Soybean seed quality and composition is receiving increased attention among farmers, agronomists, and commodity traders. An improved nutritional composition of US (United States) soybean would provide a competitive edge that can be widely exploited to produce increased economic value through the targeted marketing of each bushel. Understanding of within-field soybean quality will enable and encourage value chain disruption both up and downstream via more leveraged negotiations.
A recent survey of farmers across the US Midwest (2020-2021) led by Dr. Ciampitti at Kansas State University related to their perception of soybean quality (n = 271 responses) revealed that growers are willing to invest in the technology for spatially characterizing within-field protein variation when they can obtain a worthwhile premium for protein (Agronomy Journal, https://doi.org/10.1002/agj2.21082). Almost all, 98%, of respondents highlighted that protein levels are relevant for them, with more than 55% replying that an increase of $0.50 per bushel is enough to justify investment in this technology. In addition, more than 75% of farmers (n>180) responding to this survey are interested in learning more about quality and how to manage it spatially. Creating awareness and generating relevant data about soybean quality variation in farmer fields will support the goal of increasing the competitiveness of US farmers in both global and regional niche markets. This work will expand knowledge of soybean protein and oil quality in farmer fields in 10 states across the US North Central soybean region. This large database will be incorporated into a decision support tool developed with funding requested from USB for mapping soybean quality across the US. A simplified framework is presented below. This project will provide the base level of ground-truth field data required (Fig. 1) for integration with remote sensing provided by multi-spectral satellite imagery to build predictive models that will provide new insights derived from soybean quality spatial variation. In this regard, in a scientific article under review (Comput. Electron. Agric.), we propose a framework to predict seed protein and oil concentration using satellite data that could help farmers for in-site seed segregation at harvest.

The second year of the project has been established with a total of 10 states across the NC region collaborating to develop one of the largest databases for soybean quality globally. The target for year #2 is set to sample soybeans from more than 100 farmer fields with a similar goal for year #3 (+100 fields). In addition to soybean seed quality analysis, this project will retrieve relevant management data from farmers to refine future investigations focused on improving soybean quality for farmers across the country. The link to the field management data collection (a few captures shown below, Fig. 2) is presented here: https://forms.gle/5wBfdj9ZhsoJYsbNA

1. Continue the development of a multi-state database to allow upscaling of soybean quality predictions to regional levels and benchmark agronomic practices, soybean genetics, and environmental conditions that can lead to large-scale improvements in soybean quality.

For the development of a multi-state database in collaboration with other soybean producing states, a protocol for data collection has been developed to obtain the field management information and soybean seed samples from farmer fields that are necessary to identify overall within-field spatial variation of quality. A total of 10 states from the North Central region (excluding WI, MN, and PA) will be involved in this project.
The project will be focused mainly on the collection of soybean seed samples from farmer fields for data quality analyses and corresponding field management information. A complementary project, submitted for a third year of funding from USB, will assist this initiative providing geo-referenced aerial imagery of soybean fields during the growing season. On average, ten soybean fields in each state will be part of this large research project to collect soybean seed samples for quality analyses. A protocol for field seed quality data collection has already been developed (Fig. 3) to obtain within field variation for this relevant seed trait, in addition to connecting this determination with relevant farmer field management (Fig. 2). Fields will be scouted during the season and seed samples (~1-2 lb.) collected before harvest will be analyzed by the Kansas State University Lab.

2. Communicate the economic value of soybean quality mapping to farmers and agronomists through an online interactive simulation tool, accessible publications, and social media.

To communicate the economic value of soybean quality mapping, the second main activity will incorporate the data collected into an online tool “Soybean Quality Differentiation and Economic Simulator” (a beta version, https://analytics.iasoybeans.com/cool-apps/SoybeanQualityPremium/).
The second activity will be comprised of the following sub-activities:
- Utilize the collected information to enhance the tool to enable farmers to select soybean management practices, price points, production costs, cost savings, transportation cost to destinations that provide a soybean quality premium, among other customizable features (side-by-side comparison) that will likely produce the best economic scenario when using spatial crop quality information for soybean marketing.
- Our team will promote the new simulator and expand the reach of this tool. The target audience for our effort will be soybean farmers, the ag technology industry, seed companies, soybean processors and soybean meal consumers.
- Iowa Soybean Association Communication and Marketing departments will help promote the new simulator and expand the reach and influence of this tool. In addition, Kansas State Research and Extension personnel will help to disseminate the use of the web tool and any other relevant information for farmers across the state, and with the PIs at the regional scale.
The first version of the economic simulator allows users to estimate the optimal combination of soybean price and protein premiums while considering possible yield change values. Updates to the tool have included the option to add additional costs associated with planting a higher quality variety, and savings that could be provided by a seed company, soybean crusher or food processor. For example, the tool allows adjustments for the cost of transportation if the soybean is hauled to an alternative processor or elevator, or storage is provided by a seed company. Currently, the tool only simulates the economics of protein premium payments but the option to simulate the same for oil premium payments will be added, specifically high-oleic options such SOYLEIC and others.

Utilizing the database of soybean quality samples, predictive models will be generated to develop a “Spatial Soybean Quality Differentiation Index” to help quantify the economic feasibility of specific field locations to benefit from soybean quality differentiation segregation and marketing. Identifying a field location in the tool could employ the predictive power of such an index to estimate the probability that a given field shows potential for economic benefit from differentiation given the proper market conditions. This would enable translation from simulation into practice, by the user employing the tool to then quantify, simulate, and estimate the potential return per acre or per farm based on estimations of field level soybean protein and oil data produced by the index in conjunction with selected management and price/cost options available within the tool.

The project duration is 3 years, the funding requested now is for year #3.

Years #1-3 involved data collection from hundreds farmer fields, with the final year of the project to finish developing the final database on-farm soy quality. Fall and Winter 2023-2024 will be utilized to identify potential farmers for this project.
Reports will be submitted according to NCSRP (North Central Soybean Research Program) guidelines.

Update:
The team of all the collaborators from multiple states (Ohio, Indiana, South Dakota, Missouri, Iowa, Michigan, Illinois, North Dakota, Nebraska, Iowa, and Kansas), including John Fulton, Shaun Casteel, Peter Kovacs, Andre Borja Reis, Scott Nelson, Mark Seamon and Mani Sing, Randy Pearson, David Kramar and Michael Ostlie, and Guillermo Balboa, helped on collecting all field sites for 2023 growing season. All seeds were processed for seed quality traits, mainly protein and oil concentrations, from all fields were obtained and data share across all collaborators.

For the last two growing seasons, 2022-2023, a total of 234 fields with complete data on soybean seed quality and relevant crop management has been collected and compiled across the US soybean producing region. The states in the southern part of the US (Louisiana, Mississippi, and Alabama) were collected via a grant provided by the USB. The rest of the states are all the ones included in the current project funded by NCSRP.

The soybean quality economic simulator has been updated and modified in two key areas. The first being that the oil quality portion of the tool has been built and is functioning well. The second update was to the existing user interface to make it more intuitive for users. Based on feedback from farmers the old version was difficult to understand what yield was used and how to add yield loss properly. We also added a break-even premium price so that farmers can quickly decide on if the premium they are receiving will have a positive ROI on their farm. The last addition we made is that users can export their results in a pdf, csv, or XLSX document so they can save their work for future use.

The next steps are to develop a first database and publication of the within-field variation for soybean seed quality, protein, and oil, for farmer fields across the US. A new growing season, 2024, will be added to complete the entire data collection. For more information, please see the attached report.

View uploaded report

Updated January 29, 2025:
The team of all the collaborators from multiple states (Ohio, Indiana, South Dakota, Missouri, Iowa, Michigan, Illinois, North Dakota, Nebraska, Iowa, and Kansas), including John Fulton, Shaun Casteel, Peter Kovacs, Andre Borja Reis, Scott Nelson, Mark Seamon and Mani Singh, Randy Pearson, David Kramar and Michael Ostlie, and Guillermo Balboa, helped on collecting all field sites for 2022, 2023, and 2024 growing seasons.

All seeds were processed for seed quality traits, mainly protein and oil concentrations, from all fields were obtained and data share across all collaborators. Reports for each state were prepared every year to provide information on the soybean seed quality (mainly protein and oil) for each farmer field.

For the last three growing seasons, 2022-2023-2024, a total of 394 fields with complete data on soybean seed quality and relevant crop management has been collected and compiled across the US soybean producing region. The states in the southern part of the US (Louisiana, Mississippi, and Alabama) were collected via a grant provided by the United Soybean Board (USB). The rest of the states are all the ones included in the current project funded by NCSRP.

Most recently, a manuscript was prepared to summarize all the information collected on this project. As mentioned before, we followed a standardized protocol across all farms for collecting representative seed and soil samples in-situ for analysis. In addition, we retrieved relevant crop management and yield data via surveys and linked all datasets with seasonal weather variables to develop a large on-farm database. The main objectives of the paper were to i) assess the importance of environmental variables in predicting seed oil and protein concentration and reported yields, ii) identify regions related to yield and seed quality, and iii) explore key predictors linked to these variables across regions to further understand seed oil and protein concentration differences across defined geographical regions.

We collected yield and management data via survey, and combined growth models to summarize weather data during key crop phenological stages. The prediction of yield and oil concentration exhibited greater accuracy than that of protein concentration when seasonal variables related to weather, soil, and crop growth were considered. Yield, protein, and oil levels were within the ranges usually reported for soybean in the regions explored. However, higher protein levels in the north suggest a narrowing in the quality gap of soybeans between this region and the Corn Belt.

We are building a tool for predicting quality (oil and protein) as a next step. In addition, the soybean quality economic simulator has been updated and modified in two key areas. The first being that the oil quality portion of the tool has been built and is functioning well. The second update was to the existing user interface to make it more intuitive for users. Based on feedback from farmers the old version was difficult to understand what yield was used and how to add yield loss properly. We also added a break-even premium price so that farmers can quickly decide if the premium they are receiving will have a positive ROI on their farm. Users of the tool are also able to add any additional costs or savings that are a result from growing soybeans with an associated premium for oil and protein, such as increased seed and planting costs, or reduced financing options for inputs. This functionality can be accessed through selecting additional costs and incentives, and is captured as an aggregate sum in the downloaded pdf of added costs and incentives. Adding this utility allows farmers and other users to estimate total financial gain or loss of implementing a practice and can help in ensuring that all costs are estimated before implementing a new practice on their farm.

View uploaded report

The team of all the collaborators from multiple states (Ohio, Indiana, South Dakota, Missouri, Iowa, Michigan, Illinois, North Dakota, Nebraska, Iowa, and Kansas), including John Fulton, Shaun Casteel, Peter Kovacs, Andre Borja Reis, Scott Nelson, Mark Seamon and Mani Singh, Randy Pearson, David Kramar and Michael Ostlie, and Guillermo Balboa, helped on collecting all field sites for 2022, 2023, and 2024 growing seasons.

For the last three growing seasons, 2022-2023-2024, a total of 394 fields with complete data on soybean seed quality and relevant crop management has been collected and compiled across the US soybean producing region. All seeds were processed for seed quality traits, mainly protein and oil concentrations, from all fields were obtained and data shared across all collaborators. Reports for each state were prepared every year to provide information on the soybean seed quality (mainly protein and oil) for each field.

We collected yield and management data via survey, and combined growth models to summarize weather data during key crop phenological stages. The prediction of yield and oil concentration exhibited greater accuracy than that of protein concentration when seasonal variables related to weather, soil, and crop growth were considered. Yield, protein, and oil levels were within the ranges usually reported for soybean in the regions explored. However, higher protein levels in the north suggest a narrowing in the quality gap of soybeans between this region and the Corn Belt.

Lastly, we are building a tool for predicting quality (oil and protein) as a next step. In addition, the soybean quality economic simulator has been updated and modified in two key areas. The first being that the oil quality portion of the tool has been built and is functioning well. The second update was to the existing user interface to make it more intuitive for users. Based on feedback from farmers, the old version was difficult to understand what yield was used and how to add yield loss properly. We also added a break-even premium price so that farmers can quickly decide if the premium they are receiving will have a positive ROI on their farm. Users of the tool are also able to add any additional costs or savings that are a result from growing soybeans with an associated premium for oil and protein, such as increased seed and planting costs, or reduced financing options for inputs. This functionality can be accessed through selecting additional costs and incentives, and is captured as an aggregate sum in the downloaded pdf of added costs and incentives. Adding this utility allows farmers and other users to estimate total financial gain or loss of implementing a practice and can help in ensuring that all costs are estimated before implementing a new practice on their farm.

At the end of this project, the team expects to have the largest dataset on the within-field variation soybean quality at farmer scale across the US North Central Region. This project is important and timely since it will provide relevant information to growers related to segregate quality at the field level, with the ultimate outcome of improving overall profits from the current soybean farming systems.