Name: Thais Covre Delboni
Type: MSc dissertation
Publication date: 23/12/2019

Namesort descending Role
Antonio Sérgio Ferreira Mendonça Advisor *

Examining board:

Namesort descending Role
Antonio Sérgio Ferreira Mendonça Advisor *
Diogo Costa Buarque Co advisor *
José Antônio Tosta dos Reis Internal Examiner *
Marco Aurélio Costa Caiado External Examiner *

Summary: The selection of hydrological models for flow simulation in a watershed is a challenge
and comparison between models and their performances can help researchers to
choose the appropriate model for specific conditions. In this study, distributed
hydrological models Soil and Water Assessment Tool (SWAT) and Modelo de Grandes
Bacias (MGB) characteristics were theoretically analyzed with respect to hydrologic
processes representations and complexities. Models performances were evaluated
and compared considering applications in Itapemirim River basin flow simulations, This
basin presents an approximately 6,000 km2 drainage area and is located in the south
of Espírito Santo state, Brazil. The models performances analysis was made for seven
rivers control sections, considering 1999 to 2013 daily flows records.
Models parameters sensitivity analysis allowed the identification of the most influential
parameters contributing to the models calibration. Both models were calibrated for a
six years data period and then validated by using an independent six years data period.
Simulated and observed daily flows were compared. Average Nash-Sutcliffe efficiency
(NSE and NSElog) and determination (R2) coefficients, considering the seven control
stations located in the Itapemirim basin, for calibration and validation periods,
estimated for MGB simulations were about ten percent higher than those estimated for
SWAT simulations. Considering the classification criteria developed by Moriasi et al.
(2015), these coefficient values indicate that the simulations performed with SWAT
model can be classified as satisfactory and those performed with MGB as good.
However, the percentage errors related mainly to maximum and minimum annual flows
reproductions were high for many control sections. Thus, empirical criteria for model
performance classification, such as the proposed by these authors, should be used
with caution and considering modeling objectives. Differences in model performances
can be mainly attributed to the two models flow generation and production
mechanisms. The results do not allow concluding that either of the hydrological models
is superior to the other for all sites or considering all model performance measures.
There is no consensus in rain-flow modeling research on which model structure should
be preferred, since factors such as model formulation, parameterization, and modeler
ability have large influence on the simulations accuracy

Access to document

Acesso à informação
Transparência Pública

© 2013 Universidade Federal do Espírito Santo. Todos os direitos reservados.
Av. Fernando Ferrari, 514 - Goiabeiras, Vitória - ES | CEP 29075-910