With most Kansas soybean fields already in reproductive stages, it is time to assess the yield potential. The latest USDA-NASS Kansas Crop Progress and Condition report (Aug. 18) estimates that 87% of soybeans in Kansas are blooming. Around 69% of the soybeans in the state are setting pods, which is on pace with last year and near the 5-year average of 67%.
Soybean plants can more easily compensate for stress (relative to corn), especially if the timing of the stress is before seed filling. Soybeans can compensate because of overlapping vegetative and reproductive phases due to their indeterminate nature (i.e., the plant continues to produce new nodes and leaves at the top of the plant even when flowering has begun at the bottom of the plant). Thus, flowering in soybeans can normally last between 4-6 weeks, allowing for the replacement of flowers and small pods that may have been aborted during heat or drought stress.
Since the final number of pods is not determined until the beginning of seed filling, when estimating soybean yield potential, we must remember that the estimate could change based on the timing of this estimation as well as 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 heavier seed weight.
Estimating the final yield before harvest can be a very tedious task, but a simplified method using yield components can be applied to start setting yield expectations.
From a physiological perspective, the main yield components to consider are:
- Plants per acre
- Pods per area
- Seeds per pod
- Seed size
When can I start making soybean yield estimates?
There is no precise time, but we can start making soybean yield estimates as soon as the end of the R4 stage, full pod (pods are ¾-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). Keep in mind that yield prediction is less precise at these early reproductive stages since the seed number per area, as well as the seed weight, are not yet completely defined. At this early stage of seed development, it is important to only consider the pods that are at least ¾-inch long to avoid over-optimistic estimations since smaller pods can still abort under stressful conditions.
As the crop moves into the R6 stage (full seed), the seed number (main yield component) is almost fully defined. However, the conditions during seed filling will determine the final seed number as well as the size and weight of the seed. The closer to maturity (R7 stage) we make the estimation, the more accurate the expectation and overall yield prediction.
Figure 1. Soybean phenological stages to start yield prediction using the yield components method.
How many samples are needed to account for field variation?
Perform yield estimations in at least 6 and ideally 12 different areas of the field. Variability between areas within the same field also needs to be properly accounted for. Make sure to collect enough samples from the different areas to fairly represent the entire field. Within each sample section, take consecutive plants within the row to have a good representation. The variability between plants in terms of the number of pods and seed size needs to be considered when trying to get an estimation of soybean yield. The more variability to represent, the more samples needed for a proper estimation.
The yield components equation
Soybean yield estimates following the conventional approach are based on the following components (Eq. 1):
Eq. 1
where,
- Plants/ac. (Figure 2, step 1)
- A simplified approach can be applied by using samples covering 1/10,000th of an acre. With 30-inch row widths, you would need 21 inches of row length. Thus, the average number of plants in several sections multiplied by 10,000 will give us an estimation of the number of plants per acre.
- If the soybean plants are arranged in 30-inch rows, we just need to sample a single row; 2 rows if the row spacing is 15 inches; and 4 rows if the row spacing is 7.5 inches.
- Pods/plant. (Figure 2, step 2).
- Count and average the number of pods per plant from the sample.
- Seeds/pod. (Figure 2, step 3)
- Count and average the number of seeds per pod. Soybean plants have, on average, 2.5 seeds per pod (ranging from 1 to 4 seeds per pod).
- Under severe drought and heat stress, a pessimistic expectation would be an average of 1-1.5 seeds per pod.
- Seeds/acre. (Figure 2, step 4)
- Calculate the total number of plants/acre (1), pods/plant (2), and seeds/pod (3) to obtain the total number of seeds/acre .
- Pounds/bu.
- For this estimation, you could use the average test weight for soybeans of 60 lbs./bu.
- Seeds/lbs. (Figure 2, step 5)
- The number of seeds per pound will vary depending on the seed-filling conditions, which will determine the seed size. Normally, this number could range somewhere between 2,500 (bigger seeds) to 3,500 (smaller seeds) seeds per pound for optimal to unfavorable seed filling conditions.
- Combined with a constant test weight of 60 lbs./bu, you could see a range of seeds per bushel between 150,000 seeds per bushel to 210,000 seeds per bushel.
The final step to estimating yield is dividing seeds/acre by seeds/bu to obtain the yield estimation in bushels per acre (Eq. 2).
Eq. 2
Figure 2. Example of yield estimation method using samples of 1/10,000th of an acre (21-inch x 30-inch sections) for a regular yield until seed filling, where the seed size is expected to be reduced compared to favorable seed filling conditions, increasing the #seeds/bu component.
Example representing average conditions
- Using the steps, we counted the number of plants in 10 sections of 30x21 inches and determined an average number of 12.4 plants. Since these sections are 1/10,000th of an acre, our first component, the plant density, is 124,000 plants/acre.
- Then, in those 12.4 plants, we measured, on average, 27 pods per plant. Assuming a “normal” growing season condition, we could expect to count around 2.5 seeds per pod.
- Combining these components, we obtained an expectation of 8.37 million seeds per acre.
- Finally, we have to assume or estimate the seed-filling conditions. In this case, we used regular-to-poor seed filling due to a lack of precipitation combined with heat. Thus, we divided the seeds per acre by an expectation of 200,000 seeds per bu (small seed size), giving us an estimate of roughly 42 bu/acre.
Tina Sullivan, Northeast Area Agronomist
tsullivan@ksu.edu