Mitigating Differential Troposphere Effects for Gps-based Volcano Monitoring

The UNSW-designed GPS volcano monitoring system installed on Gunung Papandayan seeks to characterise the ground deformation within the volcanic edifice. Hourly baseline solutions produce a time series of data designed to detect any ground motion. Differential tropospheric effects between GPS receivers separated by a large change in height (in this case up to 1400m) impact on the quality of this time series data, particularly in the case of the vertical component. This increased variability constrains reliable ground deformation detection to the decimetre level. The Saastamoinen troposphere model is recommended as the most suitable to account for the hydrostatic (dry) part and non-hydrostatic (wet) part of the delay. A time series of data from the SAGE (New Zealand) network is processed using Saastamoinen modelling for varying session lengths. This time series is re-processed using the Saastamoinen model as an a-priori model and estimating one additional residual relative zenith delay parameter per session during the baseline computation. GPS baseline processing is more of a challenge in this case due to the short session length and reduced number of observations from a single-frequency GPS system. In comparison to the modelling approach, the parameter estimation technique is not affected by changes in local weather (which manifest as short-term trends in the height