Evaluation of pressure extracted from NCEP and CMC global numerical weather prediction models against in-situ and GPT pressure

An earlier investigation by Urquhart et al (2011) demonstrated that ray traced hydrostatic zenith delays from NCEP's Re-Analysis I (NCEP) dataset proved to exhibit higher variability when compared to those from the Canadian Meteorological Centre's (CMC) Global Deterministic Prediction System (GDPS) and the European Centre for Medium Range Weather Forecasting (ECMWF). This investigation expands on the original by analyzing the variation of the zenith hydrostatic delay (as ray traced through the NWP) and the extracted pressure at the surface for 35 IGS reference stations for the entire year of 2010. Two NWP's were selected, NCEP's Re-Analysis I and CMC's GDPS. NCEP was selected since it forms the basis for the UNB-VMF1 service, and CMC's GDPS was selected due to availability. Both models have global coverage, but NCEP's grid resolution is 2.5 x 2.5 degrees as compared to CMC's (GDPS) 0.6 x 0.6 degrees. The location within the NWP is defined by the location of the IGS reference stations. The position of the reference stations are defined by the IGS weekly solutions where week 52's coordinate values for the year 2010 were used. The height was then adjusted by the defined meteorological sensor offset as defined the IGS stations's respective log. The investigation is based on the following comparisons: 1. Extracted Pressure from the NWP (NCEP and CMC (GDPS)) compared to measured pressure from the site, and also from pressure derived from GPT model. 2. Raytraced hydrostatic zenith delay compared to the Saastamoinen hydrostatic zenith computed from the measured site pressure. Results indicate good agreement between pressure extracted from the NWP and in-situ pressure and larger differences with respect to GPT.