Banca de QUALIFICAÇÃO: HELDER JOSÉ FARIAS DA SILVA

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
DISCENTE : HELDER JOSÉ FARIAS DA SILVA
DATA : 24/02/2017
HORA: 08:00
LOCAL: Sala de seminários de estatística no CCET
TÍTULO:

STOCHASTIC MODELING OF POTENTIAL EVAPOTRANSPIRATION IN TROPICAL BIOMES AND THE INFLUENCE OF LAND-USE ON PRECIPITATION

PALAVRAS-CHAVES:

MOD16, Amazon, Mixed Models, Penman-Monteith

PÁGINAS: 64
GRANDE ÁREA: Ciências Exatas e da Terra
ÁREA: Geociências
SUBÁREA: Meteorologia
ESPECIALIDADE: Sensoriamento Remoto da Atmosfera
RESUMO:

The precise monitoring and characterization of Evapotranspiration (ET) and Rainfall in its spatial and temporal components represent a great challenge for understanding energy and hydrological partitioning between the surface and the atmosphere in different biomes, especially in tropical regions that Are the major sources of moisture and that they exert strong control over the processes of atmospheric circulation on a regional and global scale, since they condition the energy of the atmosphere and alter the characteristics of the air masses in it. In this sense, the main objective of this research will be to model Potential Evapotranspiration (ETp) inserted in the Brazilian Amazon region, based only on the spectral data extracted from orbital sensors through Stochastic Modeling through Linear Models of Mixed Effects as well as to evaluate the characteristics Rainfall in different regions (forests and deforestation). In order to do so, we used monthly ETp data, estimated by the Penmam-Montheith parametrized standard method using data from the Instituto Nacional de Meteorologia and orbital data of the MOD16 algorithm, as well as rainfall data from the precipitation radar based on the algorithm 2A25 of the TRMM satellite And soil and deforestation data obtained from the MCD12Q algorithm and the PRODES / INPE project, respectively, for the period from 2000 to 2014. Firstly, homogeneous regions of ETp were determined by cluster analysis using Ward's agglomerative hierarchical method as well as the verification of significant differences between them through Variance Analysis (ANOVA). The sensitivity of the standardized Penman-Monteith equation was then evaluated in the homogeneous sub-regions in order to obtain seasonal sensitivity coefficients. Finally, an empirical approach was developed for estimation of ETp based on remote sensing data from the MOD16 products, being calibrated and validated using the standard Penman-Monteith method fed with climatic data from 38 surface meteorological stations. In order to characterize rainfall microphysics in regions of covered (forest) and bare (deforested) soil, a polygon will be chosen within a region that contains the two distinct areas that will probably be within the state of Rondônia, To deforestation in the Amazon, which will be adopted as a PRODES project methodology for the identification of these regions or as a second option the MODIS MCD12Q product data that will be re-mapped to a 5 km square grid that corresponds approximately to the resolution of the TRMM 2A25 product that will provide The data of rain and not for the dry season (July, August and September).From the binary maps of 5 km of forest and non-forest grid that can reach a polygon of 600 km x 800 km of domain which will be used a random sampling strategy where an analysis will be performed in a series of smaller subdomains (for example, 200 km x 200 km) that will assist in the subsequent stages of quantification and characterization (statistical techniques to be discussed) in the microphysical structure of rain in each subdomain. The preliminary results suggest that the Brazilian Amazon has seven homogeneous regions of ETp. All sub-regions formed by the hierarchical agglomerative Ward method presented statistical difference at 5% significance. The Sensitivity Coefficient indicated that the Radiation Balance and Wind Speed are the most sensitive variables in the estimation of ETp by the standard method. Preliminary evaluations have shown that the ETp of the MOD16 product differs significantly from those obtained by the simplified Penman-Monteith method, overestimating by up to 69% in some regions. It also showed that in addition to the different values, there is difference in the seasonal peaks (maximum and minimum), which contributed to weak correlation between the methods. The proposed model increased the capacity utilization of global satellite data products on a basin scale, respecting the seasonal pattern of the standard method, reducing errors to less than 5% in the study region.

MEMBROS DA BANCA:
Presidente - 1164414 - WEBER ANDRADE GONCALVES
Interno - 2086472 - BERGSON GUEDES BEZERRA
Interno - 1752417 - CLAUDIO MOISES SANTOS E SILVA
Interno - 350691 - MARIA HELENA CONSTANTINO SPYRIDES
Interno - 320597 - PAULO SERGIO LUCIO
Externo à Instituição - CRISTIANO PRESTRELO DE OLIVEIRA - UFRN