If your crop yields are accelerating more each year, is it because you are paying your seed dealer, or paying the preacher? In other words, if you are experiencing yield improvements at a higher rate than ten years ago, is that change due to the technology in the seed bag or the fact that Cornbelt weather has been more favorable? Or is there any yield acceleration at all?

Crop weather specialist Mike Tannura at the University of Illinois was joined by agricultural economics colleagues in looking at the yield trends and the weather in Illinois, Iowa, and Indiana, all of which have similar crop production practices, similar weather, and produce about 50% of the corn crop. Their analysis looks at the widespread perception that trendline yields have accelerated since the mid-1990’s, compared to the longer trendline yield from 1960 to 2007. They looked at precipitation both pre-season and growing season, as well as temperature, and utilized some of the longtime efforts of Iowa State weather specialist Louis Thompson.

Since 1996 they found that the trendline for corn yields changed very little in the three states and at most .2 bu. which was described as statistically insignificant, after adjusting for the weather. They characterized the weather as fairly benign for corn development since the mid-1990’s. Similar to the yield jump from single cross hybrids in the 1930’s and the reliance upon nitrogen fertilizer in the 1950’s, many folks in agriculture attribute yield increases to improved seed technology. That would include triple stack corn which has only been yield tested for 1 year, and the speculation that Bt corn genetics also improve its use of nitrogen.

(Along the way, the researchers found, “Unfavorable weather reduced yields by a much larger amount than favorable weather increased yields. For example, 2 inches higher than average July precipitation in Illinois increased corn yields 6 bushels per acre, while 2 inches less than average reduced yields 16 bushels per acre.” They also found that, “Corn yields increased at the fastest rate in Iowa and Illinois, with annual increases of 2.1 and 2.0 bushels per year, respectively. Trend yield increases in Indiana were slightly lower at 1.7 bushels per year.”)

Their statistical analysis indicated both that seed technology and weather really did not fully explain any yield improvement in recent years. However, they measured about a 1 bushel per acre increase attributable to the change to single cross corn, and a 1.7 bushel per acre increase attributable to nitrogen use. They were intrigued that biotech seeds may also result in a yield increase over the period of 2008 to 2030, but one that is closer to 225 bushels per acre instead of 300 bushels per acre. But leaving a question mark in their minds was whether cyclical weather patterns may also at work as Dr. Thompson had suggested.

The Tannura, Irwin, and Good analysis was part of more lengthy research which focuses on the relationship of crop yields and crop weather statistics that will be of concern if there are increases in global temperatures.
Tannura and colleagues note that analysis of weather impacts on crops can be difficult, given the fact that 4 inches of rain in a month could be spread over several showers, or could be just one gullywasher. They looked at weather studies back to 1914, which determined July rainfall was the most important for corn, and another in 1920 which said temperature played an important role. Following up on Thompson’s statistics which spanned 1930 to 1962, Tannura’s group explored the potential for yield acceleration since 1960 and including the new biotech period.

Tannura says, “The states of Illinois, Indiana, and Iowa were chosen for the present study because they represented 43% to 45% of U.S. corn and soybean production from 2000 through 2006. These states also have similar climate and planting dates. Consideration was given to including Minnesota and Nebraska since they often rank in the top five corn and soybean producing states. However, Minnesota was excluded because its northern climate is more susceptible to damaging early- and late-season frosts that may not be detected by monthly weather observations. Nebraska was excluded because a high proportion of its crops are irrigated, which skews weather-yield relationships.” Among the findings are:
1) Indiana is the wettest during the pre-season, which is defined as total precipitation from September through April.
2) Illinois is drier by a small amount.
3) However, Iowa averages approximately 6.00” to 8.00” less during the pre-season period. This is primarily due to considerably drier weather during the winter.
4) Precipitation during June in Illinois and Iowa both averaged around 4.00”, which was slightly drier than in May.
5) However, Iowa averaged 0.50” more precipitation than in Illinois or Indiana, which was slightly wetter than in May.
6) A review of the median also shows Iowa was the wettest during June. At the extremes, Illinois and Indiana have been much drier than Iowa in June with minimum values of 1.05” and 0.74”, respectively.
7) Illinois was the warmest of the three states throughout the growing season from May through August, while Iowa was always the coolest.
8) The month of May was the coolest with average temperatures around 62°F to 63°F, while July was the hottest with averages around 74°F to 75°F.

Regarding yields, corn yields climbed quickest in Iowa at 1.9 bu. per year, with Illinois and India at 1.7 bu. Soybean yields increased by 0.5 bu. in Iowa and Indiana, with Illinois at 0.4 bu. “Yields from 1960 through 2006 were adjusted to the level of technology available in 2006. Iowa averaged the highest de-trended corn yield, 159.1 bushels per acre, followed by Illinois at 155 bushels per acre and Indiana at 149.6 bushels per acre, respectively.” Tannura’s group also reports, “Monthly precipitation and temperatures during sensitive periods of crop development were stable across the sample period. Although precipitation and temperature observations showed large year-to-year variability, a clear trend was not detected. The lack of a trend and earlier evidence of low month-to-month correlations support the view that monthly temperatures and precipitation from 1960 through 2006 were random. Finally, the fact that an upward trend in temperatures was not observed in Illinois, Indiana, or Iowa is not necessarily inconsistent with global warming, because the local effects of global warming on climate and weather are poorly understood.”

Summary:
The perception that trend yields have accelerated in the past decade could be due to benign weather that was not recognized, or due to an improvement in seed genetics, or due to weather patterns that cycle through with beneficial effects and a year to year decrease in yield variation. Precipitation, temperature, and time can explain the increasing yield curve, but trend yield forecasts based on technology may lead to poor yield forecasts.