A woman carries water during a blistering hot day in New Delhi, India.

What Would Cities Look Like With 3 Degrees C of Warming vs. 1.5? Far More Hazardous and Vastly Unequal

The world recently experienced a 13-month streak of record-breaking global temperatures. And as blistering heat waves punish communities across several continents, 2024 is on track to be the hottest year on record. Global average temperatures are now perilously close to exceeding 1.5 degrees C (2.7 degrees F) above pre-industrial levels, a threshold scientists warn will bring increasingly dangerous droughts, wildfires and other impacts of climate change. Researchers project nearly 3 degrees C (5.4 degrees F) of temperature rise by 2100 without significant action to reduce greenhouse gas emissions.

That means almost 600 million people will be exposed to flooding from rising seas, food production will drop by as much as half and  habitats will suffer disastrous levels of loss. But what will 3 degrees C of warming vs. 1.5 degrees C look like in specific places — like Bengaluru, Johannesburg, Rio de Janeiro or the city you live in?

New data from WRI  finds that in most cities, the difference between 3 degrees C and 1.5 degrees C of warming is sizable. We analyzed potential climate hazards for nearly 1,000 of the world’s largest cities1 — currently home to 2.1 billion people, or 26% of the global population — using estimates based on downscaled global climate models. At 3 degrees C of warming, many cities could face month-long heat waves, skyrocketing energy demand for air conditioning, as well as a shifting risk for insect-borne diseases — sometimes simultaneously. People in low-income cities are likely to be the hardest hit.

These findings hold immense consequences for people’s lives and livelihoods, as well as for cities’ economies, infrastructure and public health systems. The implications are especially important as cities are home to 4.4 billion people globally — more than half the world’s population — and will grow rapidly over the next two decades. By 2050, as another 2.5 billion people move to urban areas, two-thirds of humanity will live in cities, with over 90% of that growth in Africa and Asia.

 

About this data

Global climate models often don’t produce data granular enough to be usable at a local level. For this article, we produced a global data set that includes data on 14 heat- and precipitation-related climate hazards for the 996 cities with populations greater than 500,000 people using a new statistical modeling method (designed by WRI) that makes it easier to project city-scale impacts from global climate models.

[Read more]

Global climate models often don’t produce data granular enough to be usable at a local level. For this article, we produced a global data set that includes data on 14 heat- and precipitation-related climate hazards for the 996 cities with populations greater than 500,000 people using a new statistical modeling method (designed by WRI) that makes it easier to project city-scale impacts from global climate models.

In consultation with city-level decision-makers from cities we work with, we chose these hazards for their relevance to public health, livelihoods, infrastructure and economic productivity. Average hazard magnitudes were calculated based on three climate models for each city (chosen based on best fit with historical data) for three global warming scenarios (1.5 degrees C, 2 degrees C and 3 degrees C), plus a recent historical reference period (1995-2014). For the sake of brevity, in this article we describe findings for the 1.5 and 3 degree C scenarios from only some of the hazard indicators, and we only report estimates from the best of the three models per city (based on historical correlation). Data from all three models are in the accompanying dataset.

Our data are all based on models. And like all models, there are limitations and unavoidable uncertainty. Here are some of the important limitations to our data:

Inherent uncertainty in climate models. Climate models are built on the research community’s most current understanding of how the climate works and will evolve, but they are far from perfect. The complexity of Earth’s systems and our inability to predict important factors like future greenhouse gas emissions mean that all climate models contain unavoidable uncertainty. The data we provide are estimates based on probability models, not definitive predictions. The full dataset includes estimates and uncertainty measures for three independent climate models used for each indicator, for each city.

Spatial resolution. Our data and the models we use have the same spatial resolution: approximately 25 km by 25 km pixels. While an improvement over earlier work, the resolution is still coarser than is ideal for urban climate planning. The estimates we provide are averages over areas more than 600 square kilometers. Therefore, our data likely underestimate extremes that might arise from more granular built environment differences: Very hot locations (locations that are warmer than average within their grid cells) will appear cooler, and cool locations will appear warmer.

Urban heat island effect. Because of pavement, building materials and relative lack of vegetation, cities tend to be warmer than rural areas. This is the urban heat island effect, and although we do include a bias correction step to partially mitigate the problem, the climate models do not account for it. Many locations in cities are therefore likely to be even warmer than is reflected in our data. Estimates of future heat wave frequency and duration and number of extreme-temperature days are likely to be low.

Simplified disease-risk hazards. We based our peak arbovirus and malaria transmission hazards on comparing estimated temperatures with published optimal temperature ranges for particular mosquito species (Aedes aegypti and Aedes albopictus for arboviruses, and Anopheles gambiae for malaria). Disease life cycles are of course much more complex than can be captured in simple temperature ranges, and diseases can be transmitted even in suboptimal temperatures and through vectors other than the species we considered. For example, Anopheles stephensi, an Asian mosquito and malaria vector, is a growing threat in Africa, particularly in urban areas. Finally, we note that our estimates examine climate only, and not the present or future geographic distribution of mosquitos or pathogens.

We are constantly refining our methods, and we expect to reduce uncertainty and improve our estimates as we gain access to better climate models. For now, our advice for interpreting these data is to understand the inherent uncertainty in this modeling work and consider our findings as a possible future, based on some of the best existing global models, as uncertain as they may be.

 

 

A young girl carries water through an informal settlement in Johannesburg, South Africa.
A young girl carries water through an informal settlement in Johannesburg, South Africa. Photo by Joe Eldrdige/Alamy Stock Photo

Here, we unpack what conditions may look like in the world’s largest cities if global warming reaches 3 degrees C, compared to 1.5 degrees C:

Longer and More Frequent Heatwaves

Study after study show that people die and economies slow down during prolonged periods of abnormally high temperatures. Even when temperatures do not approach life-threatening highs, heat waves can elevate risks to agriculture and infrastructure, such as crop failures and increased energy demand. At 3 degrees C of warming, most cities can expect both longer and more frequent heat waves than at 1.5 degrees C.2

Across the world’s largest cities, the longest heat wave per year in a 1.5 degrees C warmer world could average 16.3 days, with 3% of cities facing a heat wave every year that lasts one month or longer.

However, under the world’s current trajectory of nearly 3 degrees C of warming, the average duration of the longest heat wave in a year may jump to 24.5 days, with more than 16% of the world’s largest cities — home to 302 million people — exposed to at least one heat wave lasting one month or longer every year.3

Duration of annual longest heat wave also varies significantly by region. 

At 1.5 degrees C of warming, the Middle East and North Africa and Latin America and the Caribbean regions may see the longest heat waves on average, lasting an estimated 23-25 days.

At 3 degrees C of warming, the Middle East and North Africa may see the largest estimated jump in duration of annual longest heat wave (+13.6 days), bringing the longest heat waves in the region to 36.3 days, on average.

East Asia and the Pacific and South Asia regions may also see increases of 9-10 days longer, resulting in heat waves lasting 23-25 days.

Heat waves may become more frequent, too.

At 1.5 degrees C of temperature rise, the average city may experience 4.9 heat waves per year.

At 3 degrees C of warming, the number of heat waves may rise to 6.4 per year, with an increasing number of cities facing double-digit heat waves every year.

Cities in countries with different income levels would see different changes.

At 1.5 degrees C of global warming, cities in high-income countries may see an estimated 5.2 heat waves per year, on average, while those in low-income countries may see 4.7 heat waves.

At 3 degrees C of warming, the number of heat waves for cities in high-income countries may rise by 0.7 to an estimated 5.9 heat waves per year.

For cities in low-income countries, heat wave frequency may rise by an estimated 2.1 to 6.9 heat waves per year.

Longer, more frequent heat waves can take a massive toll on public health and social and economic development. The impacts of extreme heat can be life-threatening, especially on vulnerable groups like pregnant women, the elderly or children. And within cities, poorer neighborhoods will feel the effects more sharply, as they often have fewer green areas to cool the air, poorer quality buildings and less access to mechanical cooling.

Combined with air pollution, an all-too-common problem in cities, heat waves become even deadlier. For example, cardiovascular diseases are exacerbated by extreme temperatures and by tropospheric ozone, the formation of which increases at higher temperatures. Heat waves also exacerbate wildfires, which can increase exposure to PM2.5 pollution, a known carcinogen linked to diseases and premature death.

And beyond the impacts on human health, extreme heat can strain water, energy and transport systems, disrupting food chains and agriculture.

City Solution: Inclusive Resilience in Bengaluru, India

Bangalore.
Photo by shylendrahood/iStock

In Bengaluru, India, where the longest heat wave every year could average 13.5 days at 1.5 degrees C of warming and 37.8 days at 3 degrees C, extreme heat is already bringing severe water shortages and surging energy demand. The city is India’s fifth most populous, and 16% of its rapidly growing population live in informal settlements.

Last year, Bengaluru launched its first climate action plan, which can serve as a model for how large cities with high inequality can build resilience. The plan is inclusive by design, actively recruiting all sectors of society to participate in planning and implementing nature-based climate change solutions, like planting trees and expanding green space.

Bengaluru considers clean air and other components of human well-being in its climate plan. The plan is also broad in reach, targeting sectors like solid waste management, stormwater, greening, energy, transportation and biodiversity.

 

A man repairing air conditioners.
A man repairs air conditioners at an apartment building in Asuncion, Paraguay. As temperatures rise in the world’s cities, energy demands for cooling are projected to skyrocket. Photo by Jose Carlos Alexandre/iStock

Rising Demand for Cooling

Longer, more frequent heat waves and extremely hot days will significantly raise the demand for cooling4 in homes and workplaces — and with it, demand for energy if those needs are met through mechanical means like air conditioning. Many cooling needs can be effectively met through passive cooling — including material selection, shade, insulation and ventilation — but these solutions are too often ignored in favor of mechanical methods perceived as more convenient.

At 1.5 degrees C of warming, approximately 8.7 million people across a handful of cities could face double their historical (1995-2014) cooling demand.5

At 3 degrees C of warming, approximately 194 million people may see a doubling from their historical levels.

Places that haven't had to consider cooling strategies because of mild climates — like northern Europe or the U.S. Pacific Northwest — will need to do so more.

If met through mechanical means, the uptick in energy use may require a sizable investment in energy infrastructure and building technologies, even if cooling demand is still relatively small. It would also make it harder to achieve decarbonization goals in regions where power relies on fossil fuels.

Meanwhile, many places that are already hot — like Tehran and Marrakech — are increasing their cooling demand faster than relatively cool cities.

For many of these cities, population growth and cooling technology adoption become complicating factors. Energy-intensive air conditioning is less commonly used in lower-income cities, but demand is increasing as these countries get richer and more populous. For example, the percentage of Indian households with air conditioning is expected to match or exceed that of Europe by 2050.

At the same time, without policies and concerted public investment in passive cooling, many poorer people in soon-to-be hotter cities won’t have access to cooling solutions. This could increase social inequity and health risks.

Only about 8% of the 2.8 billion people living in the hottest and often poorest parts of the world currently have air conditioning in their homes. For example, in India, over 215 million people in cities are already estimated to be at higher risk of heat-related health problems due to lack of cooling services and infrastructure, including more than 121 million women and girls, who are affected disproportionately by extreme heat both in health and in economic outcomes. There is a massive gap between existing cooling investment and what will be needed by 2050 to save lives, especially in lower-income countries.

City Solution: Mapping Hotspots in Johannesburg, South Africa

Johannesburg.
Photo by Rich T. Photo/Shutterstock

Heat waves in Johannesburg, South Africa, are becoming longer and more frequent. Johannesburg’s demand for cooling at 3 degrees C of warming is estimated to be 69% greater than at 1.5 degrees C. Air conditioning doesn’t work well when high temperatures cause water shortages and power outages, as they did in a recent heat wave that affected most of South Africa. And for many of Johannesburg’s residents who live in informal settlements — approximately 1 in 5 people — water and electricity are unreliable year-round.

The often-deadly combination of high temperatures and high inequality in cities calls for aggressive action from governments and communities. In Johannesburg, this starts with mapping. In 2022, local government officials partnered with community volunteers to measure temperatures and create maps that show the intersection of heat and vulnerability. The effort provided decision-makers with high-quality data, built a compelling evidence base for specific cooling measures, and brought stakeholders throughout the city to the planning table. As the city proceeds with tree plantings, white roofs and other cooling initiatives, its efforts are backed by data collected by residents, improving coverage of the most vulnerable.

 

A man fumigates mosquitos in Rio de Janeiro, Brazil.
A man fumigates mosquitos in Rio de Janeiro, Brazil. Photo by Luiz Souzarj/iStock

Growing and Shifting Disease Risks

Warmer conditions often fuel mosquito-borne diseases. Geographic patterns and prevalence of illnesses6 would shift significantly in a 3 degrees C warmer world, requiring major changes in public health priorities and investments. (Our analyses are based on changing temperatures and not on other factors affecting disease lifecycles or on current or future prevalence of disease pathogens and vectors.)

Incidence of arboviruses such as dengue, Zika, West Nile, yellow fever and chikungunya will likely increase worldwide as days with optimal temperatures for disease-carrying mosquitos become more common.

Comparing 1.5 degrees C and 3 degrees C of warming, the average increase in peak arbovirus-transmission days globally is 6 days. But the picture is much more complicated than the global average.

Low-income cities will be particularly hard hit.

At 1.5 degrees C of warming, cities in low-income countries may experience 73.4 peak arbovirus days each year, compared to just 27.9 for cities in high-income countries.

And at 3 degrees C of warming, cities in low-income countries may experience 87.4 peak arbovirus days each year, compared to 32.1 days per year for the average city in high-income countries.

The incidence of malaria, on the other hand, is expected to decrease globally, as temperatures in many places become warmer than what is optimal for malaria-transmitting mosquitos.

An increase in global warming from 1.5 degrees C to 3 degrees C means fewer expected peak malaria days — dropping from 114.0 to 104.4 days on average — but it varies considerably by location. For example, at 3 degrees C, some cities in Europe, Central Asia and North America see the expected number of days with peak-transmission temperatures increase. Most of these cities do not currently harbor malaria-carrying mosquitos — but as occasional reports arise of locally acquired malaria in Europe and the U.S., there is increasing concern that malaria could creep into places that haven’t seen it in living memory.

City Solution: Hunting Dengue in Rio de Janeiro, Brazil

Rio de Janeiro.
Photo by Giongi63/Shutterstock

Arboviruses are of particularly urgent concern in Latin America. Brazil is already experiencing a dengue crisis, and at 3 degrees C of warming, 11 of its largest cities could see high arbovirus risk for at least six months of the year. Rio de Janeiro may see significantly more illness if global warming reaches 3 degrees C, as the expected number of peak arbovirus days increases by 71%, from 69 to 118 days per year.

As temperatures in Rio de Janeiro become more hospitable to dengue-carrying mosquitos, the city is investing in improving the availability of dengue vaccines and controlling the mosquitos themselves. Community health workers crisscross the city, hunting for places where standing water accumulates and mosquitos can breed. And Rio de Janeiro is using another disease — Wolbachia, a bacterium that infects insects — to inhibit dengue transmission in infected mosquitos. Rio de Janeiro and five other Brazilian cities release Wolbachia-infected mosquitos by the tens of thousands. These mosquitos behave no differently from uninfected mosquitos, but they harbor less dengue virus and infect other mosquitos. A similar Wolbachia program in Yogyakarta, Indonesia resulted in a 77% reduction of dengue incidence.

 

Slum in Mumbai.
Mumbai, India. Low-income cities, in particular, will suffer compounding risks from the interacting effects of above-average increases in temperatures and disease, potentially straining health care systems, infrastructure and budgets that are already insufficient to meet residents’ needs. Photo by Harald Spilker/Alamy Stock Photo

Low-income Cities Will Face Compounding Climate Risks

Climate hazards rarely come in isolation. Low-income cities, in particular, will suffer compounding risks from the interacting effects of above-average increases in temperatures and disease, potentially straining health care systems, infrastructure and budgets that are already insufficient to meet residents’ needs.

Sub-Saharan Africa is likely to be the hardest hit region for several indicators at 3 degrees C of global warming. The region’s cities may endure multiple severe climate impacts, potentially simultaneously. Comparing impacts at 1.5 degrees C versus 3 degrees C of warming, the region’s cities rank first on average for the greatest estimated increase in heat wave frequency (up 56%, to 6.5 occurrences per year) and peak arbovirus days (25 additional days, totaling 129 annually). They also see the third-largest estimated increase in the annual longest heat wave duration (up 58%, to 20 days) and the second-biggest estimated increase in energy demand for cooling (an additional 208 cooling degree days).

Although Africa is only responsible for 2%-3% of the world’s greenhouse gas emissions, many of its sub-Saharan cities are most affected by climate change-induced hazards. At 3 degrees C of warming, cities like Freetown, Sierra Leone and Dakar, Senegal could endure heat waves lasting longer than a month, with an average of seven heat waves in a year, demanding greater energy for cooling (13% and 20% above historical reference period averages, respectively). Many cities in the region are already dealing with rapid population growth, conflict and unregulated development. These compounding climate hazards only add to the region’s vulnerability.

Latin American cities rank second on average for the greatest estimated increase in heat wave frequency (up 48%, to 7.5 occurrences annually). The region’s cities have the second-highest estimated increase in the number of days optimal for arboviruses at 3 degrees C (13 additional days, to 91 days annually) compared to 1.5 degrees C.

Cities in Indonesia face some of the highest compounding risks of both hotter temperatures and greater disease. Three of the four cities in the world with the greatest estimated increase in peak arbovirus days from 1.5 degrees C to 3 degrees C of warming are in Indonesia: Yogyakarta (273 day increase), Jember (151 day increase) and Padang (148 day increase).

Street in Indonesia.
Yogyakarta, Indonesia. Cities in Indonesia face some of the highest compounding risks of both hotter temperatures and greater disease in a warmer world. Photo by Manfred Gottschalk/Alamy Stock Photo

Reaching 3 Degrees C Is Not a Foregone Conclusion

The future under 3 degrees C of warming looks dire, but these data are only an estimate of effects for one pathway. We have a choice about our climate future. The technology already exists; with the right political will, we can mitigate the worst of these effects.

National and local governments must work urgently to make their cities more climate-resilient while slashing emissions. Specific priorities include:

  • Double down on policies to reduce greenhouse gas emissions: Climate problems will only worsen until the world brings greenhouse gas emissions down to net zero. National governments must implement policies aligned with the international Paris Agreement on climate change; cities can set targets and adopt policies to contribute to meeting these goals.
  • Increase finance for climate adaptation: Many low-income cities have climate adaptation plans, but implementation lags due to lack of funding. Adaptation needs for low-income countries are $387 billion per year — 10-18 times bigger than current international public adaptation finance flows.
  • Invest in resilient infrastructure: Adaptation solutions like cooling infrastructure, early-warning systems and heat action plans protect people’s health and boost cities’ long-term resilience. Nature-based solutions like urban forests and wetlands also do these things while bringing the added benefits of nature to city residents.
  • Improve collaboration between cities and national governments, and across city departments: Multi-level partnerships and coordination are key to tackling greenhouse gas emissions and implementing resilience at the scale required to save and improve people’s lives. While in many parts of the world different levels of government tend to operate in silos, using separate strategies, some national governments see the value of cities and support them as climate-solution allies. More than 70 countries committed at the 2023 UN climate summit (COP28) to collaborate with cities on their next national climate plans, due in early 2025, through the CHAMP Initiative. Others should follow.
  • Take a data-driven approach: Climate-hazard forecasts for cities can help them focus on the most likely and urgent climate risks while informing targeted investments and policies. City-scale, city-relevant data like ours should be at the center of planning and budgeting for climate adaptation and resilience.
 
 

The data for all 14 climate hazards, calculated for 996 cities, are available here. For this article, unless noted otherwise, we used the data only from the single best-performing of nine climate models for each location (the model that best matches the historical meteorological record for each city in the time periods where they overlap), but the dataset includes data from the three best-performing models. The data also include standard deviations for our estimates, as well as estimates and standard deviations of probabilities that the hazard magnitude exceed certain thresholds. The methods underlying our estimates of average hazard magnitudes and threshold-exceedance probabilities are detailed in this technical note. The methods we used to model warming scenarios and our definitions of the 14 climate hazards are detailed in this technical note.

The dataset is formatted to be of most use to data analysts equipped with statistical-analysis software. For everyone else, we’ve put together this user-friendly explorer tool.

Download DataDownload Data

 


Endnotes

1 Our cities are based on the definitions, 2015 populations and spatial extents outlined by the GHS Urban Centre Database.

2 There is no universally applied definition of “heat wave.” For this project, we defined it as three or more consecutive days on which the high temperature equals or exceeds the local 90th percentile of daily high temperature, as determined from the ECMWF ERA5 dataset over the 40-year period 1980–2019.

3 Exposure estimates are based on the 2020 populations in the Global Human Settlement Layer’s GHS-POP dataset.

4 We looked at cooling degree days, the annual sum of positive daily average temperature deviations from a threshold temperature, and a commonly used indicator of energy demand for cooling. Specifically, we calculated CDD21, the annual sum of daily positive deviations from 21°C.

5 Exposure estimates are based on the 2020 populations in the Global Human Settlement Layer’s GHS-POP dataset.

6 We used the ranges 22.9 – 27.8°C and 26 – 29°C for malaria- and arbovirus-transmitting mosquitos, respectively. These are estimated to be optimal for mosquito activity, but transmission can occur well outside these ranges. Also, note that many locations have temperatures within the optimal ranges but do not have the mosquitos. Our estimates therefore may overestimate the risk of shifting disease prevalence.