Climate scientists suggest that changes in land use such as deforestation -- not just greenhouse gas emissions -- can play a significant role altering the world's climate systems.

A new study by researchers at the Massachusetts Institute of Technology (MIT) and Dartmouth College reveals how another type of land use called intensive agriculture can affect regional climate.

The researchers showed that in the last half of the 20th century, the Midwest U.S. went through an intensification of agricultural practices that led to dramatic increases in the production of corn and soybeans. Also, over the same period in that region, summers were significantly cooler and had more rainfall than during the previous half-century.

This effect, with regional cooling during a time of overall global warming, may have masked part of the warming effect that would have occurred over that period, and the new findings could help to refine global climate models by incorporating such regional effects, MIT and Dartmouth said.

The findings are being published this week in Geophysical Research Letters.

The team showed that there was a strong correlation -- in both space and time -- among the intensification of agriculture in the Midwest, the decrease in observed average daytime temperatures in the summer and an increase in the observed local rainfall.

In addition to this circumstantial evidence, they identified a mechanism that explains the association, suggesting that there was, indeed, a cause-and-effect link between the changes in vegetation and the climatic effects, the announcement said.

Elfatih Eltahir, the Breene M. Kerr professor of hydrology and climate at MIT, explained that plants "breathe" in the carbon dioxide they require for photosynthesis by opening tiny pores called stoma, but each time they do this, they also lose moisture to the atmosphere.

With the combination of improved seeds, fertilizers and other practices, the annual yield of corn in the Midwest increased about four-fold, while soybean yields doubled between 1950 and 2009, the researchers said. These changes were associated with denser plants with more leaf mass, which then increased the amount of moisture released into the atmosphere. That extra moisture served to both cool the air and increase the amount of rainfall, the researchers suggested.

"For some time, we've been interested in how changes in land use can influence climate," Eltahir said, noting that the problem is independent from carbon dioxide emissions, which have been studied more intensively.

Eltahir, recent MIT post-doctoral researcher Ross Alter and their co-authors noticed that records showed that over the course of the 20th century, "there were substantial changes in regional patterns of temperature and rainfall. A region in the Midwest got colder, which was a surprise," Eltahir said. Because weather records in the U.S. are quite extensive, there is "a robust data set that shows significant changes in temperature and precipitation" in the region.

Over the last half of the century, average summertime rainfall increased by about 15% compared to the previous half-century, and average summer temperatures decreased by about 0.5°C. The effects are "significant, but small," Eltahir said.

By introducing into a regional U.S. climate model a factor to account for the more intensive agriculture that has made the Midwest one of the world's most productive agricultural areas, Eltahir said the research team found that "the models show a small increase in precipitation, a drop in temperature and an increase in atmospheric humidity" — exactly what the climate records actually show.

That distinctive "fingerprint," he said, strongly suggests a causative association. "During the 20th century, the Midwestern U.S. experienced regional climate change that's more consistent with what we'd expect from land use changes as opposed to other forcings," he said.

This finding in no way contradicts the overall pattern of global warming, Eltahir emphasized, but in order to refine the models and improve the accuracy of climate predictions, "we need to understand some of these regional and local processes taking place in the background."

Unlike land use changes such as deforestation, which can reduce carbon dioxide absorption by trees that can help to ameliorate emissions of the gas, the changes in this case did not reflect any significant increase in the area under cultivation but, instead, a dramatic increase in yields from existing farmland.

"The area of crops did not expand by a whole lot over that time, but crop production increased substantially, leading to large increases in crop yield," Alter explained.

The findings suggest the possibility that, at least on a small-scale regional or local level, intensification of agriculture on existing farmland could be a way of doing some local geo-engineering to at least slightly lessen the effects of global warming, Eltahir said. A recent paper from another group in Switzerland suggests just that.

The findings could also portend some negative impacts,  however, because the kind of intensification of agricultural yields achieved in the Midwest are unlikely to be repeated. Also, some effects of global warming may "have been masked by these regional or local effects, but this was a 20th-century phenomenon, and we don't expect anything similar in the 21st century," Eltahir said. So, warming in that region in the future is not expected to have "the benefit of these regional moderators."