G2.34

G2.34    Uncertainties of ZTD for GNSS-PW, given by a 3rd party without full traceability

Gap detailed description

The Zenith Total Delay uncertainty is a key component of the total uncertainty in GNSS-PW measurements (Ning et al., 2016). If it is not handled in a proper way, it may drastically affect the GNSS-IPW uncertainty estimate. Fixing it equal to 4mm is just a compromise, excluding outliers from longer time series.

When discussing GRUAN GNSS-IPW uncertainties, we only discuss data analysis using Precise Point Positioning (PPP) in the EPOS software package. While suggesting GRUAN GNSS-IPW uncertainties should be implemented by other data analysis centres, we talk about implementing GNSS-IPW uncertainty analysis method as described by T. Ning et al. (AMT, 2016) in different software (i.e. not EPOS, solely used by GFZ and GRUAN data analysis).  This task is not trivial, for example, the orbital error components described by J. Dousa (GPS Solutions, 2010) and used by T. Ning et al in AMT 2015 are not delivered for end users like ZTDs from IGS (or simply obtainable from standard software for GNSS-data analysis). 

Preliminary analysis has been made (and is still in progress) on documentation and related articles published by the developers of Bernese and GAMIT/GLOBK software.   ZTD uncertainty is known as a main contributor to the GNSS-IPW uncertainty budget. Therefore, it is essential to know and to find recommendations when using uncertainty estimates obtained by different data processing software for making GRUAN-type uncertainty analysis. The goal is to investigate at least two geodetic software packages using the same GNSS-data processing method, comparing the error definition and error handling, leading to (often remarkably) different numeric values of uncertainty estimates. 

Activities within GAIA-CLIM related to this gap

It has been discussed and agreed that in the GAIA-CLIM time-frame we’ll concentrate solely on GRUAN GNSS-PW uncertainty assessment. This restricts the VO user to GRUAN GNSS-PW data only, but gives a possibility to describe the traceability chain and uncertainty estimation in a consistent way, compared to all other instruments within this project and GAID.

Comparing the results offered by different parties and processed with different software (or, even while processed with the same software, but by a different operator using different initial settings) is not as straightforward as it could be expected.

Gap remedy(s)

The gap remedy actions will continue with definitions of "GRUAN GNSS-PW uncertainties" at the level GFZ has reached with their data processing and uncertainty estimation thus far as described in Ning et al. (2016).

Remedy #1

Specific remedy proposed

Task 2.1.6 aims to clarify the nature of ZTD formal error estimation, having focus on the data analyst’s freedom in giving different initial constraints for GNSS data processing. TUT and MO continue with collaborative experiments, using the same set of sites in experimental network (sites chosen from COST Action BENCHMARK test), using different software and different experimental setups. The results give a possibility for a comparative study – how much the results may differ from different experiments and what will be the average formal error differences from different software. Using E-GVAP sites gives us a possibility to compare our results with results from processing the same sites by many other data analysis centres.

The main goal for the next steps is making analysis of experimental results illustrating the data processing and uncertainty assessment chain with suggestions on how to guarantee the maximum transparency of the full process. The results will be published in peer-reviewed literature.

Achievable outcomes

The first outcome will be making GRUAN GNSS-PW with transparent uncertainty analysis usable for the VO. It cannot be expressed in euros, but additionally it should help to make decisions for selecting and extending the ‘Virtual Observatory’ database with verified and usable non-GRUAN GNSS-data available worldwide (potentially processed with alternative software and data processing strategies compared to GFZ). In the future there could be a relatively dense global dataset for GNSS-PW data usable for the ‘Virtual Observatory’.

Technological / organizational viability: high.

Indicative cost estimate: medium (>1million)/low (<1 million).

Relevance

This remedy (Remedy #1) should be sufficient for G2.34 and is not relevant to any other gaps defined.

Timebound

The task should be mostly completed by the end of 2016, beginning of 2017.

Gap risks to non-resolution