Our group (Bill Jordan (retired from EPA), Edem Avemegah (sociologist), Haiying Tao (soil scientist), and me (tech designer/social scientist) focused on questions around adoption of digital agriculture technology. We had a wide-ranging discussion; our main conclusion was that this was a complex problem that required the integration of sociological understanding with technological work and policy development.
Some of the concrete avenues we discussed include: - Extensions: connecting researchers with farmers - On-farm research and demonstration - Reaching out explicitly to early adopters; often other farmers will pick up the technology once they see the early adopters having success, and this leverages the social circulation of knowledge already happening in farming communities - Design strategies that make it easier to 'do' conservation, figure out what applies to your situation, including decision support systems - Technological developments that make the technology more relevant and accurate; for example, there are few sensors available for aspects of soil health outside how much water it contains - Addressing social issues around e.g. trust; financial issues are definitely part of the story about why people aren't adopting, but not the only thing. In particular, it may help to: - Understand the sources of information that farmers trust and target those - Include consultants and service providers
Research idea important for realizing the power of data to advance conservation: Among the research programs/projects/ideas introduced by participants in the breakout room, the one I want to highlight focuses on understanding education and training requirements. An imagined transition to next-generation technology suggests a need to train and re-train workers on farms and in off-farm service businesses. Beyond human capital, we also touched on the broader challenge of developing new organizational capabilities to function and lead in a world characterized by digital agriculture and precision conservation. This problem statement can be linked to research questions such as, what are the education and training implications of the division of labor between on-farm and off-farm actors in different commodity sectors, on farms of different sizes, and in different regions. How can public-sector education programming complement training provided by firms? Research response to a specific policy challenge/opportunity: Our group focused discuss on issue #1 – adoption of digital tools that advance conservation Public policy, including the Farm Bill, was identified as a potential resource to support adoption (EQIP), but it was also identified as a potential impediment. Policy must allow flexibility, which farmers need to do innovative things. Risks as perceived by farmers’ is a constraint to adoption. Here we identified data pooling as a potential way to reduce risk. If farmers and other actors in production systems collect and share data, and the data can be focused on decision-support, uncertainty about adoption of novel production practices can decrease. Here we see a bridge between an ambition to advance adoption and the problems of data standards and data access. Thanks to Gary, Juan, Jubing, and Bernie. Please add and refine, as you choose.
Our panel focused on Question 4 on mobilizing data for spatial targeting and also spoke to issues of payments for ecosystem services. Much of the discussion was on the limits of mobilizing data and toward the end we were starting to discuss potential solutions. Limits : Uncertainties : We acknowledged that there are already a number of resources for targeting: SWAT, RUSLE, LIDAR etc. Some ecosystem services are easier to measure and model with tools while others are harder. Variables such as type of pollutants, soil types, soil moisture, slopes etc. complicate models. E.g. phosphorous loading presents different data challenges compared to nitrogen. To advance more evidence-based policies, we also discussed the need to make the underlying assumptions of models more explicit. This was particularly in reference to various applications of the SWAT model. Scale : In discussing the scale of the target, we discussed watershed, farm and field, and intra-field scale. It seems there was more consensus on granular watershed levels datasets but less on field-level. Spot sensing within fields would also involve more extensive data gathering capacity. Also raised a question of the role of farmers in adopting sensor technologies, not only in adopting precision agriculture technologies that lower external inputs, but in supplying information for targeting like installing and maintaining sensors. Education and training could play a role here. Opportunities/Research areas : Supporting farm-level adoptio n: We recognized that adoption of conservation technologies or BMPS will entail costs and risks. What are possible mechanisms for supporting farm adoption of these solutions and could insurance play a role? We discussed farm bill payments, subsidies, and private sector payments. An important area for further analysis may be a possible role for farm insurance companies. This would help buffer potential yield losses from adoption and provide assurances. What role might farm equipment providers have? We saw opportunity within the existing interactions among farmers and equipment companies. We observed that many equipment providers provide additional but optional data services. These are not their core business but when consolidated, they may provide information for targeting and a channel to reach farmers. In supporting farm-level adoption, are there some ways to tap into the optional services farm equipment companies are already providing to farmers?