Table 3 Data processing strategy

Model

DF, FF1, FF2, FF3, FF4, FF5

Observations

Code and phase observations on B1C, B2a, B1I, B3I, B2 from BDS-3

Sampling interval

30 s

Satellite elevation mask

10°

Estimator

Kalman filter

Weighting scheme

Elevation-dependent weight; 6 mm and 0.6 m for raw phase and code, respectively

Satellite orbit and clock

Products from GFZ analysis center

Satellite DCB corrections

Corrected with MGEX DCB products

Satellite PCO and PCV

PCO is corrected with the igs14.atx and PCV is not applied

Receiver PCO and PCV

PCV is corrected with the igs14.atx and PCO is not applied

Tide effect

Solid Earth, pole and ocean tide (G. Petit and B. Luzum, 2010)

Relativistic effect

Corrected (Leick, A. et al., 2015)

Phase windup

Corrected (Wu J.T., 1992)

Earth rotation

Corrected (G. Petit and B. Luzum, 2010)

Station coordinates

Static: estimated as constants; simulated kinematic: estimated as white noise process

Receiver clock

Estimated as white noises

Tropospheric delay

Zenith Hydrostatic Delays (ZHD) are corrected with Saastamoinen model, and Zenith Wet Delays (ZWD) are estimated as random walk (1 × 10^–9 m²/s) (J. Saastamoinen, 1973)

Ionospheric delay

Estimated as random walk (1 × 10⁴ m²/s) only in FF1 (Su & Jin, 2019)

Receiver inter-frequency bias

Absorbed by receiver clock in FF5 models or estimated as constants in other models

Ambiguities

Estimated as constant