Connecting soil characteristics to edge-of-field water quality in Ohio.

Soil health and water quality improvement are major goals of sustainable agricultural management systems, yet the connections between soil health and water quality impacts remain unclear. In this study we conducted an initial exploratory assessment of the relationships between soil chemical, physical, and biological properties and edge-of-field water quality across a network of 40 fields in Ohio, USA. Discharge, dissolved reactive P (DRP), total P (TP), and nitrate (NO3 ) losses associated with precipitation events via surface runoff and tile drainage were monitored. Agronomic soil tests and a suite of soil health indicators were measured, then predictive relationships between the field average soil properties and tile drainage and surface runoff discharge and DRP, TP, and nitrate loads were explored with random forest and multiple linear regression approaches. Among the soil health indicators, water extractable C and N were consistently found to be positively related to tile nitrate loads, but other soil health indicators had little or inconsistent importance for water quality impacts. Several other soil properties were important predictors, particularly soil P pools for surface and tile DRP and TP losses as well as Mehlich-3 (M3) extractable Fe and Al for surface and tile discharge. Thus, we did not observe strong evidence that soil health was associated with improved edge-of-field water quality across the edge-of-field monitoring network. However, additional studies are needed to definitively test the relationships between a broader array of soil health metrics and water quality outcomes.

Assessing the Accuracy of Farmers' Nutrient Loss Risk Perceptions.

Use of nutrient management practices to reduce nutrient loss from agriculture and its associated water quality consequences, including hypoxia and eutrophication, is widely encouraged. However, little is known about which factors influence farmers' risk perceptions associated with nutrient loss, and thus possibly influence their decisions to adopt such practices. To determine which factors were associated with relative "accuracy" of nutrient loss-associated risk perceptions, specific farm field management information was used as inputs to a Soil and Water Assessment Tool model of the study watershed to produce water quality outputs for each modeled farm field. This information was paired with farmers' risk perceptions associated with nutrient loss on their farm to assess relative "accuracy" of each farmer's perceptions compared to the rest of the farmers in the study. We then investigated characteristics of the farm and farmer that are associated with comparative "overprediction" and "underprediction" of risk, and found that characteristics of the individual (conservation identity, prior conservation practice adoption, efficacy beliefs, and perceived seriousness of the consequences of nutrient loss) are more important in determining whether farmers are likely to "overpredict" or "underpredict" risk than is the objective (modeled) vulnerability of their land to nutrient loss.