Models to estimate potassium and nitrate in changes of ion concentration in soil solution in banana cultivation

Abstract

The adequate use of fertigation is based on the application of nutrients by water according to crop needs throughout its cycle. The evaluation of nutrients in the soil solution may be faster if based upon soil electric conductivity and moisture by using parametric models. The objectives of the study were: (i) to evaluate models for estimating the concentration of nitrate and potassium in the soil solution as a function of moisture and apparent electrical conductivity; (ii) to define the need to adjust the models for possible potassium and nitrate concentration ranges in the soil solution throughout the banana crop cycle for the purpose of monitoring the concentrations of these ions. The Vogeler model was used, in which the concentration of ions is a function of moisture and electrical conductivity of the soil; the other model was the potential in which the concentration is a function of the electrical conductivity of the soil solution. The potential model presented a better performance compared to that of adapted from Vogeler, as for estimation of ion concentrations as for the sensitivity of this model to changes in concentration of this ion in the soil solution. The models showed greater sensitivity when fitted to data corresponding to all possible occurrences in the soil during the crop cycle. The results show that the variations in potassium concentration over time. The model must consider a range of concentrations with the possible values of the soil solution throughout the harvest cycle. The model presented an average of normalized errors (MEN) ranging from 0.01 to 0.02; square root of the mean of squares of errors (RMSE) from 0.01 to 0.03; and R² from 67.0% to 91.0%.The model should consider a range of concentrations with the possible values of the soil solution along crop cycle.