2:45 PM - 3:00 PM
[SGD01-17] Optimization of water vapor analysis using ship-borne GNSS measurement
Keywords:GNSS, Kinematic positioning, Water vapor
1.Vertical positioning error and PWV error
GNSS analysis is affected by various factors such as interference with reflected waves (multipath), clock error, antenna phase center variation (PCV). It is expected that the kinematic positioning which treats the three-dimensional position of each GNSS antenna as a time variation parameter will be more affected than the static positioning. It is said that the error of vertical coordinates and the error of precipitable water (PWV) are inversely proportional (Beutler et al. 1988, Shoji et al. 2000). In the statistical comparison for a GEONET station "02P212" for eight months from May to December 2019, when there was a vertical positioning error of about 7 cm, a PWV error of about 1 mm occurred (Figure not shown). In this study, we examined the relationship between PWV error and vertical positioning error, and considered the optimum method for analyzing PWV from ship-borne GNSS measurements.
2.Vertical positioning error in kinematic GNSS analysis
We performed 10Hz GNSS observation at a fixed GNSS station at the Meteorological Research Institute for 16 days from December 1 to 16, 2020, performed kinematic positioning and static positioning while changing analysis frequency and analysis duration, and compared the vertical coordinates. As a result, it was found that the vertical coordinates by kinematic positioning tend to be higher than the static positioning solution within the range of about 1 to 2 cm. It was also found that the longer the analysis time and the higher the analysis frequency, the stronger the tendency.
3.Comparison of PWV obtained from ship-borne GNSS measurements with JMA’s meso-scale objective analysis and radiosonde observation
From October 20th to November 22nd, 2020, we carried out 10Hz GNSS observations on the two cargo ships "Wakanatsu" and "Ryunan". RTKLIB ver.2.4.2 was used for GNSS analysis. GNSS analysis was performed by changing the analysis frequency in the range of 0.1 to 30 seconds and increasing the analysis time from 0.5 to 8.5 hours in 1-hour units. As a result of comparison with PWV expressed in JMA’s operational meso-analysis (MA) (Figure 1), the tilt of the regression line tended to become smaller as the analysis time became longer. The lower the analysis frequency, the stronger the tendency. When the analysis time is less than 1 hour, the convergence of the solution is insufficient and the standard deviation (SD) becomes large. Considering BIAS and SD, it can be said that among the tried conditions, the 1.5-hour analysis at 2-second intervals has the highest degree of agreement with MA. We also compared GNSS PWV obtained by 2-sec interval analysis adopted on a GNSS observation data recorded on an observation vessel “Ryo-fu Maru” with radiosonde observations launched from the vessel. The slope of the regression line, BIAS, RMS, and SD (standard deviation), all showed a slightly higher degree of agreement in the 1.5-hour analysis than the 8-hour analysis, which shares the similar characteristics with the comparison against MA.
Acknowledgement
We thank "RYUKYU KAIUN KAISHA" for their support on “Wakanatsu” observation and "KAGOSHIMA NIYAKU KAIRIKU UNYU" and "MINAMINIHON KISEN" for their support on “Ryuunan” observation.
This study was partly supported by JSPS KAKENHI Grant Number 20H02420.
RTKLIB ver. 2.4.2 (Takasu 2013) was used for GNSS analysis.
GEONET observation data were acquired from the ftp server of the Geospatial Information Authority of Japan (GSI) in RINEX format.
MADOCA real-time prod- uct was provided by JAXA via the Internet (https://ssl.tksc.jaxa.jp/madoca/ public/public_index_en.html).
References
Beutler, G., I. Bauersima, W. Gurtner, M. Rothacher, T. Schildknecht, and A. Geiger, Atmospheric refraction and other important biases in GPS carrier phase observation, in Atmospheric Effects on Geodetic Measurements, Monograph, 12, pp. 15–43, School of Surveying, Univ. of New South Wales, Kensington, Australia, 1988.
Shoji, Y., H. Nakamura, K. Aonashi, A. Ichiki, H. Seko, and Members of GPS/MET Japan summer Campaign 1997 in Tsukuba (2000): Semi-diurnal and diurnal variation of errors in GPS precipitable water vapor at Tsukuba, Japan caused by site displacement due to ocean tidal loading, Earth Planets Space, 52, 685-690.
Takasu T (2013) RTKLIB 2.4.2 manual. http://www.rtklib.com/prog/ manual_2.4.2.pdf.
GNSS analysis is affected by various factors such as interference with reflected waves (multipath), clock error, antenna phase center variation (PCV). It is expected that the kinematic positioning which treats the three-dimensional position of each GNSS antenna as a time variation parameter will be more affected than the static positioning. It is said that the error of vertical coordinates and the error of precipitable water (PWV) are inversely proportional (Beutler et al. 1988, Shoji et al. 2000). In the statistical comparison for a GEONET station "02P212" for eight months from May to December 2019, when there was a vertical positioning error of about 7 cm, a PWV error of about 1 mm occurred (Figure not shown). In this study, we examined the relationship between PWV error and vertical positioning error, and considered the optimum method for analyzing PWV from ship-borne GNSS measurements.
2.Vertical positioning error in kinematic GNSS analysis
We performed 10Hz GNSS observation at a fixed GNSS station at the Meteorological Research Institute for 16 days from December 1 to 16, 2020, performed kinematic positioning and static positioning while changing analysis frequency and analysis duration, and compared the vertical coordinates. As a result, it was found that the vertical coordinates by kinematic positioning tend to be higher than the static positioning solution within the range of about 1 to 2 cm. It was also found that the longer the analysis time and the higher the analysis frequency, the stronger the tendency.
3.Comparison of PWV obtained from ship-borne GNSS measurements with JMA’s meso-scale objective analysis and radiosonde observation
From October 20th to November 22nd, 2020, we carried out 10Hz GNSS observations on the two cargo ships "Wakanatsu" and "Ryunan". RTKLIB ver.2.4.2 was used for GNSS analysis. GNSS analysis was performed by changing the analysis frequency in the range of 0.1 to 30 seconds and increasing the analysis time from 0.5 to 8.5 hours in 1-hour units. As a result of comparison with PWV expressed in JMA’s operational meso-analysis (MA) (Figure 1), the tilt of the regression line tended to become smaller as the analysis time became longer. The lower the analysis frequency, the stronger the tendency. When the analysis time is less than 1 hour, the convergence of the solution is insufficient and the standard deviation (SD) becomes large. Considering BIAS and SD, it can be said that among the tried conditions, the 1.5-hour analysis at 2-second intervals has the highest degree of agreement with MA. We also compared GNSS PWV obtained by 2-sec interval analysis adopted on a GNSS observation data recorded on an observation vessel “Ryo-fu Maru” with radiosonde observations launched from the vessel. The slope of the regression line, BIAS, RMS, and SD (standard deviation), all showed a slightly higher degree of agreement in the 1.5-hour analysis than the 8-hour analysis, which shares the similar characteristics with the comparison against MA.
Acknowledgement
We thank "RYUKYU KAIUN KAISHA" for their support on “Wakanatsu” observation and "KAGOSHIMA NIYAKU KAIRIKU UNYU" and "MINAMINIHON KISEN" for their support on “Ryuunan” observation.
This study was partly supported by JSPS KAKENHI Grant Number 20H02420.
RTKLIB ver. 2.4.2 (Takasu 2013) was used for GNSS analysis.
GEONET observation data were acquired from the ftp server of the Geospatial Information Authority of Japan (GSI) in RINEX format.
MADOCA real-time prod- uct was provided by JAXA via the Internet (https://ssl.tksc.jaxa.jp/madoca/ public/public_index_en.html).
References
Beutler, G., I. Bauersima, W. Gurtner, M. Rothacher, T. Schildknecht, and A. Geiger, Atmospheric refraction and other important biases in GPS carrier phase observation, in Atmospheric Effects on Geodetic Measurements, Monograph, 12, pp. 15–43, School of Surveying, Univ. of New South Wales, Kensington, Australia, 1988.
Shoji, Y., H. Nakamura, K. Aonashi, A. Ichiki, H. Seko, and Members of GPS/MET Japan summer Campaign 1997 in Tsukuba (2000): Semi-diurnal and diurnal variation of errors in GPS precipitable water vapor at Tsukuba, Japan caused by site displacement due to ocean tidal loading, Earth Planets Space, 52, 685-690.
Takasu T (2013) RTKLIB 2.4.2 manual. http://www.rtklib.com/prog/ manual_2.4.2.pdf.