Mathematical modeling for prediction of banana rust thrips from atmospheric variables

Abstract

The following article presents a mathematical model to describe and predict the level of incidence of rust thrips on organic banana, with the aim of assisting growers in making decisions to reduce the effects of this pest on their crops. The model is based on the Campbell model (a model that depends only on temperature), but a modification is also proposed with the addition of variables such as relative humidity, wind speed and rainfall rate to describe the rate of insects' development. For the estimation of the parameters of the equations, random cross-validation of 100 sets, validated with real measurements of thrips, was used. Using the Campbell model, accuracies greater than 50% were obtained and with the addition of the variables the accuracy increased to 81%.