Banca de DEFESA: VITOR HUGO MAUES MACEDO

Uma banca de DEFESA de DOUTORADO foi cadastrada pelo programa.
STUDENT : VITOR HUGO MAUES MACEDO
DATE: 31/08/2021
TIME: 14:00
LOCAL: meet.google.com/qew-uvzv-wys
TITLE:

Modeling the biomass production of the Tanzania guinea and Massai grasses under the influence of agrometeorological variables in the Amazon and semi-arid regions of Brazil.

KEY WORDS:

forage. empirical models. multivariate analysis. penalized regression. agricultural meteorology.

PAGES: 80
BIG AREA: Ciências Agrárias
AREA: Zootecnia
SUBÁREA: Pastagem e Forragicultura
SPECIALTY: Manejo e Conservação de Pastagens
SUMMARY:

Forage production results from the systemic action of environmental factors that act on grassland ecosystems. The study of the biophysical processes involved and the decision making based on predictive models, must take into account the conditions of the environment in which the plant develops. In the quest to understand how environmental variables act in the production of forage, and to obtain models applied to equatorial regions, the objective was to evaluate the biomass production of the cultivars Tanzania and Massai, in experiments carried out under the influence of both the Amazon biome and the climate. Brazilian semiarid, from the perspective of multivariate analysis and generation of empirical models. Data from five experiments: two with Tanzania guinea grass and one with Massai grass, carried out in the municipality of Igarapé-Açu-PA; two more experiments: one with Tanzania guinea grass from the municipality of Pentecoste-CE and another with Massai grass from the city of Fortaleza-CE; underwent canonical correlation analysis (CCA), discriminant function (DF), and principal components (APC). Measurements of total forage accumulation (AcFT), leaf depth (AcLF), stems (AcH), green biomass (AcBv), canopy height, tiller population density (DPP) and leaf area index (IAF) were used , as agronomic variables; and temperature data (Tmax, Tmin and Tmed), precipitation (Prec), solar radiation (Rs), real and reference evapotranspiration (ETR and ETo), degree-days of growth (GDC), photothermal units (UF), nitrogen supplied (NF), light index (IL), thermal index (IT), water index (IH), water storage (ARM) and climate growth index (ICC), as agrometeorological variables. Data from experiments carried out in the Amazon region were used to generate models using penalized regression methods (LASSO, LASSO adaptive, and elastic network), and traditional methods for model selection (forward, backward and stepwise), later validated with data from the semi-arid region. ACC generated a 95% correlation coefficient between canonical variable 1 - which represented a contrast between variables related to the hydrological cycle and variables related to energy supply - and canonical variable 2, which reflects a contrast between AcMM and AcFT with the height. The groups of experiments were strongly separated, based on the cultivar studied, by means of the discriminant function 1, characterized by greater AcLF in relation to the other types of production, and a contrast between IH and ICC in relation to the other agrometeorological variables. As for the studied regions, experiments carried out under irrigated conditions in the semi-arid region and little meteorological variation, present greater production potential - separated by the discriminant function 2 - of experiments carried out under seasonal conditions in the Amazon region. The ACP was able to characterize the observations based on the environment through the main component that reflects the contrast between hydrological cycle variables and energy supply variables, while the characterization based on the biomass production occurred through the component that reflects the distinction between AcBv and AcMM. Models generated from penalized regression methods showed a lower mean square error in the validation process compared to traditional methods, which allows the use of models applied to equatorial climates that are similar in terms of energy supply, but different in terms of water regime.

BANKING MEMBERS:
Interno - 1331235 - THIAGO CARVALHO DA SILVA
Interno - 1974294 - ANIBAL COUTINHO DO REGO
Externo à Instituição - MARCOS NEVES LOPES - IFRN
Externo à Instituição - RODRIGO GREGÓRIO DA SILVA - IFRN
Externo à Instituição - WILTON LADEIRA DA SILVA - UFG
Externo à Instituição - MÁRCIO ANDRÉ STEFANELLI LARA - UFLA
Externo à Instituição - FELIPE NOGUEIRA DOMINGUES - UFPA