Abstract

LOPEZ-CRUZ, Irineo Lorenzo; RUIZ-GARCIA, Agustín; RAMIREZ-ARIAS, Armando  and  VAZQUEZ-PENA, Mario Alberto. Uncertainty analysis of a greenhouse lettuce crop (Lactuca sativa L.) model. Rev. Chapingo Ser.Hortic [online]. 2013, vol.19, n.1, pp.33-47. ISSN 2007-4034.  https://doi.org/10.5154/r.rchsh.2011.09.049.

An uncertainty analysis for a crop growth model allows to quantitatively evaluate the variability of the model's parameters by deducing an uncertainty distribution for the model's predicted variables. These studies only rarely have been applied to greenhouse crop growth models. In the present work a methodology to carry out an uncertainty analysis for a greenhouse crop model is described and it is applied to determine the variability of the NICOLET model parameters, which is a model developed to account for the growth of a greenhouse lettuce (Lactuca sativa L.) crop. Firstly, probability density functions were defined for all model parameters. Next, parameter values were chosen using Monte Carlo sampling. Both random and Latin Hypercube sampling and N = 2000 samples were used. Subsequently, 2000 computer simulations were performed in order to calculate the outputs of the NICOLET model for each scenario. Finally, an analysis of the distribution of the variables carbon in the vacuoles, carbon in the structure, total dry weight and nitrate concentration was carried out, by calculating their histograms and statistic measures. For all the simulations the software package for uncertainty and sensitivity analysis Simlab was used, which is available for the programming environment Matlab. The results showed that carbon in the vacuoles has the greater uncertainty given that its coefficient of variation (CV) for both random and Latin hypercube sampling was 35.27 and 35.67 %, respectively, then the nitrate content (CV = 18.16 % and CV = 19.07 %), the carbon in the structure (CV = 5.52 % and CV = 5.67 %) and the total dry weight (CV = 4.80 % and CV = 4.82 %).

Keywords : Uncertainty distribution; parameters variability; Monte Carlo simulation; sampling method; dynamic model.