Parametrização do modelo AquaCrop e simulação da transpiração e produtividade do algodoeiro sob lâminas de irrigação e nitrogênio

Abstract

The cotton plant Gossypium L. is an agricultural species of great economic and social importance for Brazil, whose demand for technologies that enhance productivity combined with sustainability, should be constantly sought. Productivity simulation models, such as AquaCrop, allow the determination of the response of the crop as a function of variation in water availability and nitrogen fertilization. Thus, this work aimed to parameterize the AquaCrop model to evaluate the yield of cotton as a function of water and nitrogen levels. The experiment was conducted in the experimental area of the Federal University of Rondonópolis. The soil is classified as dystrophic Red Latosol and the cultivar used was the IMA5801B2RF. Irrigation was by drip system with irrigation rates of 30%, 90%, and 150% of crop evapotranspiration. Nitrogen was applied via fertigation at doses of 20, 110, and 200% of the recommended dose for cotton. The model input parameters determined locally were soil physical parameters: field capacity (10 kPa), permanent wilting point (1500 kPa), saturated hydraulic conductivity (constant load permeameter); vegetative parameters: canopy cover (%); duration of phenological stages (days); maximum depth of the root system (m); meteorological parameters: maximum and minimum air temperature (ºC), relative humidity (%), wind speed (m s-1 ), global solar radiation (MJ m2 d -1 ). After parameterization, the model was calibrated to simulate the effect of nutrient (N) stress. Seed cotton yields were 3495 kg ha-1 (ET90N110), 2027 kg ha-1 (ET90N20), and 2075 kg ha-1 (ET30N110). Regarding the simulation, it was observed that the simulation of the productive variables, biomass, and seed cotton, by the AquaCrop model were higher than the measured data. In general, the largest errors occur for the simulation of treatments with water or nutrient stress. Despite this, the statistical parameters R2 , d and c, were greater than 0.90, which classifies the model performance as "optimal".