BALZARINI MONICA GRACIELA
Congresos y reuniones científicas
Título:
Two Step Procedure to Model Site Specific Herbicide Soil Persistence
Autor/es:
GIANNINI KURINA, FRANCA; HANG, SUSANA; BALZARINI, MÓNICA;
Lugar:
Seul
Reunión:
Congreso; International Biometric Conference; 2020
Institución organizadora:
International Biometric Society
Resumen:
Soil herbicide persistence is the length of time the herbicide molecule remains active in soil and it is crucial to describe risks of diffuse contamination in agriculture. Persistence is characterized by ?half-life?, which is the time it takes to reach half of the initial concentration supplied to soil. Half-life is estimated as a function of the dissipation curve parameters. Analytic quantification is costly for obtaining dissipation curves at many sites. Methodological tools to predict half-life in a continuous spatial domain, from a sample of dissipation curves, become crucial in regional studies. Since herbicide persistence in the environment depends on sites variables, model-based predictions of half-life as function of environmental features, are pursuit. The objective of this work was to design a statistical workflow for digital modeling of soil herbicide persistence at regional scale. From a regional soil survey, a sample of sites was drawn using the cLHS method. Samples were fortified with the herbicide atrazine and incubated for 21 days. Herbicide concentrations were measured at days 0,3,7,14 and 21 on each soil by liquid chromatography coupled to tandem mass spectrometry (LC-MS / MS) using QuEChERs. A two-step procedure was proposed for digital mapping of herbicide persistence in the environment. First, an exponential model with a random site effect, associated to the decay rate, was fitted to derive atrazine half-life for each sampled soil. Second, a Bayesian regression with a site random effect relating the resulting half-life values with soil and land-use values was adjusted to predict the spatial distribution of atrazine persistence at un-sampled sites for mapping. The addition of a random effect on the decay rate produced a better fit than a fixed exponential model and allowed us to explore half-life variability among soils. Atrazine persistence was mainly explained by the agricultural use of land (sites with previous grass crops had higher decay rates than other land-uses). The two-step procedure made possible to a