Simulating potato tuber yield in subtropical and temperate environments

Abstract

Crop models are important tools for decision-making in different agricultural crops. Potato (Solanum tuberosum) is an important crop in Brazil, but studies with potato simulation models in Brazil, mainly in the South, are scarce. This study aimed to calibrate and evaluate statistical models for simulating potato tuber yield in Santa Maria, Rio Grande do Sul (subtropical climate) and São Joaquim, Santa Catarina (temperate climate). Six models from literature were evaluated. In Santa Maria, the planting dates of 02/28, 03/27, 08/12, 09/15, 10/15, 11/20/2003 were used to calibrate the models, and the planting dates of 01/21, 02/12, 12/24/2003 and 01/28, 02/27, 03/26, 04/26/2004 were used to evaluate the models. Yield data of 35 cultivars grown in five years in São Joaquim were also used to evaluate the original version of the models. The statistics used for selection of models was the root mean square error. Models with the original coefficients had the same performance in simulating potato yield in Santa Maria and São Joaquim, with the Hartz and Moore, Johnson and others, and Pereira and others models having the best performance. After calibration, the Sands and others model had the best performance followed by the models of Pereira and others, and MacKerron and Waister.