Best Management Practices are farming practices that help increase soil health, decrease soil erosion, and stop excessive amounts of agricultural runoff. This means less phosphorous and solids enter the streams, rivers, and lakes, thus lowering the amounts of blue green algae blooms that are causing the toxicity, cloudiness, and smell of the lake.
There are many predictors that have a significant input on increasing or decreasing BMP implementation but, for the sake of this summer’s project, I looked into farmers’ social networks. My research suggests that to increase BMP use, farmers need to become part of a social network, most effectively through Farmer-Led Councils, which would help deal with other problems that hinder the adoption of BMPs.
The LAKES REU program was a great experience because in an 8-week program I learned so much about survey creation, data input, and statistics. The farmer survey that we created consisted of about 15 minutes worth of questions that ranged from personal values to fairness to education on farming practices and implementation of best management practices.
There were also questions about willingness to work with local and state agencies to implement BMPs. All of the questions were shaped to get a feel about what stands in the way and what is helpful for the adoption of BMPs.
The results for my project were two different regression models with an interaction and a social network graph. The first regression model was created with BMP usage as a dependent variable, meaning the variable that is dependent upon the influence of other variables, called the independent variables. The independent variables for the model were soil test frequency, closeness centrality, ecological impact, perception that farmers are unfairly targeted as problems of water quality, the value of organic matter, and farm size.
All of these variables came together as valuable predictors for the BMP use, with soil test frequency, closeness centrality, unfairly targeting farmers, and value of organic matter all being statistically significant. This brings in the second regression model and where it gets really interesting.
In the second regression model, we used closeness centrality as the dependent variable (since this is a good predictor of BMP usage, we wanted to know what predicts high closeness centrality). Essentially, closeness centrality measures if somebody is highly connected to a network yet not directly connected to many people (e.g. so they may trust farming advice from only two other people in the network, but through those two people they are connected to a dozen others who may relay knowledge to them).
We found that age, social connection information (like Farmer Led Councils), capital costs, farm size, and an interaction between farm size and capital costs were predictors of whether or not a farmer has high closeness centrality (thus leading to higher BMP use).
The really interesting part of these results was the interaction which showed that while all farmers worry about capital costs, for farmers of fewer acres, capital costs have much more of an impact on their closeness centrality than larger acre farm producers. This suggests smaller acre farmers need financial help if they are going to get integrated into the network of farmers in the watershed and if they are going to adopt BMPs.
What does this mean? Farmers need help with the finances behind implementing BMPs, but it is not just about money. Farmers also need to be brought into the networks, because word alone within a network has impact on the farmers BMP index.
More Farmer Led Councils working with UW-Extension, NRCS, and Land Conservation Divisions could tackle both of these problems up to a point by getting more farmers into larger social networks and creating more suitable incentive programs to help ease the capital costs of BMP implementation.
Ultimately, relatively easy changes can be made if farmers work together to create the infrastructure necessary to implement BMPs and remove phosphorous from our waterways.