Addressing alternative approaches for spatial modeling of herbicide retention in soil
Date
2019Author
Giannini Kurina, Franca
Hang, Susana Beatriz
Rampoldi, Edgar Ariel
Córdoba, Mariano Augusto
Macchiavelli, Raúl E.
Balzarini, Mónica Graciela
Metadata
Show full item recordAbstract
Glyphosate retention coefficient (Kd) is modeled as function of soil variables, from a regional sampling, using: Ordinary and Partial Least Square regression, Random Forest, Generalized Boosted regression (GB), and Bayesian modelling with INLA; all regressions were fitted using spatial constraint on residuals. INLA produced the best fit, but GB the best spatial prediction.