RESUMEN
HYDRUS 2D was used to simulate chloropicrin (CP) emissions across a range of expected application and environmental conditions present within California, where CP is widely used in the pre-plant treatment of soils for high-value specialty crops. Simulations were developed based on field calibration work and physicochemical parameters from literature with additional consideration of application rate-dependent degradation and applicator practices including application depth, application mode, and tarp material. Model output was compared to the distribution of indirect whole-field flux estimates derived from field monitoring studies using measures of maximum 8-h, maximum 24-h, and cumulative emissions due to their relevance to public health. We observed a strong linear relationship (R2 ≥ 0.80, p < 0.001) between HYDRUS-simulated and field-based maximum flux estimates and no evidence of statistical difference depending on the estimation source for maximum 24-h flux. A linear relationship of similar strength (R2 = 0.82, p < 0.001) was observed between simulated and field-based cumulative emission estimates, although mean HYDRUS estimates were lower than field-estimated values for some high-emission application methods. Analysis of simulation output demonstrated large differences in CP emissions in response to application method and a non-linear increase in CP emissions with increasing application rate, with considerable interaction between application variables including application depth, tarp types, and field layout. The findings generally support the use of simulated CP emission estimates as a tool to address gaps in field-based flux estimates, particularly where characterization of short-term peak emissions is needed.
RESUMEN
The fumigant 1,3-dichloropropene (1,3-D) is widely used for control of soil-borne pests and pathogens, but post-application emissions may lead to off-site transport and possible human exposure. The fraction of applied material emitted into the atmosphere and the magnitude of peak emissions are two quantities used by regulators to protect public health and are typically based on field estimates. However, the current body of field studies covers only a narrow subset of the broad range of application practices and soil conditions under which applications are performed and is subject to an unknown level of estimation error. Here we use the HYDRUS model to estimate cumulative and peak emissions of 1,3-D for 17 application methods used in California. The simulations are parameterized with soils data from 16 fields sampled immediately prior to fumigation in order to establish a representative distribution of initial soil conditions. The results demonstrate a wide range in cumulative emissions, with mean losses of initial applied mass between 10 and 58% over two weeks depending on application method. Emissions are highly variable in response to soil conditions, with coefficients of variation ranging from 16 to 54% for cumulative flux and 26 to 67% for peak three-hour flux depending on application method. The simulated distributions show similarities to the available field study estimates in terms of the mean and spread of distributions, particularly in the case of cumulative emissions, indicating that the modeling approach could be a useful tool to support regulatory decision-making in cases where field data is limited.
