Ten years ago, , an international treaty designed to cut greenhouse gas emissions and curtail global warming. Under the treaty, most nations made a 15-year promise to reduce emissions. Now, armed with a decade of data, a new 91̽��-led study shows global progress, but not enough to compensate for the environmental cost of economic growth.

With its adoption in 2015, the Paris Agreement kicked off a concerted effort to reduce emissions by prioritizing renewable energy, reducing waste and tracking impact. The treaty aimed to keep the planet from warming more than 2 degrees Celsius by 2100, with a target of 1.5 C. A new statistical analysis of the data shows that the agreement has helped some nations cut emissions, but the net impact is still too high to curb warming.

“The efforts made in response to the Paris Agreement did change the course of things, but the effects were knocked out by an increase in gross domestic product,” said , a 91̽��professor emeritus of statistics and sociology who led the study.

The results were .

This study is the third in a series of papers by Raftery and colleagues. The first, published in 2017, provided a probability-based assessment of what would happen if every country satisfied its , as defined by the Paris Agreement. The researchers’ predictions were based on three components: total world population, GDP and carbon intensity, a measure of carbon emissions per dollar.

“One of the key findings from that work was that basically, it’s not going to be enough,” Raftery said. “Even if every country met their goal, there was just a 5% chance that we would stay below the 2-degree mark.”

Meeting the collective goal would require everyone to do more, but just how much more was unclear.

Four years later, in 2021, the authors came back to say that emissions goals needed to be about 80% more ambitious to keep warming below 2 C. If each country increased its emissions reduction goal by 1.8% annually, and continued doing so after the Paris Agreement lapses in 2030, temperatures might stay just under the threshold.

This new study updates the statistical methods developed in 2021 and applies them to data collected in the past decade. It reveals that carbon intensity is trending downward by 3.1% each year, compared to 1.1% before the Paris Agreement was signed in 2015.

“It’s an improvement,” Raftery said, but the net result is not positive because world economies grew faster than expected. Even though less carbon was released to produce each economic unit, global GDP increased enough to drive total emissions up.

“Reducing economic growth is unpopular,” Raftery said, and the world population is growing. “So realistically, carbon intensity is the only factor under some kind of policy control.”

Still, the data does contain promising trends. The chance of “the most catastrophic climate change,” where temperature increases by 3 or more degrees, has gone down since 2015, from 26% to 9%. The likelihood of keeping warming below 2 degrees has also increased from 5% in 2015 to 17%.

All countries can and should continue looking for ways to contribute but some have more power than others, according to Raftery. The bigger the economy, the greater its impact on carbon intensity.

China, which is responsible for nearly one-third of total global carbon emissions, saw dramatic economic growth over the past decade. Although the country managed to reduce its carbon intensity 36% by 2024, its emissions shot up as GDP rose.

Both India and Russia followed similar trends, with observed emissions climbing far above their projected goals.

“China and the U.S. have the biggest economies and are among the most wasteful countries,” Raftery said. The carbon intensity for China was three times that for Germany, which is the largest economy in the European Union. Carbon intensity for the U.S. was 50% higher than that for Germany.

When President Donald Trump took office earlier this year, from the Paris Agreement for a second time. After he withdrew during his first term, former President Joe Biden rejoined and pledged that the U.S. would reduce its emissions 60% from peak levels by 2035. To evaluate how the climate will change if the U.S. makes no contribution, the researchers ran the numbers again without it.

If all the other participating countries meet their goals, the projected temperature increase goes up by 0.1 C and the chance of staying below 2 C decreases from 34% to 27%. The researchers note that this is an optimistic projection. It assumes that the U.S. will stop making cuts, but does not account for the possibility that it will reverse the trend by increasing carbon intensity.

“The fact that the Paris Agreement did work in reducing at least carbon intensity is good news,” Raftery said. “There need to be bigger efforts made to offset economic growth, but there is reason for hope.”

Coauthors include and , both former 91̽��graduate students of statistics. This research was funded by the National Institutes of Health.

For more information, contact Raftery at raftery@uw.edu.

Record-breaking heat waves have occurred recently from Delhi to the Pacific Northwest, and the number of these deadly events is expected to increase. New research from the 91̽�� and Harvard University gives a range of heat impacts worldwide by the end of this century, depending on future emissions of greenhouse gases.

The was published Aug. 25 in the open-access journal Communications Earth & Environment.

“The record-breaking heat events of recent summers will become much more common in places like North America and Europe,” said lead author , who did the research as a doctoral student at the 91̽��and is now a postdoctoral researcher at Harvard. “For many places close to the equator, by 2100 more than half the year will be a challenge to work outside, even if we begin to curb emissions.”

“Our study shows a broad range of possible scenarios for 2100,” he added. “This shows that the emissions choices we make now still matter for creating a habitable future.”

The study looks at a combination of air temperature and humidity known as the “” that measures impact on the human body. A “dangerous” heat index is defined by the National Weather Service as 103 F (39.4 C). An “extremely dangerous” heat index is 124 F (51 C), deemed unsafe to humans for any amount of time.

“These standards were first created for people working indoors in places like boiler rooms — they were not thought of as conditions that would happen in outdoor, ambient environments. But we are seeing them now,” Vargas Zeppetello said.

The study finds that even if countries manage to meet the Paris Agreement goal of keeping warming to 2 C, crossing the “dangerous” threshold will be three to 10 times more common by 2100 in the U.S., Western Europe, China and Japan. In that same scenario, dangerous days could double by 2100 in the tropics, covering half the year.

In a worst-case scenario in which emissions remain unchecked until 2100, “extremely dangerous” conditions, in which humans should not be outdoors for any amount of time, could become common in countries closer to the equator — notably in India and sub-Saharan Africa.

“It’s extremely frightening to think what would happen if 30 to 40 days a year were exceeding the extremely dangerous threshold,” Vargas Zeppetello said. “These are frightening scenarios that we still have the capacity to prevent. This study shows you the abyss, but it also shows you that we have some agency to prevent these scenarios from happening.”

The study uses a probability-based method to calculate the range of future conditions. Instead of using the four future emissions pathways included in the Intergovernmental Panel on Climate Change reports, the authors use a statistical approach that combines historical data with population projections, economic growth and carbon intensity — the amount of carbon emitted for each dollar of economic activity — to predict the likely range of future CO2 concentrations.

The statistical approach “gives plausible ranges for carbon emissions and future temperature and has been estimated statistically from and validated against historical data,” said co-author , a 91̽��professor of statistics and of sociology with an adjunct appointment in atmospheric sciences.

The authors translated the higher carbon dioxide levels into a range of global temperature increases, then looked at how that would affect global monthly weather patterns.

“The number of days with dangerous levels of heat in the mid-latitudes — including the southeastern and central U.S. — will more than double by 2050,” said co-author , a professor of atmospheric sciences at the UW. “Even for the very low-end estimates of carbon emissions and climate response, by 2100 much of the tropics will experience ‘dangerous’ levels of heat stress for nearly half the year.”

The results underline the need to both reduce future greenhouse gas emissions and to protect populations, especially outdoor workers, against dangerous heat. The research was funded by the National Institutes of Health, the James S. McDonnell Foundation and the Tamaki Foundation.

For more information, contact Vargas Zeppetello at lzeppetello@fas.harvard.edu or Battisti at battisti@uw.edu.

Now, two 91̽�� scientists have developed a statistical framework that incorporates key COVID-19 data — such as case counts and deaths due to COVID-19 — to model the true prevalence of this disease in the United States and individual states. Their approach, July 26 in the Proceedings of the National Academy of Sciences, projects that in the U.S. as many as 60% of COVID-19 cases went undetected as of March 7, 2021, the last date for which the dataset they employed is available.

This framework could help officials determine the true burden of disease in their region — both diagnosed and undiagnosed — and direct resources accordingly, said the researchers.

“There are all sorts of different data sources we can draw on to understand the COVID-19 pandemic — the number of hospitalizations in a state, or the number of tests that come back positive. But each source of data has its own flaws that would give a biased picture of what’s really going on,” said senior author , a 91̽��professor of sociology and of statistics. “What we wanted to do is to develop a framework that corrects the flaws in multiple data sources and draws on their strengths to give us an idea of COVID-19’s prevalence in a region, a state or the country as a whole.”

Data sources can be biased in different ways. For example, one widely cited COVID-19 statistic is the proportion of test results in a region or state that come back positive. But since access to tests, and a willingness to be tested, vary by location, that figure alone cannot provide a clear picture of COVID-19’s prevalence, said Raftery.

Other statistical methods often try to correct the bias in one data source to model the true prevalence of disease in a region. For their approach, Raftery and lead author , a 91̽��doctoral student in statistics, incorporated three factors: the number of confirmed COVID-19 cases, the number of deaths due to COVID-19 and the number of COVID-19 tests administered each day as reported by the . In addition, they incorporated results from random COVID-19 testing of Indiana and Ohio residents as an “anchor” for their method.

The researchers used their framework to model COVID-19 prevalence in the U.S. and each of the states up through March 7, 2021. On that date, according to their framework, an estimated 19.7% of U.S. residents, or about 65 million people, had been infected. This indicates that the U.S. is unlikely to reach herd immunity without its ongoing vaccination campaign, Raftery and Irons said. In addition, the U.S. had an undercount factor of 2.3, the researchers found, which means that only about 1 in 2.3 COVID-19 cases were being confirmed through testing. Put another way, some 60% of cases were not counted at all.

This COVID-19 undercount rate also varied widely by state, and could have multiple causes, according to Irons.

“It can depend on the severity of the pandemic and the amount of testing in that state,” said Irons. “If you have a state with severe pandemic but limited testing, the undercount can be very high, and you’re missing the vast majority of infections that are occurring. Or, you could have a situation where testing is widespread and the pandemic is not as severe. There, the undercount rate would be lower.”

In addition, the undercount factor fluctuated by state or region as the pandemic progressed due to differences in access to medical care among regions, changes in the availability of tests and other factors, Raftery said.

With the true prevalence of COVID-19, Raftery and Irons calculated other useful figures for states, such as the infection fatality rate, which is the percentage of infected people who had succumbed to COVID-19, as well as the cumulative incidence, which is the percentage of a state’s population who have had COVID-19.

Ideally, regular random testing of individuals would show the level of infection in a state, region or even nationally, said Raftery. But in the COVID-19 pandemic, only Indiana and Ohio conducted random viral testing of residents, datasets that were critical in helping the researchers develop their framework. In the absence of widespread random testing, this new method could help officials assess the true burden of disease in this pandemic and the next one.

“We think this tool can make a difference by giving the people in charge a more accurate picture of how many people are infected, and what fraction of them are being missed by current testing and treatment efforts,” said Raftery.

The research was funded by the National Institutes of Health.

For more information, contact Raftery at raftery@uw.edu.

The number of people who live past the age of 100 , up to nearly half a million people worldwide.

There are, however, far fewer “supercentenarians,” people who live to age 110 or even longer. The oldest living person, Jeanne Calment of France, was 122 when she died in 1997; currently, the world’s oldest person is 118-year-old Kane Tanaka of Japan.

Such extreme longevity, according to new research by the 91̽��, likely will continue to rise slowly by the end of this century, and estimates show that a lifespan of 125 years, or even 130 years, is possible.

“People are fascinated by the extremes of humanity, whether it’s going to the moon, how fast someone can run in the Olympics, or even how long someone can live,” said lead author , a 91̽��doctoral student in statistics. “With this work, we quantify how likely we believe it is that some individual will reach various extreme ages this century.”

Longevity has ramifications for government and economic policies, as well as individuals’ own health care and lifestyle decisions, rendering what’s probable, or even possible, relevant at all levels of society.

The , published June 30 in Demographic Research, uses statistical modeling to examine the extremes of human life. With ongoing research into aging, the prospects of future medical and scientific discoveries and the relatively small number of people to have verifiably reached age 110 or older, experts have debated the possible limits to what is referred to as the maximum reported age at death. While some scientists argue that disease and basic cell deterioration lead to a natural limit on human lifespan, others maintain there is no cap, as evidenced by record-breaking supercentenarians.

Pearce and , a professor of sociology and of statistics at the UW, took a different approach. They asked what the longest individual human lifespan could be anywhere in the world by the year 2100. Using Bayesian statistics, a common tool in modern statistics, the researchers estimated that the world record of 122 years almost certainly will be broken, with a strong likelihood of at least one person living to anywhere between 125 and 132 years.

To calculate the probability of living past 110 — and to what age — Raftery and Pearce turned to the most recent iteration of the , created by the Max Planck Institute for Demographic Research. That database tracks supercentenarians from 10 European countries, plus Canada, Japan and the United States.

Using a Bayesian approach to estimate probability, the 91̽��team created projections for the maximum reported age at death in all 13 countries from 2020 through 2100.

Among their findings:

• Researchers estimated near 100% probability that the current record of maximum reported age at death — Calment’s 122 years, 164 days — will be broken;
• The probability remains strong of a person living longer, to 124 years old (99% probability) and even to 127 years old (68% probability);
• An even longer lifespan is possible but much less likely, with a 13% probability of someone living to age 130;
• It is “extremely unlikely” that someone would live to 135 in this century.

As it is, supercentenarians are outliers, and the likelihood of breaking the current age record increases only if the number of supercentenarians grows significantly. With a continually expanding global population, that’s not impossible, researchers say.

People who achieve extreme longevity are still rare enough that they represent a select population, Raftery said. Even with population growth and advances in health care, there is a flattening of the mortality rate after a certain age. In other words, someone who lives to be 110 has about the same probability of living another year as, say, someone who lives to 114, which is about one-half.

“It doesn’t matter how old they are, once they reach 110, they still die at the same rate,” Raftery said. “They’ve gotten past all the various things life throws at you, such as disease. They die for reasons that are somewhat independent of what affects younger people.

“This is a very select group of very robust people.”

The study was funded by the National Institute for Child Health and Human Development.

For more information, contact Pearce at mpp790@uw.edu or Raftery at raftery@uw.edu.

In 2017, a widely cited study used statistical tools to model how likely the world is to meet the Paris Agreement global temperature targets. The analysis found that on current trends, the planet had only a 5% chance of staying below 2 degrees Celsius warming this century — the international climate treaty’s supposed goal.

Now, the same authors have used their tools to ask: What emissions cuts would actually be required to meet the goal of 2 C warming, considered a threshold for climate stability and climate-related risks such as excessive heat, drought, extreme weather and sea level rise?

The 91̽�� study finds that emissions reductions about 80% more ambitious than those in the Paris Agreement, or an average of 1.8% drop in emissions per year rather than 1% per year, would be enough to stay within 2 degrees. The were published Feb. 9 in Nature’s open-access journal .

“A number of people have been saying, particularly in the past few years, that the emissions targets need to be more ambitious,” said lead author , a 91̽��professor of statistics. “We went beyond that to ask in a more precise way: How much more ambitious do they need to be?”

The paper uses the same statistical approach to model the three main drivers of human-produced greenhouse gases: national population, gross domestic product per person and the amount of carbon emitted for each dollar of economic activity, known as carbon intensity. It then uses a statistical model to show the range of likely future outcomes based on data and projections so far.

Even with updated methods and five more years of data, now spanning 1960 through 2015, the conclusion remains similar to the previous study: Meeting Paris Agreement targets would give only a 5% probability of staying below 2 degrees Celsius warming.

Assuming that climate policies won’t target population growth or economic growth, the authors then ask what change in the “carbon intensity” measure would be needed to meet the 2 degrees warming goal.

Increasing the overall targets to cut carbon emissions by an average of 1.8% annually, and continuing on that path after the Paris Agreement expires in 2030, would give the planet a 50% chance of staying below 2 degrees warming by 2100.

“Achieving the Paris Agreement’s temperature goals is something we’re not on target to do now, but it wouldn’t take that much extra to do it,” said first author , who did the research as part of his doctorate at the UW.

The paper looks at what this overall plan would mean for different countries’ Paris Agreement commitments. Nations set their own Paris Agreement emissions-reductions . The United States pledged a 1% drop in carbon emissions per year until 2026, or slightly more ambitious than the average. China pledged to reduce its carbon intensity, or the carbon emissions per unit of economic activity, by 60% of its 2005 levels by 2030.

“Globally, the temperature goal requires an 80% boost in the annual rate of emissions decline compared to the Paris Agreement, but if a country has finished most of its promised mitigation measures, then the extra decline required now will be smaller,” Liu said.

For the top 10 emitters, this is the percentage increase in the country’s emissions-reductions targets that would be required to achieve the Paris Agreement’s goal of keeping warming to 2 degrees C this century. Countries that set more ambitious goals or that have made substantial progress already need a smaller boost now. Russia is not included, since it has exceeded its Paris Agreement commitments.
Credit: P. Liu and A. Raftery/Communications Earth & Environment

The authors also suggest that countries increase their accountability by reviewing progress annually, rather than on the five-year, 10-year or longer timescales included in many existing climate plans.

“To some extent, the discourse around climate has been: ‘We have to completely change our lifestyles and everything,’” Raftery said. “The idea from our work is that actually, what’s required is not easy, but it’s quantifiable. Reducing global emissions by 1.8% per year is a goal that’s not astronomical.”

From 2011 to 2015, Raftery says, the U.S. did see a drop in emissions, due to efficiencies in industries ranging from lighting to transportation as well as regulation. The pandemic-related economic changes will be short-lived, he predicts, but the creativity and flexibility the pandemic has required may usher in a lasting drop in emissions.

“If you say, ‘Everything’s a disaster and we need to radically overhaul society,’ there’s a feeling of hopelessness,” Raftery said. “But if we say, ‘We need to reduce emissions by 1.8% a year,’ that’s a different mindset.”

This research was funded by the National Institutes of Health.

For more information, contact Raftery at raftery@uw.edu.

Widespread use of contraceptives and, to a lesser extent, girls’ education through at least age 14 have the greatest impact in bringing down a country’s fertility rate.

Education and family planning . But new research from the 91̽�� analyzes those factors to determine, what accelerates a decline in otherwise high-fertility countries.

In a published July 23 in Population and Development Review, , a doctoral student in statistics at the UW, and , a 91̽��professor of statistics and sociology, explore two nuanced questions: Is increasing contraceptive use or reducing demand more effective in family planning? And, is it the number of years girls attend school or the overall enrollment of children in school that makes education a factor in fertility?

“Policymakers in countries with high fertility rates are often interested in accelerating their fertility decline, since rapid population growth can lead to a number of unwanted economic, environmental and public health consequences,” said Liu. “Policies that increase access to education and family planning are generally thought to accelerate fertility decline by empowering individuals, particularly girls and women, to achieve their own desires in life. Our work aims to explore what aspects of a country’s education and family planning have the greatest impact on fertility decline.”

As the world’s population builds toward a projected , much of that growth is expected to occur in high-fertility countries of Latin America, Asia and sub-Saharan Africa. The United Nations’ note the role sustainable fertility can play in a country’s environmental, economic and population health, alongside the ways family planning can enable .

Higher fertility rates can stretch a country’s available resources, while rates lower than the “replacement rate” of 2.1 births per woman can lead to a long-term lack of economic growth. Today’s global fertility rate of 2.5 births per woman is down from 3.2 in 1990, but is higher in parts of the world where some countries report fertility rates of at least 4 births per woman.

Liu and Raftery’s study uses UN data on fertility rates since 1970 and combines it with data on education and contraception to determine which factors have the greatest effect. All the countries in their study sample were categorized as transitioning downward, however slowly, from a period of high fertility.

Within the category of family planning, Liu and Raftery looked at two factors over time: contraceptive prevalence, which is the percentage of women using modern contraception; and unmet need, the percentage of women who say they want to delay or stop childbearing but are not using contraception. While the difference between the two metrics may appear small, Liu pointed out that unmet need can reflect hypothetical interest in family planning, whereas contraceptive prevalence reflects actual use. The study found that contraceptive prevalence had a significantly greater effect.

For example, data from El Salvador shows that the link between an increase in contraceptive use and a corresponding decline in fertility rate is especially pronounced. The country’s total fertility rate went from 5.44 births per woman in the mid-1970s — when 28% of women used birth control — to 2.72 births in the mid-2000s, when contraceptive prevalence had more than doubled.

Liu and Raftery also wanted to look at the effect of education on fertility changes. For this, they examined two different aspects of education, both tied to cultural values and economic outcomes: school enrollment and the highest level of education girls typically attain. The latter stems from the academic and professional opportunities available to women and girls, which may affect their childbearing decisions. The former has been hypothesized to affect fertility because if more children go to school, it is more expensive to bring them up, which may discourage families from having more children.

Liu and Raftery found that education affected fertility mostly through the educational attainment of girls, particularly through their early teens (the “lower secondary” level of schooling). Generally considered the last stage of basic education, completing at least the lower secondary level had a greater effect on fertility decline than completing only primary schooling.

Kenya showed a substantial increase in girls’ educational attainment, from 12% reaching the lower secondary level in the mid-1970s to 59% in the mid-2010s. Contraceptive prevalence in Kenya also grew steadily, from 5% to 51%, while the total fertility rate dropped from 7.64 births per woman to 4.06.

Still, of the two factors — family planning and education — family planning played a bigger role in accelerating the transition. “It is important to know that family planning is so critical,” said Raftery. “However, both factors are important and they work together. Education gives women more opportunities as alternatives to having large families, while family planning gives them the means to achieve their goals.”

Overall, sub-Saharan Africa, where the highest-fertility countries are located, showed reductions in fertility but at a slower pace than other high-fertility regions of the world. This may be associated with economic development and cultural values around family size, as well as the quality of education. In line with the UN , policymakers and NGOs should continue to focus on education and on availability and acceptance of contraceptives for women, the researchers said.

The study was funded by the National Institute for Child Health and Human Development. For more information, contact Liu at dhliu@uw.edu.

“The COVID-19 pandemic is generating many different types of data about this disease in communities — ­things like the number of confirmed cases or the number of deaths in a particular area,” said Raftery. “None of these data sources on their own are perfect in terms of capturing a complete and accurate summary of the prevalence of COVID-19 and the risks of doing certain things like opening businesses or schools. All have their own strengths and weaknesses.”

Raftery is lead author of a new published June 11 by the National Academies of Sciences, Engineering and Medicine that is intended to help officials nationwide make sense of these different COVID-19 data sources when making public health decisions.

Officials looking for COVID-19 statistics have plenty to choose from: confirmed cases, deaths, hospitalizations, intensive care unit occupancy, emergency room visits, antibody tests, nasal-swab tests and the ratio of positive test results – to name a few of the more common data points collected and distributed by hospitals and public health agencies. But officials don’t necessarily have all of these statistics on hand when making decisions, or have enough information to interpret them.

“We intend for this guide to help these decision-makers and their advisors interpret the data on COVID-19 and understand the upsides and downsides of each data source,” said Raftery.

For example, the number of positive test results for the novel coronavirus is likely an underestimate of its true prevalence in a community. Many people who have the virus are asymptomatic and aren’t likely to seek out a test, and even people with symptoms may not have access to tests and medical care, according to Raftery. As another example, the number of COVID-19 deaths in a region does not reflect the disease’s current prevalence because the number of deaths lag behind the number of cases by several weeks. In addition, some deaths may be misattributed to COVID-19, Raftery said.

The guide highlights some criteria for officials to take into account when assessing the usefulness of particular COVID-19 data points, including:

• Assessing how representative the data are for a community or region
• Whether there may be systemic biases in some data sources
• Thinking about the types of uncertainties in data sources, due to factors like sample size, how data were collected and the population surveyed
• Whether there’s a time lag due to delays in reporting data, the course of the disease and other factors

“There are no perfect data sources, but all of these data sources are still useful for making decisions that directly impact public health,” said Raftery.

Raftery has worked extensively on statistical methods to measure and estimate the prevalence of other viruses, including HIV in Africa. Though HIV and the novel coronavirus cause different types of diseases, there are similarities in how the two viruses spread among susceptible populations, as well as how types of social distancing — condom use for HIV and physical distancing and mask usage for the novel coronavirus — can decrease transmission. COVID-19 is also generating the same types of data sources, with the same limitations, as HIV/AIDS, such as test results, hospitalization rates and deaths.

Over time, it may be possible to collect more revealing data about COVID-19 from what are known as “representative random samples” within a population. In representative sampling, people are surveyed at random for a disease, and certain populations can be more heavily sampled than others based on what scientists and officials have learned about a disease’s prevalence and susceptibility. Representative sampling avoids biases and can more accurately estimate the disease’s prevalence in a region, according to Raftery.

“As we learn more about COVID-19, how it spreads, how different populations are more or less susceptible, we may be able to move more in the direction of representative sampling,” said Raftery. “The State of Indiana has already done a survey of this kind, and others should follow suit. But there is also a lot that officials can do with the statistics and data sources that hospitals and agencies are providing right now — provided that officials can be made aware of the strengths and weaknesses of each piece of data.”

The guide is the first completed by the National Academies’ — or SEAN — an eight-member committee tasked by the National Academies to provide rapid expert assistance on issues related to the social and behavioral sciences during the pandemic. Raftery is a member of the SEAN and spearheaded this inaugural project.

Co-authors on the guide are , a professor of economics and public affairs at Princeton University; , a professor of public health at Harvard University; and , the provost of Georgetown University. The SEAN and its efforts are funded by the National Academies and the National Science Foundation.

For more information, contact Raftery at raftery@uw.edu.

On today’s increasingly crowded globe, human migration can strain infrastructure and resources. Accurate data on migration flows could help governments plan for and respond to immigrants. Yet these figures, when available, tend to be spotty and error-ridden, even in the developed world. Researchers have developed approaches to estimate migration rates, but even the best of these rely on unrealistic assumptions about the mass movement of people and yield migration rates that can fall far below reality.

In a published online Dec. 24 in the , two scientists at the 91̽�� unveiled a new statistical method for estimating migration flows between countries, using the so-called pseudo-Bayes approach. They show that rates of migration — defined as an international move followed by a stay of at least one year — are higher than previously thought, but also relatively stable, fluctuating between 1.1 and 1.3 percent of global population from 1990 to 2015. In addition, since 1990 approximately 45 percent of migrants have returned to their home countries, a much higher estimate than other methods.

These more accurate estimates of migration will ultimately help both migrants and the people who assist them, said senior author , a 91̽��professor of statistics and sociology.

“Planning for migration is no simple task,” said Raftery. “You need everything from medical infrastructure and trained personnel to elementary schools — and governments rely on accurate demographic estimates to help them put the right plans and responses into place.”

Countries collect migration data through immigration forms at ports of entry, but answers on these forms may contain mistakes, and often fail to collect the types of comprehensive information that demographers need to measure migration accurately. Census forms also tend to ask people where they were born, but usually not when they moved, information which does not accurately reflect the true level of movement.

“Migration is much more than the place you left and the place you ultimately settled in,” said Raftery. “Researchers have tried for years to develop statistical methods that capture the comprehensive picture of human migration across the globe.”

Raftery developed these new migration rate estimates along with his former doctoral student Jonathan Azose, a 91̽��affiliate assistant professor of statistics and the paper’s lead author. They applied the pseudo-Bayes method to migration estimates by incorporating elements of other methods, and calibrated their estimates against a relatively reliable model of migration among 31 European countries.

Azose and Raftery tested the accuracy of pseudo-Bayes against a set of reliable migration figures and discovered that its estimates were typically accurate to within a factor of three, better than many existing estimates.

“For the migration field, this level of accuracy is a significant improvement,” said Azose. “Even when you look at data from European countries, it’s not uncommon for a single migration flow to have estimates that differ by a factor of three or more depending on whether the sending or the receiving country collected the data.”

They also discovered that, compared to other approaches, pseudo-Bayes can more accurately account for return migration, in which migrants go back to their home countries.

“Our estimate shows a higher global flow of migrants in large part because it indicates that return migration is much higher than previously thought,” said Azose.

The researchers estimate significantly higher migration rates than prior methods — between 67 million and 87 million migrants over each five-year period from 1990 to 2015. By comparison, one widely used alternative method, which only estimates the minimum rate of migration, calculates just 34 to 46 million migrants in each five-year period. In addition, while the total number of migrants estimated by pseudo-Bayes increased from 1990 to 2015, the migration rate over that period remained relatively stable as a proportion of global population — between 1.1 and 1.3 percent.

Azose and Raftery also broke down migration rates by emigration, return migration and transit migration, in which migrants move between two countries that are not their countries of birth. In general, from 1990 to 2015, more than 60 percent of migration was emigration. Transit migration never topped 9 percent. Return migration accounted for 26 to 31 percent of migrants, more than twice the rate of other migration estimates. That high rate of return migration added up over time. From 1990 to 2015, approximately 45 percent of migrants ultimately returned to their home countries.

Largest emigration, return migration and transit migration flows in 2010-2015

Credit: Azose and Raftery, PNAS, 2018.

“We estimate a rate of return migration that is significantly higher than other methods, but it is also supported by history,” said Raftery. “For example, during the Rwandan genocide in 1994, more than a million migrants left the country, but most returned within three years after the conflict ended.”

The top migration rates that they observed more recently between countries were consistent with ongoing world events. Migration out of Syria, for example, accounted for two of the top three emigration flows from 2010 to 2015.

The researchers would like to refine their method by differentiating between refugees, which are counted by their method, and other types of migrants. In addition, they would like to incorporate data from non-governmental sources, such as social media records, to improve the accuracy of their estimates, as well as develop approaches to forecast future migration.

The research was funded by the National Institutes of Health and the Center for Advanced Study in the Behavioral Sciences at Stanford University.

###

For more information, contact Raftery at raftery@uw.edu and Azose at jazose@gmail.com.

Grant numbers: R01 HD54511, R01 HD70936.

But the Earth is very likely to exceed that change, according to new 91̽�� research. A study using statistical tools shows only a 5 percent chance that Earth will warm 2 degrees or less by the end of this century. It shows a mere 1 percent chance that warming could be at or below 1.5 degrees, the target set by the 2016 Paris Agreement.

“Our analysis shows that the goal of 2 degrees is very much a best-case scenario,” said lead author , a 91̽��professor of statistics and sociology. “It is achievable, but only with major, sustained effort on all fronts over the next 80 years.”

The , published July 31 in , show a 90 percent chance that temperatures will increase this century by 2.0 to 4.9 C.

“Our analysis is compatible with previous estimates, but it finds that the most optimistic projections are unlikely to happen,” Raftery said. “We’re closer to the margin than we think.”

The most recent report from the Intergovernmental Panel on Climate Change included future warming rates based on four scenarios for future carbon emissions. The scenarios ranged from “business-as-usual” emissions from growing economies, to serious worldwide efforts to transition away from fossil fuels.

“The IPCC was clear that these scenarios were not forecasts,” Raftery said. “The big problem with scenarios is that you don’t know how likely they are, and whether they span the full range of possibilities or are just a few examples. Scientifically, this type of storytelling approach was not fully satisfying.”

The new paper focuses instead on three quantities that underpin the scenarios for future emissions: total world population, gross domestic product per person and the amount of carbon emitted for each dollar of economic activity, known as carbon intensity.

Using statistical projections for each of these three quantities based on 50 years of past data in countries around the world, the study finds a median value of 3.2 C (5.8 F) warming by 2100, and a 90 percent chance that warming this century will fall between 2.0 to 4.9 C (3.6 to 8.8 F).

“Countries argued for the 1.5 C target because of the severe impacts on their livelihoods that would result from exceeding that threshold. Indeed, damages from heat extremes, drought, extreme weather and sea level rise will be much more severe if 2 C or higher temperature rise is allowed,” said co-author , a 91̽��associate professor of atmospheric sciences. “Our results show that an abrupt change of course is needed to achieve these goals.”

Raftery previously worked on United Nations projections for future world population. His 2014 study used Bayesian statistics, a common tool used in modern statistics, to show that world population is unlikely to stabilize this century. The planet likely will reach .

In the new study, Raftery expected to find that higher populations would increase the projections for global warming. Instead, he was surprised to learn that population has a fairly small impact. That is because most of the population increase will be in Africa, which uses few fossil fuels.

What matters more for future warming is the carbon intensity, the amount of carbon emissions produced for each dollar of economic activity. That value has dropped in recent decades as countries boost efficiency and enact standards to reduce carbon emissions. How quickly that value drops in future decades will be crucial for determining future warming.

The study finds a wide range of possible values of carbon intensity over future decades, depending on technological progress and countries’ commitments to implementing changes.

“Overall, the goals expressed in the Paris Agreement are ambitious but realistic,” Raftery said. “The bad news is they are unlikely to be enough to achieve the target of keeping warming at or below 1.5 degrees.”

The research was funded by the National Institutes of Health. Other co-authors are , a 91̽��graduate now at Upstart Networks in Palo Alto, California; , a 91̽��professor emeritus of economics who now holds a position at the University of California, Santa Barbara; and , a 91̽��doctoral student in statistics.

###

For more information, contact Raftery at 206-543-4505 or raftery@uw.edu and Frierson at 206-685-7364 or dargan@uw.edu. Note: Raftery is in Baltimore and best reached via email.

NIH grants: HD054511, HD070936

On March 15 at a ceremony in Washington, D.C., Prime Minister Enda Kenny of Ireland honored , a professor of statistics and sociology at the 91̽��, for his diverse contributions to the field of statistics. Kenny presented Raftery with the St. Patrick’s Day Medal, which is awarded each year by Science Foundation Ireland (SFI).

Raftery, who has worked at the 91̽��since 1986, develops new statistical methodology, with a focus on the social, environmental and health sciences. His methods have been applied to problems ranging from weather forecasting to population growth projections.

“I am very happy to be recognized by Science Foundation Ireland,” said Raftery. “It means a lot to me to be honored by my own country.”

SFI, the Irish government’s primary agency for funding and promoting STEM research, engagement and education, established the St. Patrick’s Day Medal in 2014 to honor Irish-born scientists who live and work in the United States. SFI initially awarded the medal to just one scientist per year, but last year SFI started to honor two scientists each year, one from academia and one from industry.

Raftery was SFI’s 2017 medal recipient for academic research, while T. Pearse Lyons received the medal this year from industry. Lyons is founder and president of Alltech, Inc., a Kentucky-based agricultural products and food science company.

In released by the foundation for the ceremony at the United States Institute of Peace, Kenny, who also uses the Irish title of his office, “An Taoiseach,” praised the impact that Raftery and Lyons have had on their fields.

“They have demonstrated how academic and industry based scientific research can create jobs, tackle global problems and impact positively on people and society,” said Kenny. “These distinguished medal recipients are driving globally significant innovation in the areas of agriculture, food production, health, and population and weather forecasting, to name just a few.”

Raftery was born in Dublin and became interested in statistics as a secondary student at Dublin’s St. Conleth’s College. He earned a bachelor’s degree in mathematics in 1976 and a master’s degree in statistics and operations research in 1977, both from Trinity College Dublin. After obtaining his doctoral degree in mathematical statistics from the Université Pierre et Marie Curie in Paris in 1980, Raftery served as a lecturer at Trinity College Dublin before joining the 91̽�� in 1986.

At the UW, Raftery has worked to develop new statistical methods for scientific problems. A wide range of scientific fields, particularly in the “big data” era, rely on statistical tools to draw conclusions about the world at large from the results of specific experiments or series of observations. Raftery has developed, tested and applied statistical models and methods to improve and sharpen these tools for better forecasting and estimates of probability.

“To a large extent, this award is for the work of my amazing Ph.D. students and collaborators at UW, in Ireland and elsewhere over the years,” said Raftery. “It reflects the strategic impact of statistical thinking on science and policy. I am happy to have been able to help colleagues in a variety of scientific fields with new statistical methods to interpret their data.”

Raftery served as founding director of the 91̽�� from 1999 to 2009. He became a fellow of the American Academy of Arts and Sciences in 2003 and a member of the National Academy of Science in 2009. Over the course of his career, Raftery has authored over 200 articles in scientific journals, and Thomson Reuters named him the most-cited researcher in mathematics in the world from 1995 to 2005.

The wide reach of Raftery’s work is also evident in his long list of collaborators across scientific fields. For example, he has worked with marine biologists to estimate the population of bowhead whales in the northern Pacific and Arctic oceans, and with U.N. scientists to how HIV infections spread among vulnerable, hard-to-reach populations in parts of Europe and Africa. He has also applied his models to studies of air quality, , social mobility, gene expression in cells, fertility, and family structure, among other topics. He also found time to help the Irish government estimate the Emerald Isle’s production capacity for wind power.

That is the type of global impact that SFI seeks to honor with the St. Patrick’s Day Medal, according to SFI Director General Mark Ferguson.

“The research undertaken by these two Irish leaders has had a profound and diverse impact across the globe,” said Ferguson in a statement.

###

For more information, contact Raftery at 206-543-4505 or raftery@uw.edu.