Issue No. 7, Article 7/May 8, 2009
Have You Considered Reducing Your Soybean Seeding Rate This Year?
Every year soybean producers make major decisions about the planting process. The first decision is variety selection (see issue 4 of the Bulletin, April 17, 2009.) Second is planting date (issue 6, May 1, 2009), which is taking on added importance this year as corn planting progress remains slow. A third important decision is seeding rate. Seeding rate has become more economically critical as soybean seed prices have increased with improved varieties, the addition of biotechnology traits, and the increasing popularity of more costly soybean seed treatments.
Soybean seeding rates in Illinois typically range from 150% to 200% of the number of plants needed at harvest (75,000 to 100,000 plants per acre) to maximize yield under most conditions. High seeding rates are used to provide insurance against conditions that reduce soybean emergence. At the same time, soybean plants compensate well to low stands by adjusting yield components (pod and seed number per plant) and expanding plant growth for more complete canopy closure to suppress weed competition. Historically, the cost of soybean seed has been a relatively minor expense in the cropping operation, so it made sense to plant far more seeds than the number of plants needed. Today's higher seed costs have increased interest in lowering seeding rates in order to maximize economic return.
Several of my colleagues in surrounding states have looked at the relationship between soybean seeding rates and yield and have promoted lowering seeding rate recommendations in their states. In the past decade, there have been a couple of large multiyear studies investigating desired soybean seeding rates in Illinois, and I was privileged to get these data to analyze.
I will discuss two different studies here. Before I do, I want to give credit--for the first study to the University of Illinois Department of Crop Sciences and research farm superintendents Eric Adee (Monmouth), Lyle Paul (DeKalb), and Bob Dunker (Urbana), and for the second study to the Department of Crop Sciences Variety Testing Program and Emerson Nafziger. Also, let me make clear that I will be discussing soybean seeding rates and not soybean plant populations. Data related to plant populations are also critical, but it is the seeding rate that you have the most control of. Second, the data from all of the field trials I will be discussing were from "normal" conditions, and general plant population counts in these trials would indicate that plant stands were uniform and generally between 85% and 95% of the number of seeds planted.
My first thought about data from these trials was that even though we know soybeans can yield very well with plant populations of less than 100,000 plants per acre, the risk of reduced yields would be higher when starting with lower seeding rates. As I analyzed the data I paid particular attention to the magnitude of changes in the group variances, not just to the changes in mean yield while accounting for the group variances. From these two datasets, a reduction in variance did not occur in relation to seeding rates. Therefore, reducing risk by increasing seeding rate does not intuitively appear to be logical, but I will discuss this further.
The first study examined 125,000, 175,000, and 225,000 seeds per acre planted in row spacings of 7.5, 15, and 30 inches at Monmouth, DeKalb, and Urbana, Illinois, in 1998 through 2000. There was no yield difference between the 7.5-inch rows (53.6 bushels per acre) and 15-inch rows (53.0 bushels per acre); however, 30-inch rows yielded 2.1 and 2.7 bushels per acre less than the narrower row spacings, respectively (Figure 1). An increase in seeding rate produced a modest, albeit significant, increase in yield. The rate of 125,000 seeds per acre yielded 51.6 bushels per acre. An increase of 50,000 seeds per acre increased yield 1.2 bushels per acre, while increasing another 50,000 seeds per acre (175,000 to 225,000) increased yield only 0.2 bushels per acre (Figure 2).
Figure 1. Soybean yield data from drilled (7.5-inch), 15-inch, and 30-inch row spacings pooled over 125,000, 175,000, and 225,000 seeds per acre planted at Monmouth, DeKalb, and Urbana, Illinois, in 1998 through 2000.
Figure 2. Soybean yield data from three seeding rates pooled over drilled, 15-inch, and 30-inch row spacings planted at Monmouth, DeKalb, and Urbana, Illinois, in 1998 through 2000.
In the second study, yields from 50,000 seeds per acre averaged 55.8 bushels per acre; successive increments of 50,000 seeds per acre increased yields by 5.2, 1.1, and 0.3 bushels per acre, again showing a rapidly diminishing return to added seed (Figure 3).
Figure 3. Soybean yield data from four seeding rates examined at 33 field locations throughout Illinois from 2005 to 2008.
Remember, however, that before I made many conclusions from these data I also wanted to look at the change in magnitude of variance as an indication that risk could be decreased as seeding rate increased. I hypothesized that average yields would be more stable when a higher seeding rate was used. By visual inspection of the data range (high yields minus low yields) in both studies (Figures 2 and 3), we see the variation was similar. The coefficient of variation (CV), which is the standard deviation divided by the group mean multiplied by 100, ranged from 10% to 14% for these studies and was relatively constant (Figures 4 and 5). Variance between the seeding rates in both trials was thus uniform. This means that under normal conditions, you are just as likely to end up with below-average yields when you plant at higher seeding rates. I guess a higher rate may still "save" you under catastrophic conditions that greatly and selectively reduce plant stands--soil crusting, frost, bird feeding, hail (to name a few)--but as seed costs increase, so does the cost of protection against these often-isolated occurrences. So the question is how much you are willing to pay for insurance against isolated stand reduction events.
Figure 4. Coefficient of variation from soybean yield data from three seeding rates pooled over drilled, 15-inch, and 30-inch rows examined at Monmouth, DeKalb, and Urbana, Illinois, field locations from 1998 to 2000 (see Figure 2). Coefficient of variation is the standard deviation divided by the group mean multiplied by 100.
Figure 5. Coefficient of variation from soybean yield data from four seeding rates examined at 33 field locations throughout Illinois from 2005 to 2008 (see Figure 3). Coefficient of variation is the standard deviation divided by the group mean multiplied by 100.
Economic considerations. Data from these trials would indicate that effects of changing seeding rates occur uniformly regardless of the soybean row widths of your operation. In the first trial, if you were paying 25 cents per 1,000 seeds ($35 for a bag with 140,000 seeds), the increase of 1.2 bushels per acre between 125,000 and 175,000 would probably have been just enough to break even with modest economic gains or losses, depending on the value of your soybean crop. However, an increase from 175,000 to 225,000 seeds per acre would have been economically negative, with possible losses of $10 an acre with a soybean price of $10 per bushel. The second study, which was conducted a little more recently, had very similar yield response for the addition of 50,000 seeds per acre between 100,000 and 150,000, but economic losses would have occurred above 150,000.
Data from the second study were also fit to a regression model to calculate the economic optimum seeding rate. To account for extremes, the model was used to estimate the economically optimum seeding rates for a range of seed input costs (5 cents to 65 cents per 1,000 seeds and soybeans at $7 to $16 a bushel). Fifteen years ago seed input costs were close to the price of 10 cents per 1,000 seeds. In that price range, the optimum seeding rate fluctuated under 4,000 seeds per acre from an increasing soybean value of $7 to $16 a bushel. Today, however, with various biotechnology trait fees and seed treatment options, the value of seed input is much wider and likely between 15 and 45 cents per 1,000 seeds for most farmers. Economically optimum seeding rates fluctuated by 26,000 seeds per acre in that seed price range. From this research trial, optimum rates in the range of 25 to 45 cents per 1,000 seeds would have been between 100,000 and 129,000 seeds per acre. This is much lower than current recommendations for 30-inch rows in Illinois, and a little lower than I would be comfortable recommending.
More work may be needed to fine-tune these recommendations, but I present the data as an indication that seed input costs have a much larger impact on optimum seeding rates today than they did over a decade ago. I would suggest that most operations may benefit from modest decreases in seeding rates, provided the seed used has good viability and is planted under normal conditions. Really high seeding rates are likely to provide a negative return on your seed investment.
Seed costs are likely to continue to increase, so seeding rates may need continuing adjustments to reflect the price of both seed and soybeans. If you currently plant at much higher rates than were found to be optimal, you might as a first step lower your rates some to levels you still feel comfortable with. Another suggestion would be to plant as you intended but to also plant some test strips with lower seeding rates and see how harvest plant populations and yields compare on your farm. This would be very easy for those with yield monitors, but remember that the more strips separated in space with your normal seeding rate (a creation of replications), the better.--Vince M. Davis