Analysis of parameter uncertainty and sensitivity in PCPF-1 modeling for predicting concentrations of rice herbicides

Abstract

is paper demonstrates the procedures for probabilistic assessment of a pesticide fate and transport model, PCPF-1, to elucidate the modeling uncertainty using the Monte Carlo technique. Sensitivity analyses are performed to investigate the in uence of herbicide characteristics and related soil properties on model outputs using four popular rice herbicides: mefenacet, pretilachlor, bensulfuron-methyl and imazosulfuron. Uncertainty quanti cation showed that the simulated concentrations in paddy water varied more than those of paddy soil. is tendency decreased as the simulation proceeded to a later period but remained important for herbicides having either high solubility or a high 1st-order dissolution rate. e sensitivity analysis indicated that PCPF-1 parameters requiring careful determination are primarily those involve with herbicide adsorption (the organic carbon content, the bulk density and the volumetric saturated water content), secondary parameters related with herbicide mass distribution between paddy water and soil (1st-order desorption and dissolution rates) and lastly, those involving herbicide degradations.

Inverse analysis to estimate site‐specific parameters of a mathematical model for simulating pesticide dissipations in paddy test systems

“…The upper and lower boundaries of the varying parameters were set as follows: from maximum to minimum values of corrected or calculated data, M /2 to 2* M , to the numerator of M is the initial value of the varying parameter . The sample size generated from the uniform distribution of the parameter range was set to 250, which was proven to be sufficient for the PCPF‐1 model with Latin‐hypercube sampling . The visual assessment was again conducted to check if the sensitivity range of C PW generated from the ‘sensRange’ function included or traced the observed data.…”

“…Similarly, the effect of the parameter variations on the model cost was visualized by the ‘modCRL’ function and quantified by the standardized rank regression coefficient (SRRC), which is the robust sensitivity measure estimated from the rank‐transformed regression model. A detailed explanation of SRRC is available elsewhere . In this study, the SRRCs of the varying parameters were calculated using the ‘src’ function in the R package ‘sensitivity.’…”

Development and validation of the SPEC model for simulating the fate and transport of pesticide applied to Japanese upland agricultural soil

“…This sample size proved sufficient for a pesticide fate and transport model in the case of pesticide applied in rice paddies. 32) Uniform distributions were given to all investigated parameters. All parameters except the Q 10 parameter were allowed to vary a maximum of ±10% from the values used in the deterministic scenario presented in Table 1.…”

“…32,33) The method relies on a stepwise regression analysis that computes standard rank regression coefficients (SRRCs) for the predictors (inputs) that have the most significant influence on the predictions (outputs). By ranking the input parameters by absolute values of SRRCs, the model's most sensitive parameters can be highlighted.…”

Probabilistic assessment of herbicide runoff from Japanese rice paddies: The effects of local meteorological conditions and site-specific water management

“…20) For this uncertainty application, all physicochemical inputs were xed and uncertainty inputs were char- acterized only for corresponding weather conditions and water management practices. e uncertainty of the physicochemical inputs of herbicides can be referred to Boulange et al 31) In the MC uncertainty assessment, one of the most critical limitations is the misspeci cation of sampling distributions for the uncertainty inputs due to lack of knowledge, which may gives false uncertainty information of output. 22,25,32) Although there are various options for selecting probability density functions (PDFs) on the basis of experimental data, hypothesis, and expert judgment, 32) no recommended distribution may be found in terms of water management parameters.…”