Modelagem e análise da garantia natural da produção de gramíneas forrageiras em sequeiro no semiárido brasileiro

Abstract

The arid and semi-arid regions of the world with their peculiar climatic characteristics are the main causes of volatility in plant and animal production, along with other factors. This climatic variability affects forage production, but techniques for forage conservation are now known in order to mitigate the effects of seasonality. Management techniques, food storage, forage planning and the use of modeling are tools that support on the planning and consequently on the success of animal production on pasture in arid regions. Thus, the objective was to predict the growth of five forage plants by the PHYGROW model and associate such predictions with certain guarantee levels, to guide producers in the management and use of their forage resources. The work was carried out based on data collected during the years 2018, 2019, 2020 and 2021 in 10 Technological Reference Units (URT'S), most located in the Brazilian semi-arid region. Data from 5 forage grasses, 4 perennials (Buffel, Piatã, Massai, Corrente) and 1 annual (Sorgo Ponta Negra) were used. To determine the cutting time, the pasture height of 60 was taken into account; 40; 55 and 50 cm for buffel, piatã, massai and current grasses, respectively. To harvest the biomass, a 0.50 x 0.50 m frame was used. All material over 20 cm present in the frame was weighed. Sorghum collection was carried out by cutting the three central lines of the plot at a height of 20 cm from the ground. This cut was performed when the grains presented a silage point. The cut-off point was defined as the moment when the plants had grains in pasty consistency. The material was placed in a forced ventilation oven at 55°C to obtain the pre-dry weight. The production of Total Forage Biomass (BFT) was used to carry out simulations of plant biomass production, Phytomass Growth (PHYGROW). Field data collected from the URTs were used to parameterize and calibrate the model. The performance of the model was performed through the tests: mean forecast error (BIAS %), square root mean error (RMSE), mean absolute error (MAE) and Willmott's concordance index (d). Then, on the PHYWEB 2.0 platform, a 71-year time series related to dry biomass (kg ha-1) simulated for each URT was downloaded. The highest simulated biomass value was selected from these series. The highest values for each year of the analyzed time series were evaluated in the @Risk implement regarding their probability distribution. After defining the probability density function, 3 synthetic series of 1000 values each were generated. Finally, calculations of the probability of biomass production were carried out, based on different levels of guarantee. From these guarantee levels, regression calculations were performed between the verified guarantee levels and their respective biomasses with the aid of the SigmaPlot software (11.0). For buffel grass, model calibrations were obtained for Carlos Chagas-MG, Itapetinga-BA, Montes Claros-MG, São Raimundo Nonato-PI, Tenório-PB, all with Willmott index greater than 0.85. Considering the simulations for Ibaretama-CE as very good with an index of 0.85. For the URT of São João-PE, the simulation was classified as good with an index of 0.72. As for Fortuna-MA, it was determined as poor with an index of 0.55 and terrible for Baixa Grande-BA with an index of 0.37. For piatã grass, the simulations were excellent for Baixa Grande-BA, Batalha-AL, Carlos Chagas-MG, Itapetinga-BA, Montes Claros-MG, São João-PE, São Raimundo Nonato-PI, Fortuna-MA, all with the index greater than 0.88. Very good for Tenório-PB with an index of 0.77. Ibaretama-CE was bad with an index of 0.42. For massai grass, the simulations were excellent for Batalha-AL, Carlos Chagas-MG, Ibaretama-CE, Fortuna-MA and São João-PE. Itapetinga-BA, Montes Claros-MG, São João-PE all with an index greater than 0.85. Very good for Itapetinga-BA with an index of 0.79. Good for Tenório-PB with an index of 0.68. Bad for Montes Claros-MG with an index of 0.47. For current grass, the simulations were excellent for Montes Claros-MG and Ibaretama-CE. Very good for Itapetinga-BA and Baixa Grande-BA. Tenório-PB had a bad simulation. For Sorghum Ponta Negra, the simulations were excellent for Montes Claros-MG and Tenório-PB. Very good for Itapetinga-BA. Carlos Chagas-MG and Fortuna were good. They were terrible for Baixa Grande-BA, Ibaretama-CE and São João-PE. It is concluded that the forage biomass simulations were efficient through PHYGROW. The use of technological tools such as PHYGROW can be of great help in planning the plant and animal production of producers in the long term.