Jason Ward

The cost of soybean production is increasing. Nutrient inputs are at historically high costs at a time where there is increased variability in growing season conditions. Crops can experience multiple kinds of stresses and better information is needed on strategies that allow the crop to perform as well as possible while using a minimum of inputs. In North Carolina three of the most substantial production costs are seed, nutrient inputs, and water management infrastructure. This project further develops relationships between water stresses and nutrient uptake so that the response to stress conditions will protect the variety���s yield potential, optimize nutrients, and suggest water management approaches ��� addressing all the major cost centers. The larger outcome of this research and extension effort is in-season response to water challenges such that nutrient input may be reduced, applications timed to environmental stress and profitability improved. The ability to utilize UAV data to assess plant water and nutrient stress in-season as it occurs and ultimately use the data to prescribe an appropriate course of action is necessary to protect soybean crop profits. This process will provide a framework to quickly assess areas in the field and allow for corrective measures to protect the crop from climatic conditions that ultimately would cause severe yield reductions.

Precision agriculture technology is being adopted in many crop production enterprises to optimize crop yields, minimize their associated costs and reduce environmental impacts of excessive crop inputs. Precision agriculture technology enhances management through variable-rate application of lime, fertilizers, pesticides, seeds and tillage. The management protocol is based on determining soil variability by georeferenced soil sampling/mapping, determining yield potential by georeferenced yield monitoring, and using agronomic recommendations for variable-rate application of inputs. Technologies available include georeferenced soil sampling, on the go crop sensors, variable rate control systems, and yield monitors.

The cost of corn production is increasing. Nutrient inputs are at historically high costs at a time where there is increased variability in growing season conditions. Crops can experience multiple kinds of stresses and better information is needed on strategies that allow the crop to perform as well as possible while using a minimum of inputs. In North Carolina three of the most substantial production costs are seed, nutrient inputs, and water management infrastructure. This project further develops relationships between water stresses and nutrient uptake so that the response to stress conditions will protect the variety��������s yield potential, optimize nutrients, and suggest manage water approaches �������� addressing all the major cost centers. The larger outcome of this research and extension effort is in-season response to water challenges such that nutrient input may be reduced and profitability improved.

Modern cotton harvesting equipment has the capability to weigh, estimate moisture content of, and discretely identify round modules when completed. These data along with bale location should allow tracking of gin fiber quality data all the way back to a field area via the permanent bale indicator (PBI) assigned at the gin and the radio-frequency Harvest ID (HID) tag assigned at harvest. The ultimate purpose of this project is a proof-of-concept linking fiber quality data to specific field locations within a cotton field. The objectives of this research are: 1. To identify data workflows and best management practices to connect USDA fiber quality data to specific round modules, 2. to map fiber quality to spatial locations within production fields, and 3. to better understand logistics management in modern cotton harvest such as in-field transport and prioritizing high-moisture round modules.

Unmanned Aircraft (drones) are marketed in the agricultural sector as a ���������������revolutionary������������������ technology. Although the technology and corresponding data are truly unique, the application of data outputs for agricultural management decisions (e.g., re-plant, pest management) remain unclear. This interdisciplinary project will investigate the use of drones in five key production areas 1) re-plant decisions, 2) incidence of fungal disease, 3) severity of insect-related defoliation, 4) weed identification and management, and 5) nutrient deficiencies. The project will evaluate common commercial drone technology to document baseline potential for decision support in soybean. The information generated by this project will be used to provide robust training to County Extension Agents and farmers across North Carolina on the use of these technologies to enhance profitability.

We propose to develop a pilot Data Science Extension program at North Carolina State University, with a primary focus on data-driven soybean production. In the near-term, we will leverage our networks to interview stakeholder groups and identify grower interests and needs in data services and training, and use this information to prepare a vision and series of recommendations for Extension programs across the country regarding data science Extension programming. Additionally, we will develop a series of data science Extension products -- including tutorials, workshop materials, social media releases -- tailored for soybean growers. These products will be advertised through the Soybean Research Information Network to ensure national visibility and access. In the long-term, we will pursue additional funds to establish additional data-driven soybean research and Extension projects that serve the priorities identified through interviews held during this project, and advance the field and practice of data science Extension. USB funding is specifically sought to support personnel who will interview growers, summarize findings in a white paper, and prepare data-focused soybean Extension products.

One of the most expensive labor costs associated with sweetpotato production occurs during transplanting. Current approaches require multiple (between 4 and 16) human operators to manually sort and singulate individual slips from a pile and feed them into a planter that buries the slip into the ground. The ability to automate the slip singulation process would allow growers to save on prohibitively expensive labor costs while increasing transplanting efficiency. This project will take an integrated computer vision, mechanization, and robotics approach to automatically separate, identify, and grasp a single slip for transplanting operations.

Severe weather, especially during Atlantic hurricane season, will continue to impact corn production in North Carolina. After the negative impacts have occurred, quickly assessing the extent and severity of the damage can aid in recovery and mitigation of losses. UAVs or manned aircraft provide opportunities to survey locations where crop lodging may have occurred, but often significant additional analysis was needed. Rapid assessment of the collected imagery can help to direct effort to where it is most needed.

Farmers across Eastern North Carolina are becoming increasingly dependent on precision agriculture (PA) technologies to sustain profit and compete nationally and globally. Schimmelpfennig (Farm Profits and Adoption of Precision Agriculture, USDA ERS 2016), Ward et. al (Precision Agriculture Technologies for Cotton Production in North Carolina, Cotton, Inc. 1/01/12 - 12/31/20) and others have documented that PA technologies including telematics on agricultural machinery, in-field sensor networks and other automated data centric applications provide farmers with increased efficiencies and reduced operating costs by helping detect and manage the impact of crop damage, invasive species, erratic weather patterns, irrigation levels and other key elements to improve yields and reduce risk. Auto-steered combines and tractors help reduce operator errors by determining precise field locations (that are often difficult to determine accurately by sight) and compensating for operator fatigue. Field operators using guidance systems have real-time positional and operational information in the equipment cab which saves money by reducing over- and under application of crop inputs and off-target seeding of field crop rows. Advanced guidance systems allow equipment to return to the exact same position within the field which improves all in-season field operations including harvest. PA also enables farmers to determine the ideal time to plant and manage resources such as water, soil, fertilizer, seeds and labor to increase profits. As food borne illnesses become more prevalent, PA also helps farmers track and trace the production and delivery of their crops from farm to markets to consumers to mitigate food safety issues, spoilage and waste across the value chain. Consumers are demanding sustainability and to know the origin and production system used to grow the food they eat and the fibers they wear. Only through traceability throughout the value chain, enabled by broadband data streams, can those requests be met in a meaningful way.

As North Carolina corn producers have been reminded in recent years, harvest season is hurricane season. After a severe weather event crops must be rapidly assessed for producer decision making, disaster declarations, and insurance claims. Producers need to know how to assess losses and plan remediation. Disaster response agencies need to quickly estimate agricultural losses. Insurance providers need to verify claims and prepare for financial outlays. Agricultural statisticians need to estimate crop damage to report to federal agencies. This project will identify underlying relationships to describe crop damage from UAV imagery, automate the analysis process, and engage stakeholders to better understand data. This project will build on a preliminary project to use unmanned aerial vehicles (UAVs) to build tools to detect crop damage from imagery and to then automate the process using machine learning tools. Imagery from unmanned aerial vehicles emphasize early to mid-season analyses. The late season environment has not received similar efforts in data collection or analytical tools. New data-driven tools are needed to skillfully identify the presence and severity of crop damage. Long term outcomes of this project will allow better response, recovery, and resiliency when crop damage occurs. Understand the underlying damage metrics and automating the damage detection processes can put UAV-based crop damage detection to work in a region already heavily impacted by severe weather.