Estimation of precipitable water vapour from GPS measurements in Argentina: Validation and qualitative analysis of results

• Published in: Advances in space research Vol. 46; no. 7; pp. 879 - 894
• Main Authors: Fernández, L.I., Salio, P., Natali, M.P., Meza, A.M.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01-10-2010; Elsevier
• Subjects: Astronomy; Earth, ocean, space; Exact sciences and technology; External geophysics; Global Positioning System; Mean tropospheric temperature; Precipitable water vapour; Argentina; Argentina, Salta; Argentina, Buenos Aires; Argentina, Cordoba; Precipitable water vapour; Mean tropospheric temperature; Global Positioning System; Validation; NOAA satellite; Numerical integration; Bias; Northern hemisphere; Troposphere; Calibration; Space research; Polar orbiting satellite; Regional study; Extreme value; Steam; Precipitable water; Atmospheric precipitations; Models; Dry spell; Radio sounding
• ISSN: 0273-1177; 1879-1948
• DOI: 10.1016/j.asr.2010.05.012

This paper presents PWV estimates from GPS data computed at four continuously operated GPS stations in Argentina established at Buenos Aires, Córdoba, Rosario and Salta over a 1 year period (2006–2007). The objective is to analyze the behaviour of the GPS PWV estimation using mean tropospheric temperature (
T
m
) values from the Bevis model, Sapucci model and obtained by a numerical integration of variables provided by the operational analysis of the National Centre of Environmental Prediction (NCEP). The results are validated using PWV values from nearest radio soundings. Moreover, a comparison between PWV values determined from microwave sensors deployed on the NOAA-18 satellite and PWV from GPS observations is also presented.
From the analysis we can see that the computation of GPS PWV using the
T
m
from the Bevis model, originally deduced for the northern hemisphere, shows similar behaviour to the respective computation using a Sapucci model inside 0.5
mm. The differences between the
T
m
values computed from the Sapucci model and the numerical integration of NCEP variables are of the order of 15
K, although it does not represent a significant error in PWV.
Nevertheless, differences in bias are imperceptible during the dry period and they are as big as 3
mm during the moist or high precipitation period. This behaviour could not represent an improvement when comparing radio soundings with respect to the GPS PWV values using different estimations of
T
m
. Thus, we conclude that the usage of
T
m
estimated from the Bevis model is the best choice for regional studies, considering the simplicity and dissemination of the method, unless some more studies taking into account the geographical and climatological characteristic of the region are performed. As expected, GPS PWV values show very good agreement with radio sounding determinations, small differences can be observed especially during extreme precipitation periods. In general the NOAA PWV values denote an over estimation of the available water vapour. It is important to note that the determination of PWV is not the mean product of the NOAA-18 satellite mission. These results show the potential of such a product although a preliminary calibration using GPS PWV is still necessary.