Effects and disturbances on GPS-derived zenith tropospheric delay during the CONT08 campaign

Abstract

The zenith total delay (ZTD) can be retrieved from space geodetic techniques, e.g., Global Positioning System (GPS) and Very Long Baseline Interferometry (VLBI), which plays a key role in climatological and atmospheric sciences. However, ZTD estimates still have lots of effects and uncertainties, particularly in GPS model errors. The continuous VLBI observations provide an opportunity to assess GPS ZTD estimates during the Continuous VLBI Campaign 2008 (CONT08) at 11 co-located stations from August 12 to 26, 2008. In this paper, the effects on GPS ZTD estimate and its disturbances are investigated using different mapping function models (GMF, NMF, and VMF1), Phase Center Variation (PCV) models (AZEL and ELEV) and Ocean Tide Loading (OTL) models (FES2004, CSR4.0 and GOT00). It has shown that the ZTDs from VLBI and GPS have an agreement in –3.88–3.74 mm with correlation coefficients of higher than 0.87. For mapping function models, there are no obvious differences, while the PCV model of ELEV is always a little better than AZEL for large scale network with mixed antenna types. For stations near to the coastlines, ocean loading effects must be corrected. While for short period, the effects with OTL models of FFES2004, CSR4.0 and GOT00 are always at the same level. In addition, significant diurnal cycles S₁ (24 h period) and semidiurnal cycles S₂ (12 h period) of GPS ZTD are found with amplitudes between 0.82 and 13.84 mm and 0.30 and 5.23 mm, respectively, which are closer to VLBI ZTD estimates. The correlation coefficients between VLBI and GPS ZTD are 0.85 and 0.95 in S₁ and S₂, respectively.

Introduction

The tropospheric delay is one of the major error sources of space geodetic techniques while their radio signals propagate through the atmosphere, e.g., Global Positioning System (GPS) and Very Long Baseline Interferometry (VLBI). Nowadays the total zenith tropospheric delay (ZTD) can be determined by GPS and VLBI through mapping function (e.g., Niell, 1996, Behrend et al., 2000, Niell et al., 2001, Pacione et al., 2002, Snajdrova et al., 2006), which plays an important role in climatological and atmospheric sciences. The GPS and VLBI-derived ZTDs are the integrated refractivity in the zenith direction, and can be expressed as the sum of the zenith hydrostatic delay (ZHD) related to the surface pressure (Elgered et al., 1991), and the zenith wet delay (ZWD) related to the water vapor (Jin and Luo, 2009).

A number of studies have been carried out on the effects and accuracies of ZTD derived from GPS, VLBI and other ground-based techniques. For example, Behrend et al., 2000 analyzed GPS and VLBI data for 2 weeks in December 1996 at Spain, and found that ZTD differences were smaller than 1 cm. Snajdrova et al. (2006) analyzed continuous VLBI data for 15 days during the Continuous VLBI Campaign 2002 (CONT02) and found that the ZTD differences were about 3 to 10 mm. Meanwhile, a lot of works have also demonstrated that VLBI can provide ZTD with high precision and accuracy for meteorological and climatological applications (e.g., Niell et al., 2001, Hatanaka et al., 2001, Heinkelmann et al., 2007). Therefore, the ZTD determined by VLBI can be used as an independent and high-accuracy reference to assess the accuracy and reliability of GPS-estimated ZTD.

Although GPS can provide precise and high temporal resolution ZTD as a highly precise, continuous, all-weather and near-real-time technique, there are lots of effects on GPS ZTD estimates, e.g., mapping functions (Boehm et al., 2007, Won et al., 2010) and Ocean Tide Loading (OTL) models (Vey et al., 2002, Vey et al., 2006, Tregoning and Herring, 2006, Tesmer et al., 2007). For example, Won et al. (2010) analyzed GPS ZTD by testing GMF, NMF, and VMF1 models and found the maximum difference occurred in February and August. Fund et al. (2011) processed 1-year GPS data with different mapping functions, and found significant differences between VMF1 and GMF models due to the GMF’s low spatial resolution. Vey et al. (2002) investigated the effects of ocean loading on GPS ZTD and concluded that unmodeled ocean loading has significant effects on GPS ZTD, which must be properly corrected for GPS ZTD estimating. The Continuous VLBI Campaign 2008 (CONT08) was a follow-on campaign of the CONT94, CONT95, CONT96 CONT02 and CONT05 with 11 co-located stations equipping with GPS receivers (Fig. 1). The goal of CONT08 was to acquire state of the art VLBI data over a two-week period (August 12 to 26, 2008) and to provide the highest accuracy of VLBI products, such as the Earth orientation parameters (Nilsson et al., 2010) or tropospheric delay (Teke et al., 2011). Therefore, the independent VLBI provides a unique chance to investigate the reliability and effects of GPS ZTD estimates. In addition, the ZTD has significant diurnal and semidiurnal oscillations, but with a number of possible effects or unknown factors, e.g., ocean tides (Vey et al., 2002) or atmospheric tides (Jin et al., 2008). In this paper, the effects and disturbances on GPS ZTD with different models are investigated, including mapping function models, Phase Center Variation (PCV) models and OTL models. The comparisons between ZTDs derived from VLBI and GPS are performed as well as surface pressure data. A detailed data analysis and post-processing are shown in Section 2. Section 3 presents the results and discussions, including the relations between VLBI/GPS ZTD and altitude, the effects on GPS ZTD and Precipitable Water Vapor (PWV) with different models in GPS data processing, diurnal and semidiurnal cycles of VLBI/GPS ZTD. Finally, conclusions are given in Section 4.