A common Excel worksheet has been elaborated and distributed for the NMAs as the general data exchange and evaluation format. The final EUVN-DA worksheets include the checked and analysed GPS and levelling data provided by the countries. Following the detailed analysis and tests the completed worksheets were sent back to the data providers. The national datasets are available on request from the data providers, they listed in the table EUVN-DA data of the countries.
These worksheets are considered as the direct product of the project and they may be the starting point for further studies e.g. the computation of the height reference surface (HRS). The HRS is an alternative term of the modified gravimetric geoid / quasigeoid, which can serve and support the physical height determination using the GNSS technique. Although the realization of the HRS is out of the actual EUVN-DA tasks the feasibility of the continental scale gravimetric geoid – GPS / levelling combination is demonstrated. The height anomaly differences of EGG2008 and the EUVN-DA GPS / levelling data have been modelled by the "Local Polynomial" approach selected from well known interpolation methods (e.g. minimum curvature, kriging, polynomial fitting).
On the contour map below the modelled differences and the remaining residuals are shown. The fitted surface is smooth and the residuals are well below ±7 cm over the greatest part of Europe. Differences over this level can be observed only at regions discussed in the following "Open Issues" section (Iberian and Balkan peninsula), where either the quality of the geoid is lower or the GPS / levelling data do not conform all quality standards.
These results are clearly demonstrate the feasibility of the realization of a high accuracy HRS based on EUVN-DA and EGG2008. Further, in-depth analysis is foreseen in cooperation with the EGGP.
The picture shows the modeling of the EUVN-DA-EGG2008 height anomaly differences (contour interval: 2 cm). The fit residuals (black and white sticks) are in agreement with the estimated accuracy of EGG2008.
The gravimetric geoid models and the GPS / levelling data well agree over the greatest part of the continent, but at some regions inconsistencies were still observed, those should be further investigated in the frame of the EVRS project:
- The Iberian part shows moderate improvement and noisier comparison results. This reflect the high age of the present Spanish levelling data (coming from the 1930ies) included in UELN and EUVN-DA. With the completion of the new national levelling network of Spain in 2008 and later providing the data to UELN and EUVN-DA this situation should be improved.
- A small trend in the Portuguese data is observed. Part of the data is also from the 1930's, and additionally the EVRF2007 heights are derived by transformation.
- The North-South trend over France is also remarkable. This trend is in agreement with the tilt IGN found in the French levelling network (Duquenne, 2007). IGN will measure further primary levelling lines to verify the tilt.
- The EUVN-DA / EGG comparisons over Bulgaria and Romania are noisier, because the gravimetric geoid solutions contain only sparse and old gravimetric data from that region.
Over the British Isles all comparisons show negative differences (higher EGG values with an average of -50 cm) and a clear negative trend (dN = 70 cm) against the continent, which already starts on the territory of France. The observed tilt may caused by several factors:
- tilt in the gravimetric geoid solutions, which already seen to start on the continent
- the current "official" connection is done by hydrodynamic levelling (done in 1963); the connection of the Ordnance Survey's levelling network (ODN) to UELN is not yet finalized through the EuroTunnel, therefore a datum offset may still exists
- a tilt in the UK levelling network may also be present as concluded by Edge (Edge, 1959) from comparison of the 2nd and 3rd levelling solutions, where the suspected main contribution is from the 3rd epoch. The current ODN realization is based on the 2nd levelling epoch, while the UELN uses data from the 3rd levelling epoch.
- In Italy the differences are mainly positive and a trend along the peninsula is appearing. If this trend is removed the RMS of the residuals decrease to 6 cm, which is in agreement with the rough error estimate of the underlying EGG2008 geoid model. The source of the observed tilt could not be identified.