Extremely Bad Weather

Studies start linking climate change to current events

6:35am, November 2, 2012
Sponsor Message

Texas spent 2011 baking. About half the state was gripped by what climate scientists described as an “exceptional” drought, one that goes beyond their categories of severe, or even extreme.

Texans are used to dry, but this was worse than the Dust Bowl and drier than the crippling decade-long drought of the 1950s. In fact, it was the driest year since record-keeping began in 1895. As rivers dried up and farmers scrambled to irrigate, many public water systems reported that they were within six months of running out of water. Agricultural extension agents pegged crop and livestock losses at a staggering $5.2 billion.

Problems that people hadn’t anticipated started to pop up. In a state obsessed with football, Dallas had to close more than two dozen athletic fields as clay soils shrank and formed cracks more than half a meter deep. One county hired some 100 bridge and road crews to fix pavement cracks all summer long. The shifting soil twisted water mains, breaking more than 200 in Fort Worth alone, 20 in a single day.

And 2011 didn’t bring bad weather just for Texas. At least 1,625 tornadoes touched down across the United States. Half a world away, heavy monsoons in Asia triggered record floods, causing tens of billions of dollars in damages.

Most recently in 2012, the U.S. Northeast was pummeled by Hurricane Sandy. Earlier in the year the nation faced blazing heat and a crop-slaying drought across the Midwest. From January through September, U.S. cities matched or broke more than 29,300 high-temperature records. By the end of September, nearly two-thirds of the contiguous United States was suffering moderate to exceptional drought — including 80 percent of the nation’s farmland.

In the wake of such events, everyone from insurance companies to Congress, homeowners and government agencies has been asking whether global warming might have played some role. And scientists have been working on ways to figure out just how much of this weird weather has come from natural variability, and how much might be driven by warming from greenhouse gases.

“The breaking of records is the best indication that we’re outside the normal range of simply weather,” argues Kevin Trenberth of the National Center for Atmospheric Research in Boulder, Colo. Under normal variability, regions should experience the same number of high and low temperature records in any given year or decade, on average. “But we did some statistics on the first six months of this year,” he says, “and there were around nine times as many records being broken on the high side as there were on the cold side.”

Such a short-term trend could be a fluke, but the same pattern shows up when scientists look back over decades. Work published in Geophysical Research Letters in 2009 showed that record highs and lows happened about equally in the 1950s, but from 2000–2009 record-high temperatures outnumbered record lows 2-to-1.

An approach that climate scientists call “attribution” may help explain this apparent rise in extremes. Initially, attribution studies looked for the drivers of climate change over decades, using computer simulations of climate and statistical analyses of past conditions. Such studies have helped show that human activities are contributing to rising temperatures, by establishing that observed changes cannot be explained without including the effect of greenhouse gases. Now, a twist on attribution studies is looking at whether a warming climate may be shifting the odds of extreme weather events such as floods, droughts and heat waves.

Knowing if extreme events will recur every few decades instead of once in a century is becoming vital for government agencies in charge of building codes, water supplies and emergency planning, says climatologist Thomas Peterson of the National Climatic Data Center in Asheville, N.C. “There’s much each can do to adjust how they manage the environment and their budgets,” he says, “as we understand how the odds of extreme weather are changing.”

Altered odds

Climate scientists have long warned that no single weather event — say, Hurricane Katrina — could be attributed to climate change. After all, even if warming doubled the number of category 5 hurricanes per year, it would be impossible to say whether Katrina was one of the extras. But in some cases researchers can now calculate whether climate change upped the odds of a major hurricane, heat wave or other type of event.

Some scientists have likened the effect of global warming on weather to the use of steroids by a baseball slugger, intensifying a hitter’s power. If the player hits 10 percent more home runs, it’s impossible to know which particular home runs were the result of drug use, but the chances of his hitting any one of them was 10 percent higher because of the steroids. Other scientists use a similar analogy: Climate change, by bumping the average temperature up, moves the ballpark fences closer so that a home run is easier. Either way, actually doing the attributing works the same way: Scientists calculate how climate change affects the odds of a given event happening.

This change is most evident so far for extreme heat. Many regions of the globe have begun experiencing spells of anomalous warmth — a trend that’s only expected to build. An analysis in the September 11 Proceedings of the National Academy of Sciences, for instance, found that unusual summer heat — extremes characterized by seasonal temperatures at least 3.2 degrees Fahrenheit above the regional average — covered roughly 1 percent of the planet on average between 1951 and 1980. Over the next 30 years, the land area affected by extreme temperatures mushroomed to roughly 10 percent. Within a decade, the value could spike to 16.7 percent, concluded James Hansen of NASA’s Goddard Institute for Space Studies in New York City and his coauthors (SN: 9/8/12, p. 10).

People may not notice warmer temperatures or other anomalies, however, when they fall in seasons that normally aren’t the hottest, coldest, rainiest, driest or windiest.

As Trenberth explains it, when temperatures in the U.S. midsection from March to May are a few degrees above normal, people might see it as a welcome respite from the chill. At a minimum, these are temperatures people will normally encounter at some point during the year. “It’s when those record-breaking temperatures occur in June and July — as we had this year — that you exceed the bounds of previous experiences, and people suddenly take notice,” he says. But the elevated heat on summer days does not necessarily have a greater effect on ecosystems than higher temperatures in other months, he says.

Markus Donat and Lisa Alexander of the University of New South Wales in Sydney pored over global temperature records for the same years as Hansen’s study, but concentrated instead on daily temperatures. Compared with the earlier period, “we saw a skewing in the distribution of temperatures to warmer values,” Donat says. The threshold temperature needed to qualify a day as being in the top 5 percent between 1951 and 1980 now occurs 40 percent more often, they reported July 31 in Geophysical Research Letters.

Neither of those studies tried to link these trends to global warming or associated human activities. That’s where attribution science steps in, and it’s challenging.

Scouting for a climate-change signature within transient events lasting just months and affecting only a small part of the globe is equivalent to hunting for a faint fingerprint amid the heavy boot prints of weather’s natural variability.

The first big advance came in 2004 when Peter Stott of the Met Office, which is the United Kingdom’s weather service in Exeter, England, and his colleagues linked Earth’s globally warming surface temperatures to a savage heat wave in Europe the year before.

Over the summer of 2003, more than 70,000 excess deaths across 12 European nations resulted from what may have been the region’s hottest protracted period in 500 years. Using temperature readings collected since the 1920s, computer analyses by Stott’s group identified an increased chance of excessive warming in 2003 — at least a doubling — “attributable to human influence on climate.” And by the end of the century, those analyses projected, “2003 would be classed as an anomalously cold summer.”

Russian heat

The poster child for attribution science, however, may be the 2010 Russian heat wave. Blistering summer temperatures devastated regional crop yields, fostered widespread wildfires and killed thousands. It took quite a while to crunch the numbers, but two recent attribution studies now indict climate change as a coconspirator.

Earth’s slowly growing fever played an increasing role in breaking July temperature records in Moscow over the last decade, according to climate simulations by Stefan Rahmstorf and Dim Coumou of the Potsdam Institute for Climate Impact Research in Germany. Their work, reported late last year in the Proceedings of the National Academy of Sciences, suggested an 80 percent likelihood that global warming intensified the 2010 Russian heat wave.

A Sept. 5 report in the Journal of Geophysical Research by Trenberth and John Fasullo, also of the National Center for Atmospheric Research, points to factors that may have set the stage.

Previous analyses had looked in Russia for evidence of what caused a huge high-pressure air mass to stall over the region, baking it dry with warm and stable sunny weather. “What we show,” Trenberth says, “is that’s really a mistake. A lot of things going on all around the world now appear very much related to the Russian heat wave.” These included record-breaking high sea-surface temperatures covering an area from the Caribbean to the northern Indian Ocean and Australia. The analysis also linked Russia’s heat to anomalous monsoon circulation patterns from Pakistan to South America and northeastern Australia.

Although a large share of these predisposing conditions trace to natural variability, Trenberth says, “it’s also fairly clear that there was a significant climate change aspect,” particularly in exaggerating sea-surface temperatures. The upper ocean’s record heat set up atmospheric patterns that pushed conditions known collectively as the Mediterranean climate from southern Europe to Russia, where those conditions persisted for months.

Measuring that fingerprint

But most scientists aren’t content to merely probe whether the severity of weather extremes can be attributed in some part to climate change. They’d also like to quantify how much of the blame climate change deserves.

One problem: Natural variability in Earth’s climate makes some years warmer than others, says Richard Seager of Columbia University’s Lamont-Doherty Earth Observatory in Palisades, N.Y. “The dominant features,” he notes, “are El Ni±os, during which the global mean temperature goes up, and [countervailing] La Ni±a events where the global mean temperature goes down.” But other factors — from volcanic eruptions and local cloudiness to human-caused changes like urban development and crop irrigation — can also affect winds, rainfall and sunlight.

This variability tends to smooth out on the global scale, as local effects blend together. “But as you move to smaller spatial and temporal scales,” Seager says, “the amount of random variability in climate gets bigger and bigger relative to any global-warming signal.”

He says that’s why attribution studies have largely focused on weather events that tend to be regionally broad and long-lived.

As in Texas’ scorching 2011 drought.

Meteorologist Martin Hoerling of NOAA’s Earth System Research Laboratory in Boulder, Colo., and collaborators probed what drove Texas’ recent drought, which set in months before a massive heat wave gave the region a double whammy. They found that drought has tended to precede extreme hot spells in the region.

For Texas, he says, “the dry antecedent conditions made the heat wave that much more extreme.” Once soils dry out, there’s no moisture to cool the environment through evaporation, so the heat just bakes the air and ground hotter and hotter. The 2011 extreme developed in part because of La Ni±a conditions, which as they cool the ocean surface in the eastern equatorial Pacific also tend to lead to drier weather in Texas. His team’s findings have been submitted for publication.

Those results are consistent with findings published in the July Bulletin of the American Meteorological Society, which concluded that compared with the 1960s, global warming appears to have increased by 20 times the risk that extreme heat will accompany La Ni±a conditions in Texas(SN: 8/11/12, p. 14).

“Dryness alone would have explained more than half of how hot it was that summer,” Hoerling says. It now appears that climate change contributed another 1 degree Fahrenheit — “or about 20 percent of the magnitude of the heat wave.”

Don’t underestimate the impact of just one extra degree, Seager warns. “It may be all you need to exceed some threshold, such as the temperature at which railway tracks buckle.”

When it rains

Seager’s team and others have also been probing recent changes in precipitation extremes against the backdrop of Earth’s warming atmosphere. He says these studies show that wet years get wetter and dry years get drier — and the length of dry spells increases.

Overall, roughly the same amount of water is available to evaporate and rain down in any year. But because a warmer atmosphere can hold and transport more moisture, warming can ramp up the activity of the whole precipitation cycle, Seager notes. Not surprisingly, he says, rain gauge data have begun showing that “an increasing proportion of rain is falling in the heaviest rain events.”

In the May Journal of Climate, Seager and his colleagues mapped the variability of rains and evaporation. Their attribution analyses showed increased variability in monsoons and other precipitation. The researchers concluded that these trends, “no longer natural but a mixed hybrid of [natural] variability and human-induced climate change, will only become more extreme.”

But not every monster flood carries a climate-change signature. Take the unusually wet 2011 monsoon that submerged much of Bangkok and other parts of Thailand under 2.5 meters of water for up to two months, causing an estimated $45 billion in damage. Though the scale of the flooding was unprecedented, human development and water management policies bore most responsibility, according to an attribution study by Geert Jan van Oldenborgh of the Royal Netherlands Meteorological Institute in De Bilt and his colleagues.

Reported in the July Bulletin of the American Meteorological Society, their analysis showed that 2011 monsoon rainfall differed little from some in the recent past. What set last year apart, they concluded, was how locks and dam levels along catchment streams had been managed, together with a recent decision to foster the development of high-value properties within the region’s floodplain.

Gales versus twisters

Among rainy extremes, hurricanes pose the biggest risks. Meteorological analyses predict that the total number of tropical cyclones should diminish dramatically in a warmer world, says climate scientist Kerry Emanuel of MIT. But those that do develop will increasingly be strong ones, he adds. And that’s important, “since more than 80 percent of the damage produced by tropical cyclones in the United States come from storms category 3 or higher — the ones projected to increase.”

In the Atlantic, hurricane power has more than doubled since the 1980s, Emanuel has found. This change tracks the uptick in sea-surface temperatures there over the last 30 years, which he says “is almost entirely due to greenhouse gases.”

But despite the correlation, there isn’t a long enough track record of hurricane numbers and intensities to establish causality, even if it is expected, Seager warns. Unlike temperature and rain data, which go back a century or more in many places, good tropical cyclone data date back only to the late 1970s and the widespread use of weather satellites to spot these storms.

What’s more, reliable intensity data exist only for Atlantic hurricanes, Emanuel notes, where airplane reconnaissance has been able to measure wind speeds. Still, he argues, “the Atlantic link between hurricane power and sea-surface temperature for the tropical summer­time is really spectacular — and hard to deny.”

In an analysis in the March Nature Climate Change, Emanuel and colleagues used attribution to project trends for Atlantic hurricane development over the next century. Climate change, they showed, “has a negligible effect on common small storms, but increases the intensity of large [hurricanes].” Warming will make storms causing monster damage more frequent, they say.

The role of global warming in tornadoes is far less certain.

Yet after the United States was hammered last year by a record number of twisters for a single month — 758 in April — and seven that caused at least $1 billion in damages each, the public and policy makers have been clamoring for an explanation, says Harold Brooks of NOAA’s Severe Storms Laboratory in Norman, Okla. Unfortunately, he says, the limitations on attributing hurricanes to climate change pale in comparison to what little can be said about tornadoes.

2011 twister numbers “were ridiculously high — absolutely off the charts,” Brooks says. But this year, tornado rates are about the lowest on record. “That’s what we mean by high interannual variability,” he says.

As with hurricanes, historical data on tornado occurrence was spotty until widespread radar began in the 1970s and the National Weather Service started rating tornado intensities. Yet even today, those intensity ratings remain a judgment call, Brooks notes, because wind speeds have been measured directly for only a handful of twisters. The rest have been inferred from the damage caused. And that’s problematic because the structural integrity of what’s in a tornado’s path can vary greatly, as can the experience of those who evaluate storm damage.

Global warming may be driving real changes in tornado numbers or intensities, Brooks concedes. But because their numbers vary so much from year to year, his analyses suggest that it could take another century before scientists can gather enough data for a compelling signal to emerge.

On the horizon

For now, running mammoth climate models and analyzing their outputs to find a global-warming signal takes nine months to a year — even for events as big as the Texas and Russian heat waves, notes Hoerling. “It’s tough to push it faster than that.”

But scientists would like to. And an experimental program has begun spitting out predictions for weather events expected in days to weeks. The analyses were developed at the University of Cape Town, South Africa and are now being fine-tuned in collaboration with researchers at the Met Office and Lawrence Berkeley National Laboratory in California.

“Our forecast system is not just looking for major weather events, but for ‘impact’ events,” explains climate researcher Dáithí Stone of the Berkeley lab. “Like what’s the chance of flooding in a certain area, and has that changed because of our emissions?”

These forecasts attempt to address in real time whether greenhouse gas emissions share some blame for a severe weather event. During the first few years of testing, the simulations failed to forecast extreme events in advance with accuracy at the regional level, Stone says. So now the team is refining these analyses in hopes of discerning future details of such regional events.

“We’re not entirely sure if that is possible,” Stone concedes. “But we also won’t know without trying.”

Attributing extremes
Recent attribution studies look for the fingerprint of climate change on weather events. Not all types of extreme weather have been linked to climate change, and some are harder to study than others.

Heat versus chill Regional heat waves have shown the strongest signature of global warming. A stable climate exhibits as many warm as cold extremes on average, yet in recent decades new heat records vastly outnumber those for extreme cold snaps — a trend that has been growing.

Drought and floods Studies have been linking global warming with extremes in rainfall. So when it rains, it can really pour — and when it doesn’t, a dry spell can persist far longer than was typical in former decades. Drought then boosts the impact of local warming, fostering especially prolonged, blistering heat spells. Attribution studies have also found that an extended drought conspired with global warming to foster the unprecedented Texas heat wave of 2011.

Hurricanes Recent attribution analyses find that warming sea-surface temperatures in recent decades have preceded a suspicious increase in hurricane intensity — at least in the Atlantic, for which the best data exist. Projections of future hurricanes indicate that their actual numbers will fall with increased warming, although the incidence of the most extreme storms will increase.

Wildfires The average size of wildfires in the United States so far in 2012 (182.4 acres for January through September) is the largest since 2000 for those months. The total area burned, 8.8 million acres, is the second highest since 2000. Studies have linked climate change to more persistent droughts, which up wildfire risk.

Tornadoes The most extreme among storms, tornadoes currently show the greatest year-to-year variability. The United States saw 1,625 twisters in 2011, racking up more than $28 billion in damage. So far, 2012 has proven an exceptionally quiet tornado year. But if there is a link between twister numbers or intensities and climate change, it may not show up for perhaps another century, owing to a lack of reliable long-term data on these storms. — Janet Raloff

More from this issue of Science News