@InProceedings{NascimentoHerdSouzAnge:2010:EvPrOv,
author = "Nascimento, M G do and Herdies, Dirceu Luis and Souza, Diego
Oliveira de and Angelis, Carlos Frederico",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto
Nacional de Pesquisas Espaciais (INPE)} and {Instituto Nacional de
Pesquisas Espaciais (INPE)} and {Instituto Nacional de Pesquisas
Espaciais (INPE)}",
title = "Evaluation of precipitation over South America in the reanalysis
MERRA",
booktitle = "Abstracts...",
year = "2010",
organization = "The Meeting of the Americas.",
publisher = "AGU",
keywords = "hydrological cycles, budgets, precipitation, climatology, data
assimilation.",
abstract = "The reanalysis data are an essential tool in the study of weather
and climate variability in the past 15 years. The reanalysis are
nothing more than a merger of a large amount of observational data
with data from numerical modeling. One of the key utilities in a
reanalysis is that the output generated from the model physics
provides data not easily observed, but is consistent with the
analyzed observed data. Knowing that precipitation is one of the
critical components of the balance of moisture and energy, and
that this variable on reanalysis data is highly correlated with
the model physical parameterizations, this work attempts to
quantify the uncertainty of rainfall data between the MERRA
reanalysis and data of precipitation obtained from the GPCP and
TRMM for the period 1999 to 2008. For a better understanding of
the observed data, TRMM and GPCP data were compared, where make
clear a large spatial correlation (0.93) between these two
datasets, with major differences in the tropical region, where
TRMM data have higher values on the Amazon river mouth, equatorial
Pacific Ocean and the Caribbean. On the whole South America this
analisys showed an average error of -0.33 mm/day, showing that the
values observed by TRMM are higher than GPCP. Analysis of the
MERRA data for the entire period showed a good spatial correlation
between the reanalysis data and TRMM (0.77) and GPCP (0.78). It
was also noted that the MERRA data tend to overestimate the values
of precipitation over much of the equatorial region, mainly on the
tropical Atlantic Ocean, and to underestimate the values of
precipitation over southern Brazil. When analyzed seasonally it
was found that the MERRA data had lower spatial correlation in the
seasons that show more convective activity, mainly in summer (0.75
TRMM and GPCP 0.76) and spring (0.72 TRMM and GPCP 0.73). In all
seasons, especially in spring and summer, it was observed that the
MERRA data have difficulties in representing the values of
precipitation over southern Brazil and the La Plata Basin
(underestimating the values) and over the region of tropical
Atlantic Ocean (overestimating the values). More detailed studies
should be conducted to better estimate the errors presented in
this work, but the results make clear that the use of MERRA data
for the southern and northern South America deserve more
attention.",
conference-location = "Foz do Igua{\c{c}}u, BR",
conference-year = "8-12 aug 2010",
language = "en",
urlaccessdate = "28 mar. 2024"
}