Articles
Uncertainty analysis of modified VegSyst model applied to a soilless culture tomato crop
Article number
1182_30
Pages
249 – 256
Language
English
Abstract
Over the last decades, the soilless culture technique has rapidly progressed in several developed countries linked to crop growth control environment and automation.
Several crop growth models have been developed for decision support systems.
Thus it is important to quantify the uncertainty associated to the predicted variables of these models previously to their application.
An uncertainty analysis aims to know quantitatively the variability of model components for a specific situation and the derivation of an uncertainty distribution for each state variable and model output.
Recently, the VegSyst model was developed to assist the nitrogen (N) supply and irrigation management for some horticultural crops.
The basic input data are measurements of air temperature, relative humidity, and solar radiation which are climatic data that are commonly measured, by growers, in the greenhouse.
The model was developed assuming non-limiting conditions of water and N use.
The aim of this research was to modify the VegSyst model including a leaf area index (LAI) sub-model, in order to improve the prediction of dry matter production (DMP), and N uptake (Nup) for a soilless culture using plastic bags filled with tezontle (volcanic sand) as substrate.
LAI was modeled using accumulated normalized thermal time and photosynthetically active radiation.
An experiment with a tomato crop was carried out during the autumn-winter 2015 in a greenhouse located at University of Chapingo, Mexico.
The collected data were used to carry out an uncertainty analysis in which the inputs were the model parameters and the outputs were the predicted DMP, LAI, and crop N content.
Probability density functions were defined for each model parameter to calculate the corresponding statistics and histograms of the model outputs.
Also the generalized likelihood uncertainty estimation (GLUE) Bayesian method was used.
Results showed that LAI can be predicted better by the model than DMP and Nup.
Several crop growth models have been developed for decision support systems.
Thus it is important to quantify the uncertainty associated to the predicted variables of these models previously to their application.
An uncertainty analysis aims to know quantitatively the variability of model components for a specific situation and the derivation of an uncertainty distribution for each state variable and model output.
Recently, the VegSyst model was developed to assist the nitrogen (N) supply and irrigation management for some horticultural crops.
The basic input data are measurements of air temperature, relative humidity, and solar radiation which are climatic data that are commonly measured, by growers, in the greenhouse.
The model was developed assuming non-limiting conditions of water and N use.
The aim of this research was to modify the VegSyst model including a leaf area index (LAI) sub-model, in order to improve the prediction of dry matter production (DMP), and N uptake (Nup) for a soilless culture using plastic bags filled with tezontle (volcanic sand) as substrate.
LAI was modeled using accumulated normalized thermal time and photosynthetically active radiation.
An experiment with a tomato crop was carried out during the autumn-winter 2015 in a greenhouse located at University of Chapingo, Mexico.
The collected data were used to carry out an uncertainty analysis in which the inputs were the model parameters and the outputs were the predicted DMP, LAI, and crop N content.
Probability density functions were defined for each model parameter to calculate the corresponding statistics and histograms of the model outputs.
Also the generalized likelihood uncertainty estimation (GLUE) Bayesian method was used.
Results showed that LAI can be predicted better by the model than DMP and Nup.
Authors
A. Martinez-Ruiz, I.L. López-Cruz, A. Ruiz-García, J. Pineda-Pineda, A. Ramírez-Arias
Keywords
mineral nutrition, simulation model, decision support system, Monte Carlo, Bayesian method
Online Articles (34)