Optimization of GPS water vapor tomography technique with radiosonde and COSMIC historical data

Abstract

Abstract. The near-real-time high spatial resolution of atmospheric water vapor distribution is vital in numerical weather prediction. GPS tomography technique has been proved effectively for three-dimensional water vapor reconstruction. In this study, the tomography processing is optimized in a few aspects by the aid of radiosonde and COSMIC historical data. Firstly, regional tropospheric zenith hydrostatic delay (ZHD) models are improved and thus the zenith wet delay (ZWD) can be obtained at a higher accuracy. Secondly, the regional conversion factor of converting the ZWD to the precipitable water vapor (PWV) is refined. Next, we develop a new method for dividing the tomography grid with an uneven voxel height and a varied water vapor layer top. Finally, we propose a Gaussian exponential vertical interpolation method which can better reflect the vertical variation characteristic of water vapor. GPS datasets collected in Hong Kong in February 2014 are employed to evaluate the optimized tomographic method by contrast with the conventional method. The radiosonde-derived and COSMICderived water vapor densities are utilized as references to evaluate the tomographic results. Using radiosonde products as references, the test results obtained from our optimized method indicate that the water vapor density accuracy is improved by 15 and 12 % compared to those derived from the conventional method below the height of 3.75 km and above the height of 3.75 km, respectively. Using the COSMIC products as references, the results indicate that the water vapor density accuracy is improved by 15 and 19 % below 3.75 km and above 3.75 km, respectively.

Assessing water vapor tomography in Hong Kong with improved vertical and horizontal constraints

“…The SWV was then obtained using the Niell mapping function (Niell, 1996) and the calibrated humidity conversion coefficient. The WVLT was determined as 9.5 km from COSMIC historical data and Ye et al's (2016) method. Following the tomography model proposed by Flores et al (2000), we estimated the 3-D water vapor distribution using the GPS tomography technique with the horizontal constraint from Eq.…”

“…Currently, there is no accurate model function to fit the spatial distribution of water vapor. The vertical constraint of atmospheric tomography can be obtained using an exponential model (Jiang et al, 2014;Ye et al, 2016), as follows:…”

“…Later, Wang et al (2011) established the T m model in Hong Kong using radiosonde products. Ye et al (2016) also assessed the relationship between T m and surface temperature based on radiosonde and COS-MIC products. However, these three models only consider the parameter of surface temperature.…”

“…Traditionally, Eq. (14) was used as a vertical constraint and the parameter H we could be obtained using COSMIC or radiosonde historical data products (Ye et al, 2013(Ye et al, , 2016. Due to H we changes over time are obvious, so they need to be obtained once for each tomography epoch.…”

A new approach for GNSS tomography from a few GNSS stations

“…The size and location of the rectangular cubes are determined based on the distribution of GNSS signals only on the top boundary of the tomographic field -the rectangular cube that best fits the top boundary is adopted (Bastin et al, 2005;Bender et al, 2009;Champollion et al, 2005;Ding et al, 2017;Gradinarsky and Jarlemark, 2004;Hoyle, 2005;Rohm et al, 2014;Seko et al, 2000;Troller et al, 2006;Xia et al, 2013;Ye et al, 2016). In fact, the field that GNSS signals cover has roughly the shape of an upside-down cone, meaning that in the part near the edge of the cube, especially in the lower part, none of the GNSS signals cross through.…”