Estimating soybean yield potential: A web-based app
Many producers would like to estimate the yield potential of their soybeans well before reaching the end of the season. In contrast with corn, soybean can easily compensate for abiotic or biotic stresses. The final number of pods is not determined with finality until close to the end of the season (beginning of seed filling, R5 stage). In corn, the final kernel number is attained during the 2-week period after flowering. Thus, when estimating soybean yield potential, we have to keep in mind that the estimate could change depending on the growth stage at the time the estimate is made and weather conditions. For example, wet periods toward the end of the reproductive period can extend the seed-set period, promoting greater pod production and retention, with larger seed size and heavier seed weight.
From a physiological perspective, the main yield driving forces are: 1) plants per acre, 2) pods per area, 3) seeds per pod, and 4) seed size. Estimating final yield in soybean before harvest can be a very tedious task, but a simplified method can be used for just a basic yield estimate. This method is based on an article by Dr. Shaun Casteel (Soybean Specialist, Purdue University). For details, see:
When can I start making soybean yield estimates?
There is not a precise time, but as the crop approaches the end of the season (R6, full seed or R7, beginning of maturity) the yield estimate will be more accurate. Still, you can start making soybean yield estimates as soon as the end of the R4 stage, full pod (pods are 3/4 inch long on one of the top four nodes), or at the onset of the R5 stage, beginning seed (seeds are 1/8 inch long on one of the top four nodes), knowing that the yield prediction is less precise at those early stages.
Is plant variability within the field an issue in soybean?
Variability between plants relative to the final number of pods and seed size needs to be considered when trying to get an estimation of soybean yields. In addition, variability between areas within the same field needs also to be properly accounted for (e.g. low vs. high areas in the field). Yield estimations should be made in different areas of the field, at least 6 to 12 different areas. It is important to properly recognize and identify the variation within the field, and then take enough samples from the different areas to fairly represent the entire field. Within each sample section, take consecutive plants within the row so as to have a fairly good representation.
Conventional approach to estimating soybean yields
In the conventional approach, soybean yield estimates are based on the following components:
- Total number of pods per acre [number of plants per acre x pods per plant] (1)
- Total number of seeds per pod (2)
- Number of seeds per pound (3)
- Total pounds per bushel, or test weight, which for soybeans is 60 lbs/bu (4)
The final equation for the estimation of the potential soybean yield is:
[(1) x (2) / (3)] / (4) = Soybean yield in bushels/acre
Simplified approach to estimating soybean yields
The main difference between the “conventional” and “simplified” approaches is that the conventional approach uses the total number of plants per acre in its calculation; while in the simplified approach, a constant row length is utilized to represent 1/10,000th area of an acre (Figure 1).
For the simplified approach, sample 21 inches of row length in a single row if the soybean plants are spaced in 30-inch rows; in 2 rows if the row spacing is 15 inches; and in 4 rows if the row spacing is 7.5 inches.
Figure 1. In the “simplified” approach to estimating yields, sample 21 inches row length to equal 1/10,000th of an acre. The number of rows to sample will depend on the row spacing. With 30-inch row spacings, sample one row. With 15-inch row spacings, sample two rows. With 7.5-inch row spacings, sample four rows. Photo by Ignacio Ciampitti, K-State Research and Extension.
This procedure should be repeated in different sections of the field to properly account for the natural field variability.
Driving forces of soybean yield
1) Total number of pods per acre:
The total number of pods (Figure 2) within this constant row length should be counted. After counting all the plants within the 21-inch row sections that represent 1/10,000th of an acre, a final pod number per acre can be estimated. A similar procedure should be used in different areas of the field to get a good overall estimate at the field scale. One good criterion is to only consider pod sizes that are larger than ¾ or1 inch long. Smaller pods can be aborted from this time on in the growing season until harvest.
Figure 2. Total number of pods per plant (only consider the pod sizes larger than ¾ or 1 in). Photo by Ignacio Ciampitti, K-State Research and Extension.
2) Total number of seeds per pod:
Soybean plants will have, on average, 2.5 seeds per pod (ranging from 1 to 4 seeds per pod), primarily regulated by the interaction between the environment and the genotypes (Figure 3). Under severe drought and heat stress, a pessimistic approach would be to consider an average of 1-1.5 seeds per pod. This value is just an approximation of the final number of seeds per pod, and can change from the time the estimate is made until the end of the growing season.
Figure 3. The number of seeds per pod will vary somewhat, depending on the growing environment and genotype. Photo by Ignacio Ciampitti, K-State Research and Extension.
3) Seed Size:
Seed size can range from 2,500 (normal to large seed weight) to 3,500 (small seed size) seeds per pound. This season, conditions are mostly favorable in Kansas for promoting large seed sizes. In more stressful years, such as 2012 and 2011, seed size is normally smaller, meaning a larger number for the seeds per pound (e.g. 3,500 seeds per pound). In the simplified estimation approach published by Dr. Casteel, you do not need to actually measure the number of seeds per pound in order to estimate yields, as is done in the conventional approach. Instead, a seed size conversion factor is used. If the conditions are favorable and large seed size is expected, the conversion is 15 units; while if abiotic or biotic stresses are present during the seed-filling period, a seed size factor of 21 units is used. Further details related to the seed size factor can be found in the link to the Purdue University extension article listed at the end of this article.
Example of the simplified approach for estimating soybean yields:
Let’s say that we have 120,000 plants/acre in a 30-inch rows. Then, we should have around 12 plants in 21 inches of row. In those 12 plants, we have measured on average 22 pods per plant, with a total number of 264 pods (22 x 12).
If we assume a “normal” growing season condition, then the final seeds per pod will be around 2.5, and for the seed size factor, we can assume large seeds, and will use a conversion factor of 15 units.
Equation for a “Favorable” Season:
264 pods x 2.5 seeds per pod / 15 = 44 bushels per acre
For a “droughty” (late reproductive, from R2 to R6 stages) growing season, the final seed number and size will be dramatically affected. Thus, even if the pod number is the same as in a normal season, the yield calculation could be:
Equation for a “Drought” Season:
264 pods x 1.5 seeds per pod / 21 = 19 bushels per acre
This “simplified approach” basically relates the total number of pods in a “known” unit area (easily extrapolated to the acre unit), and is affected by the total number of seeds in the pod. This is adjusted by the estimated seed weight, which is affected by two main components: duration of seed fill and rate of dry mass allocation to the seeds.
New K-State mobile app for estimating soybean yields
If you have an Android device, there is a “free” mobile Web-App that can help estimate soybean yields before harvest. The app is called “KSUSoyYieldCalc.” Complete information can be found at:
The KSUSoyYieldCalc has only four inputs for predicting the final yield:
1. Plant population (plants/acre). This component can be estimated by counting the number of plants in a 21-inch row length for 30” row spacings (1/10,000th area), and by multiplying that number by 10,000;
2. Pods per plant. If the simplified approach is used, this factor can be obtained by counting all pods per plant in the 21-inch row length (total 10 boxes – 10 plants);
3. Seeds per pod. A good average number is 2.5 seeds per pod, but the range presented in this web-based app is from 1 to 4 seeds per pod;
4. Seed size. Seed size typically ranges from 2,500 (large) to 3,500 (small seeds) seeds/lb, with an average of 3,000 seeds/lb.
Inputs 1, 2, and 3 have been already discussed in the “simplified approach” section earlier in this article. Once all these components are estimated in the field, the numbers can be entered into the KSUSoyYieldCalc app.
The last factor “seed size” is the same as the one presented in the “conventional approach.” This factor normally varies from 2,500 to 3,500 seeds/lb. If the conditions until harvest will be favorable, then the “seed size” component should be a lower number (e.g., 2,500 seeds/lb). If conditions are likely to be unfavorable, resulting in a short seed-fill period, then this factor should be higher (e.g., 3,500 seeds/lb). This factor will be ultimately determined as the crop approaches maturity, but an estimation is needed considering the importance of this factor for influencing final soybean yields.
All steps are also highlighted in the below image:
Here is one example of how to use this web-based App:
1. Plant Population: 12 plants (measured at 12 sites within the field) in 21-inch row length x 10,000 = 120,000 plants/acre
2. Pods per plant: 24 pods per plant (average of 12 plants in 21-inch row length)
3. Seeds per pod: 3 seeds per pod (estimation)
4. Seed size: 2,800 (assuming “normal” conditions during seed-fill period)
Final yield estimation: 43 bu/acre
More examples on how to use the App and estimate yields are presented in the below figure:
You can see a video explaining the app and download the app from the Google Play link:
For more information on how to estimate soybean yields, check the following resources:
University of Kentucky
Ignacio Ciampitti, Crop Production and Cropping Systems Specialist