Date

Summer 2012

Document Type

Master's Thesis (Open Access)

Degree Name

Master of Science (M.S.)

Department

Science & Environmental Policy

Abstract

Frequent pesticide detection at toxic levels to test organisms in California's Central Coast waterbodies has motivated regulators, resource agencies and end-users to investigate and adopt management practices and technologies to diminish agricultural chemicals entering receiving waters. Treatment wetlands are a technology of special interest because of their ability to simultaneously treat multiple pollutants commonly found in agricultural and urban runoff including nitrate, suspended sediment and pesticides. We sought evidence for transformation of three highly water soluble pesticides (diazinon, methomyl and acephate) in a full-scale constructed treatment wetland located at the base of the Salinas Valley. We pumped water into the wetland from a slough containing agricultural runoff. The pumping rate was set to achieve a four-day mean residence time, and outlet samples were collected four days after inlet samples. We developed a dynamic tanks-in-series model and fit it to pesticide concentration data from the wetland, using parameters for number of tanks in series, mean hydraulic residence time, pesticide decay, and two parameters for inlet concentrations outside of the sampling period. We used a Bayesian analytical approach to determine the 95% credible intervals (CI) and most likely values for the five model parameters, and developed inference for pesticide decay based on the CI for the decay rate parameter. The CIs for the three pesticide decay parameters were positive and did not span zero, supporting the postulate that the wetland removed these pesticides to some extent. CIs for first-order decay rates were 0.097-0.289 day-1 for diazinon, 0.068-0.232 day-1 for methomyl, and 0.068-0.246 day-1 for acephate. These intervals can be used in conjunction with simple decay models to optimize the design of wetlands and to estimate size requirements.

Share

COinS