Limiting factors in tropospheric propagation delay error modelling for GPS airborne navigation

We examine the likely accuracies to which the tropospheric propagation delay at the zenith can be modelled in airborne navigation and the limiting effects this will ultimately have on an aircraft's position determination. Even though the hydrostatic component of the tropospheric zenith delay can be modelled to millimetre accuracy, this requires an accurate atmospheric pressure measurement. In aircraft navigation this may not be directly available and the pressure value used to drive the model must be derived from another source. Whatever average value is used, it is unlikely to represent the real pressure exactly and will therefore introduce a bias into the delay determination. To observe the effects of the neutral atmosphere on GPS signals, dual frequency (L1 and L2) GPS data was collected with Ashtech Z-12 receivers at a surveyed base station and simulataneously on board an aircraft. Kinematic position solutions for the aircraft have been computed from the data sets. The effects of the troposphere have been examined by looking at differences in position solutions using different tropospheric delay models. Our results indicate that the water vapour component of the tropospheric delay provides the limiting factor close to the earth's surface. If the value used for water vapour pressure is incorrect compared to the ''true'' value by the order of 10 mbar, then in the absence of other errors, the resulting zenith delay will be in error by the order of 10 cm Towards the tropopause however, incorrect scaling of the total surface pressure provides the limiting factor.