HR: 15:00h
AN: A53C-05 [Abstracts]
TI: Evaluating the Performance of the CPTEC-INPE/FCT-UNESP Troposphere Dynamic Model Using the VRS Concept
AU: * Alves, D B
EM: danibarroca@yahoo.com.br
AF: Sao Paulo State University, Roberto Simonsen, 305, Presidente Prudente, SP , Brazil
AU: Dalbelo, L F
EM: lfdalbelo@gmail.com
AF: Sao Paulo State University, Roberto Simonsen, 305, Presidente Prudente, SP , Brazil
AU: Monico, J F
EM: galera@fct.unesp.br
AF: Sao Paulo State University, Roberto Simonsen, 305, Presidente Prudente, SP , Brazil
AU: Sapucci, L F
EM: lsapucci@cptec.inpe.br
AF: CPTEC-INPE, Rodovia Presidente Dutra, Km 40, Cachoeira Paulista, SP , Brazil
AB:
Nowadays, the use of Zenithal Tropospheric Delay (ZTD) prediction from Numeric Weather Prediction (NWP)
models is a good alternative to minimize the effects of the troposphere in the radio frequency signs for real time
and/or pos-processed applications. This process is denominated here after ZTD dynamic modeling. In Brazil, the
procedure used to compute the ZTD by the NWP model was jointly developed by researchers from UNESP
(São Paulo State University) and CPTEC (Center for Weather Forecasting and Climate Studies) of the INPE
(National Institute for Space Research). The ZTD values are provided for all South America twice a day with
predictions for a period of 66 hours. The database and the quality analysis are available by CPTEC-INPE in
http:satelite.cptec.inpe.br/htmldocs/ztd/zenital.htm. In order to test the performance of the ZTD dynamic modeling
in positioning applications, some experiments were carried out. Besides, the results obtained with dynamic
modeling were compared with those obtained by Hopfield empirical model. These two tropospheric models were
used to generate a VRS (Virtual Reference Station). In the VRS concept developed in this research, a reference
station is generated near the user using data provided by a reference network station and atmospheric models.
Therefore, the VRS data are not provided by a real receiver. But, the idea is that the VRS data resemble as much
as possible a real receiver data at the same location. Therefore, the user has the possibility of using the VRS as if
it were a real reference station in your proximities, and to accomplish the relative positioning with a single
frequency receiver. This method was implemented in an software which has been developed at UNESP. In order
to test the discussed method it was accomplished two experiments using data from two different networks: (a)
Brazilian Continuous GPS Network (RBMC) and some extra stations; (b) GPS Active Network of West of São
Paulo State. Using the first network it was processed 10 days of data collected in 2005 May. The results obtained
by the dynamic modeling were better than those obtained by Hopfield model. It was obtained, on average, an
improvement of 19% using the dynamic modeling instead of the Hopfield one. Using the second network it was
processed 4 days of data collected in 2006 December. But, in this case, the results obtained were quite similar.
This probably happened because in this period occurred a problem with the dynamic modeling. Some data from
radiosondes launched over South America were not assimilated during this period deteriorating the ZTD
predictions. This fact is important because shows how the assimilation process is important to the ZTD dynamic
modeling and revels that this relationship between assimilation data and ZTD prediction quality must be better
investigated. But, in spite of this problem, in general the results from dynamic modeling are still good if compared
with empirical model.
DE: 1220 Atmosphere monitoring with geodetic techniques (6952)
DE: 1241 Satellite geodesy: technical issues (6994, 7969)
DE: 1610 Atmosphere (0315, 0325)
DE: 3238 Prediction (3245, 4263)
DE: 7959 Models
SC: Atmospheric Sciences [A]
MN: 2007 Joint Assembly