Desenvolvimento de um programa computacional para o modelo de cana-de-açúcar QCANE

Abstract

ABSTRACT

The cultivation of sugar cane (Saccharum spp) is one of the main agricultural activities in the world, having an important role in world food security, due to its main product and its by-products. Despite Brazil being the largest producer of sugar cane in the world, productivity in the State of Alagoas is low. The application of crop growth models is a fundamental tool to assist in decision making regarding agricultural practices and to increase crop productivity. The objective of this work was to develop a computational program for the sugarcane model QCANE and to evaluate the sensitivity of state variables of the model to photosynthesis parameters. QCANE is a mechanistic model based on eco-physiological processes that provides daily predictions of sugarcane growth and development. The computational program was developed in Fortran 90 programming language. The program was structured in modules and external subroutines. To read ASCII files, the TUTTIL library was used to read the input file of the model. The program consisted of two modules and several external subroutines, representing each process simulated by the model, in addition to calculations of daily rates and corrections for the simulation time. The sensitivity simulations carried out in this study showed that the state variables of the model, such as stem dry mass and LAI, were sensitive to the parameters related to the concentration of N in the leaf. Thus, the calibration of these parameters for each sugarcane variety becomes an important factor for the performance of the model.

Keywords: Sensitivity analysis, Fortran 90, Mechanistic models.