@ElectronicSource{MendonçaBona:2009:ExEOPe,
abstract = "The impact of modifications of the perturbation method based on
empirical orthogonal functions (EOF method) used operationally
upon the ensemble prediction system (EPS) at CPTEC/INPEis
evaluated. The main changes proposed in this study are: to apply
the EOF method to perturb the midlatitudes; apply additional
perturbations to the surface pressure (P) and specific humidit (Q)
fields; and, compute regional perturbations over South America.
The impact of these modifications in the characteristics of the
initial perturbations and in the quality of the EPS forecasts is
investigated. The EPS forecasts are evaluated through average
statistical scores over the period 15 December 2004 to 15 February
2005. The statistical scores used in the evaluation are pattern
anomaly correlation, root mean square error, ensemble spread,
Brier skill score and pertubation versus error correlation
analysis (PECA). Results indicate that will the inclusion of
perturbations on P and Q, EOF-based perturbations acquire a more
baroclinic structure. It is also observed that the simultaneous
applications of additional perturbations both in the extratropics
and to the P and Q fileds improves the performance of CPTEC-EPS
and enhances the quality of forecasts perturbations. Moreover,
regional EOF-based perturbations computed over South America have
positive impact on the ensemble forecasts over the target
region.",
address = "S{\~a}o Jos{\'e} dos Campos",
affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE/CPTEC)} and
{Instituto Nacional de Pesquisas Espaciais (INPE/CPTEC)}",
author = "Mendon{\c{c}}a, Antonio Marcos and Bonatti, Jos{\'e} Paulo",
keywords = "*.",
language = "en",
lastupdatedate = "2008-08-26",
publisher = "Instituto and Nacional and de and Pesquisas and Espaciais",
ibi = "6qtX3pFwXQZGivnK3Y/UQEuP",
url = "http://urlib.net/ibi/6qtX3pFwXQZGivnK3Y/UQEuP",
targetfile = "v1.pdf",
title = "Experiments with EOF-Based perturbation methods and their Impact
on the CPTEC/INPE Ensemble Prediction System",
typeofmedium = "On-line",
year = "2009",
urlaccessdate = "08 maio 2024"
}