Improved GPS tropospheric path delay estimation using variable random walk process noise

Area of Science:

• Geodesy and Geodynamics
• Atmospheric Sciences
• Satellite Navigation

Background:

• Accurate Global Positioning System (GPS) positioning necessitates precise modeling of tropospheric delay.
• Inadequate tropospheric delay variation can lead to systematic errors in position estimates, particularly vertical displacement.
• Conversely, overly loose constraints amplify noise, reducing overall precision.

Purpose of the Study:

• To investigate optimal tropospheric constraints within the GipsyX software for improved GPS positioning.
• To evaluate the impact of random walk process noise on zenith wet delay (ZWD) estimation.
• To enhance the accuracy of 24-hour GPS solutions and precise orbit parameters.

Main Methods:

• Utilized the variability of 5-minute GPS estimated positions as a proxy for tropospheric error.
• Analyzed the effect of different ZWD constraint values (random walk process noise) on position estimates.
• Compared 5-minute weighted mean positions with 24-hour solutions to assess the impact on daily products.

Main Results:

• The default ZWD constraint of 3 mm/√(hr) produced spurious wave-like vertical displacements (~100 mm) during Winter Storm Ezekiel.
• Loosening the ZWD constraint to 6-12 mm/√(hr) suppressed these spurious signals, reduced vertical variability, and improved water vapor estimates.
• Regional or station-specific optimization of constraints further enhanced results.

Conclusions:

• Loosening the default ZWD constraint from 3 mm/√(hr) to at least 6 mm/√(hr) for 5-minute data intervals is recommended.
• Optimal global ZWD constraints typically fall within the 6-12 mm/√(hr) range.
• Constraint values should be scaled for different data intervals using the formula √(x/300).