower per-capita emissions, highlighting inequity and underscoring the need for climate justice and loss-and-damage frameworks. The near-linear scaling between warming and exposure clarifies the benefits of incremental policy strengthening: every fraction of a degree avoided prevents exposure for hundreds of millions. Although colder regions may become more suitable, demographic growth is concentrated in hotter regions, increasing absolute and proportional exposure. The findings thus directly inform mitigation prioritization and adaptation planning, especially for highly exposed countries (for example, India, Nigeria, Sahelian nations).

This study reframes climate damages by quantifying how many people are and will be pushed outside the historical human climate niche as global temperatures rise. It shows that current policies (~2.7 °C) could place about one-third of humanity outside the niche by late century, while meeting the Paris 1.5 °C target reduces exposure roughly fivefold for unprecedented heat. The linear dose–response between warming and exposure provides a clear metric of the benefits of stronger mitigation. The emissions-to-exposure linkage makes inequities explicit: high emitters’ actions today disproportionately expose lower-emitting, often poorer populations in the future. Future work should integrate additional hazards (sea-level rise, compound extremes), refine niche characterization with humidity and precipitation (and their extremes), incorporate climate-induced migration and adaptation dynamics, explore urban heat island interactions, and assess socioeconomic pathways including finance constraints and adaptation investments.

Exposure estimates combining temperature and demography are upper bounds due to methodological constraints that limit the absolute density at the higher-temperature peak relative to the lower-temperature peak. The analysis assumes the temperature niche remains constant over time and does not explicitly model future technological, infrastructural, or behavioral adaptations that might expand tolerable conditions. MAT is used as a proxy; while strongly correlated with MMT and WBT, it does not capture all aspects of heat extremes. The temperature-only niche overestimates population density at very low and very high precipitation; projections using the temperature–precipitation niche yield higher exposure. Climate-induced migration is excluded from population scenarios. The study also does not quantify other climate hazards (for example, sea-level rise, permafrost thaw, or compound extremes), which would increase total risk. Uncertainty arises from climate model spreads, downscaling, and socioeconomic projections.