In order to assess the impact of ground-based GPS water vapour estimates in
meteorological forecasting applications, estimates must be available withi
n near real time and must be produced with an accuracy comparable with that
of existing meteorological measurement techniques. Before the potential be
nefit of GPS-integrated water vapour (IWV) estimates in numerical weather p
rediction (NWP) models could be determined, extensive tests were performed
to examine the effect of various process error models, constraints and esti
mation techniques involved within GPS processing procedures in order to dev
ise a recommended processing strategy. Ten months of GPS data from a select
ion of stations, which are co-located with radiosonde launch sites, within
the UK permanent GPS network, have been processed to validate these finding
s. Near real-time estimation issues have subsequently been examined, primar
ily using a sliding-window processing approach to demonstrate the feasibili
ty of producing accurate GPS IWV estimates within the required time frame.
Additionally, preliminary simulation work has been performed to assimilate
GPS IWV into the UK Meteorological Office (UKMO) numerical weather predicti
on models, in preparation for impact assessments on the quality of mesoscal
e weather forecasts. Crown Copyright (C) 2001 Published by Elsevier Science
Ltd. All rights reserved.