Effect of Observing Session Duration on Static GPS Velocity Estimation

The effect of the observing session duration on static GPS positional accuracies has previously been detected. This study further investigates this effect on the estimated velocities. This is especially important for tectonic/deformation monitoring studies that use GPS observations shorter than 24 h. A sample has been made from 13 IGS stations that cover the earth globally. GIPSY research software has been used for the processing. GPS velocities have been estimated from GPS campaigns using the time series of 15-year length on average. The sampling rate was chosen as one observation per year as still used by some of the tectonic studies today. The solutions from 24 h data were taken as the truth and the velocities estimated from the shorter sessions were assessed. We show how the coefficient of multiple determination R-squared can successfully be used to assess the accuracy of GPS velocities from campaign measurements. Results based on this assessment indicate that vertical velocities are affected most when observation sessions shorter than 24 h are used. An accuracy improvement of 12% on average on the vertical velocity is gained as the observation session is extended from 8 h to 12 h. Individual solutions show that this improvement can be as high as 32%. Similarly, the improvement in horizontal velocities is 3% on average but it could be as high as 13% for some stations. A further amount of improvement at similar levels can only be obtained when observing session duration is extended from 12 h to 18 h which is not practical when one needs to use the advantage of the day light. This suggests us that the lowest limit for the session length of GPS campaigns needs to be 12 h rather than 6-8 h which have typically been applied conventionally by many research groups up to date. Using 12 h of GPS observations, 90% of the vertical deformation pattern is recovered. The percentages for the same duration are 96 and 100 for the longitudinal and latitudinal deformations respectively.