A review and reassessment of the earlier presented information relative to the Louis M. Thompson, PhD study of corn and soybean yields.
http://ageconsearch.umn.edu/bitstream/37501/1/morr_08-01.pdf
Weather, Technology, and Corn and Soybean Yields in
the U.S. Corn Belt
by
Michael A. Tannura, Scott H. Irwin, and Darrel L. Good
Summary:
The purpose of this section is to review how the expected relationship between weather, technology, and corn and soybean yields has changed relative to previous publications by Thompson. Direct comparison of Thompson’s results and this study’s results was not possible because of specification differences. For example, the original model by Thompson (1963) used quadratics for all weather variables, so it was not possible to make direct comparisons of the expected effect of all precipitation and temperature variables on yields. In later studies (e.g., Thompson 1969 1970 1985 1986 1988), the definition of pre-season precipitation was changed from September-May to September-June, eliminating June precipitation as a stand-alone variable. The following discussion provides a qualitative rather than quantitative comparison to Thompson’s estimation results (1962 1963 1969 1970 1985 1986 1988) due to differences in model specification and variable forms.
Thompson (1963) noted that, “The most significant weather variables in the production of corn and soybeans are July rainfall and August temperature” (p. 49) In particular, above average rainfall in July and lower-than-average temperature in August was ideal for production, while August rainfall had a larger influence on soybeans than corn. He furthered noted that rainfall during August was “not significant” for corn production, though above-average rainfall in July and August was the best for high soybean production (Thompson 1969 1970). Both crops were also concluded to have highest yields with near-average June temperature and below average July and August temperatures. Subsequent studies (Thompson 1985 1986) showed that soybean yields generally produced highest yields with near-average temperatures during June, July, and August and above-average July and August rainfall, while the effect of weather on corn yields was similar to previous findings (Thompson 1969).
Thompson (1988) noted that highest corn yields were associated with: 1) above-average July rainfall, 2) below-average temperatures in July and August (though August could be too cool), 3) average pre-season precipitation (September through June), 4) slightly above-average June temperature, and 4) slightly above-average August rainfall (though August could be too wet). Highest soybean yields were associated with: 1) above-average July and August rainfall (especially important in August), 2) slightly above-average June and July temperatures, 3) average August temperature, and 4) average pre-season precipitation. This showed that corn was “… more sensitive than soybeans to high temperature in August as well as in July“ (p. 23), while August could be too cool and wet for ideal corn production.
Results from the regression models estimated for this study were generally similar to results in the aforementioned Thompson publications. Both the current and previous studies showed that corn yields were strongly influenced by the magnitude of July precipitation and July and August temperatures. However, the current study indicated that June precipitation had a larger influence on corn yields than in Thompson’s studies. As noted above, all but the first two of Thompson’s publications defined pre-season precipitation as September-June precipitation, as opposed to September through April in the current study. Therefore, inclusion of the June variable showed that the magnitude of June precipitation is more important than indicated previously – especially if the month is considerably drier than average because it can quickly reduce potential corn yields.
Additionally, Figure 47 suggests the effect of May and June temperature was minimal, as it appeared that corn can recover from unusually warm or cool conditions during its early developmental phases. Above-average pre-season precipitation was also suggested by the modified models to increase yield potential, but the effect of weather later in the growing season considerably outweighed the magnitude of precipitation prior to May.
Several differences were also found to exist with regard to relationships between weather and soybean yields. As also noted with respect to corn, June precipitation is more important to soybean yield potential than previously indicated. Although the amount of August and July precipitation is particularly important – especially during August – the magnitude of July precipitation is also high. This suggests that above-average precipitation throughout June, July, and August is best for soybean yields – as opposed to only July and August. The main difference is that below-average temperatures in July and August appear to be the most ideal for soybean yields, although Thompson (1988) had suggested slightly above-average temperatures were the most ideal during this period.
However, the models agreed that above-average warmth in June may be the most ideal for soybeans. Therefore, the magnitude of July and August temperatures may be more closely related to soybean yields than previously suggested, although the effect of precipitation is clearly more important.
