Crop irrigation could be cooling Midwest

ATLANTA — If summers seem cooler and wetter in parts of the Midwest in recent years, you can thank — or blame — farmers, two new studies contend.

While average global temperatures rose about 0.74 degrees Celsius during the past century, the U.S. Midwest has experienced a noticeable slump in summer temperatures in recent decades, reported David Changnon, a climatologist at Northern Illinois University in DeKalb, on January 19 at the annual meeting of the American Meteorological Society.

On average, daily high temperatures in Chicago rise above 90° Fahrenheit (32.2° Celsius) on 24 days each summer. But from 2000 through 2009, only two years tallied more than 24 days hotter than 90°— the lowest decadal total in 80 years, Changnon noted.

Rather than being just a statistical anomaly, the recent cool temperatures seem to be part of a steady long-term decline in summertime highs in Chicago, Changnon and his colleagues found. The last 10 years have seen a total of only 172 days above 90°; the 1930s saw more than twice as many. And Chicago wasn’t alone. The team noted a comparable decline in unusually hot days at 13 other sites in a swath stretching from western Iowa through Illinois to eastern Indiana.

From 1970 through 2009, average high temperatures at the sites in Iowa and Illinois during July and August were between 0.5 and 1.0 degrees F (0.28 and 0.56 degrees C) cooler than they were for the years 1930 through 1969, the researchers found. The amount of precipitation received in the region has changed substantially as well: Average rainfall for July and August from the 1970s through 2009 was about 0.33 inches (0.8 centimeters) higher each month than it was from the 1930s through the 1960s.

Changnon suggested that fewer hot days and more precipitation are linked, because humid air warms more slowly than dry air does. One likely source of the extra moisture is the region’s agriculture. Plants pump vast amounts of water from surface soil into the atmosphere as they grow, and thirsty row crops such as corn and soybeans are much more prevalent in the region these days — about 97 percent of farmland is planted in those crops now, versus about 57 percent in the 1930s, Changnon notes. Also, the plants are spaced more closely now (about 30 inches apart, versus the 40-inch spacing typical in the 1930s), a trend that has boosted the numbers of water-pumping plants per acre by about 60 percent.

Even if much of the extra summer rainfall in the Midwest derives from water in local soils, the original source of that moisture might be an irrigation spigot somewhere on the Great Plains. A rapid rise in irrigation in that region apparently has boosted precipitation downwind in the Midwest, Alan Robock, an atmospheric scientist at Rutgers University in New Brunswick, N.J., reported January 21 at the meeting. Rather than running off into rivers and streams or soaking back into the ground, Robock added, most of that liberated groundwater was used by plants, evaporated into the air and was carried downwind, where it condensed in clouds and then fell to the ground as rain.

In 1930, in a swath of plains that stretches from South Dakota down to Oklahoma and the Texas panhandle, farmers irrigated only about 7,500 square kilometers, an area about half the size of Connecticut. But by 1980, he notes, irrigated farmland in this same area covered about 60,000 square kilometers. During the 20th century, irrigation pulled more than 333 cubic kilometers of groundwater from aquifers beneath the Great Plains.

When Robock and his colleagues analyzed precipitation at more than 300 weather stations from central Wisconsin and Michigan down to northeastern Arkansas and northwestern Tennessee — a region that includes much of the same area studied by Changnon and his team — they too found that rainfall had increased. At sites in that swath, precipitation during a typical July late in the 20th century was between 25 and 50 percent higher that it was early in the century.

A variety of climate simulations don’t show such long-term changes in precipitation, probably because those simulations don’t take irrigation into account, Robock and his colleagues say. July, Robock noted, is when irrigation on the Great Plains is most profligate, with more than one-third of groundwater withdrawal occurring during that month.

Future analyses will compare the results of simulations that include irrigation with those that don’t, Robock said. Results of those studies might allow the team to more confidently pin the blame for the region’s increased July rainfall on Great Plains irrigation, he noted.