Now Reading: Will El Niño Affect Midwest Corn and Soybean Yield?

Will El Niño Affect Midwest Corn and Soybean Yield?

by The Climatology Team

May 4, 2016

Share:

In a previous post we showed that despite the hype, there isn’t much evidence that El Niño has a large impact on springtime precipitation in the Midwest. The major weather impact to a grower’s operations will be the shift in timing of planting. But what about its impact on yield? Should corn and soybean growers expect a smaller harvest?

This year’s El Niño was one of the strongest events on record and to better understand the possible impact on agricultural production in the Midwest, a team of Climate scientists explored El Niño’s potential impact on corn and soy harvests in Illinois, Iowa and Indiana.

What we found

Based on a detailed analysis of yield data spanning more than 60 years, we found little to no impact on crop yields. In fact, of the five strong El Niño events we studied we found yields to be on average or slightly higher than expected, with the only lower than expected yields occurring during a major drought in 1983. That year yields were about 30 bushels per acre below average – still less than the recent 50 bushel per acre average loss during the 2012 drought. While media may hype severe weather in other parts of the US or the shifts in large-scale weather patterns, farmers in the Midwest should expect to experience an average harvest, outside of unforeseen severe weather events or plant-related challenges.

Being able to understand and plan for potential impacts from El Niño or other weather-related events is key to managing your operations and minimizing risk. Climate’s team of atmospheric scientists working closely with statisticians are able to combine science and data to help drive decision making. Climate FieldViewTM Pro uses this sophisticated weather modeling to help farmers plan and make more informed decisions throughout the growing season.

About the science

For this data analysis, we used historical state-average corn and soybean yield for Illinois, Iowa, and Indiana, collected from the USDA National Agricultural Statistical Service (NASS). We have a longer record for corn, going back to before 1900. Looking at the time series in Figure 1, we can see a positive trend in yields starting in the late 1930s, largely due to a powerful combination of new hybrid development, better breeding methods, fertilizers, and new statistical techniques to evaluate hybrid performance.

Figure 1. Historical corn and soybean yield data for Illinois, Iowa, and Indiana. The data show an increasing trend from the late 1930s to present day. The data were collected from the National Agricultural Statistical Service.

Figure 1. Historical corn and soybean yield data for Illinois, Iowa, and Indiana. The data show an increasing trend from the late 1930s to present day. The data were collected from the National Agricultural Statistical Service.

Since our sample size dates back to 1950, we removed the data prior to 1950 shown in Fig. 1 and remove the positive trend in corn and soybean yield (Figure 2 shows the detrended data). Removing the positive trend allows us to focus on the impact of weather rather than the increased bushels/acre each year, allowing us to identify years in which yield deviated from what we’d expect based on long-term trends.

From 1950 to present, Iowa gained approximately 2 bushels/acre per year. Since Iowa averaged 50 bushels/acre in 1950, we’d expect that 43 years later in 1993, Iowan farmers would grow about 86 more bushels/acre than in 1950, or 136 bushels/acre. However, massive flooding in 1993 left Iowan farmers with yields closer to 80 bushels/acre. The difference between the expected and actual yields in 1993 is largely attributed to weather.

Now that the trend has been removed, you can more easily identify the effects of the significant drought and heat waves of 1988, 2003, and 2012, as well as other years that deviate from the expected trend (like 1993 in Iowa). Specifically, we would like to see if exceptionally good and bad years (in terms of yield) happen to correspond to the occurrence of El Niño or La Niña events in the Equatorial Pacific.

Figure 2. Time series showing the corn and soybean yield in Illinois, Iowa, and Indiana, after removing a trend in the data. We only considered data collected after 1950 for this analysis.

Figure 2. Time series showing the corn and soybean yield in Illinois, Iowa, and Indiana, after removing a trend in the data. We only considered data collected after 1950 for this analysis.

After 1950, there are only five El Niño events considered strong (2.7 degrees Fahrenheit above the average) or very strong (3.6 degrees Fahrenheit above the average): 1958, 1966, 1973, 1983, and 1998. Further, there are only three strong (2.7 degrees Fahrenheit below the average) La Niña events over this time period: 1974, 1976, and 1989. We compare the average yield in these years to all other years from 1950 onward. Figure 3 displays average corn yield (left) and soybean yield (right) across Illinois, Iowa, and Indiana, for El Niño years, La Niña years, and all years that were not classified as El Niño or La Niña (neutral).

Figure 3. A strip plot showing the distribution of average yield in Illinois, Iowa, and Indiana, for corn (left) and soybeans (right).

Figure 3. A strip plot showing the distribution of average yield in Illinois, Iowa, and Indiana, for corn (left) and soybeans (right).

This figure illustrates how difficult it is to draw firm conclusions from a very small amount of data. For instance, we can see that soybean yield was below average in all three La Niña years. However, there are two neutral years that showed lower soybean yield anomalies than the three La Niña years. In fact, it appears that the observed yield in the El Niño and La Niña years could reasonably be expected in a neutral year.

About the authors: Our Climatology team is a group of atmospheric scientists and statisticians who construct agronomically-relevant weather datasets. These data play an important role in the development of our weather-dependent agronomic models, for example, Nitrogen Advisor offered in FieldViewTM Pro.

Share: