Geoida Help

Combining GNSS and terrestrial observations

THIS TOPIC HAS BEEN INCORPORATED INTO THE 3D Adjustments TOPIC

Combining terrestrial observations with GNSS and immediately adjusting the whole network with the 3D Terrestrial + GPS option can result in a poor or un-converged adjustment if there are 'incompatible' observations present. In such a case an important point to start preparing for the adjustment is to isolate the broad observation types and compute separate 1D, 2D and 3D preliminary adjustments as follows:

1. First disable the GNSS observations (GPS Vectors and/or Coord. Observations) and any 3D terrestrial observations (Slope Distances and/or Zenith Angles) by un-crossing 'Active' on the corresponding tabs, and adjust the remaining terrestrial observations in 2D. Eliminate outliers, correct errors, adjust settings, and fine tune until a satisfactory 2D adjustment is made
2. Next disable the 2D terrestrial observations and activate the 3D observations made in-active in step 1 and adjust in 3D. Again eliminate outliers, correct errors, adjust settings, and fine tune until a satisfactory adjustment 3D is made
3. Finally activate all the observation types and adjust again in 3D

If there are no conflicts, the adjustment should be satisfactory but once again may need fine-tuning and refinement of parameters and settings as in the preceding steps.

However if the adjustment now immediately diverges or there are excessive residuals, corrections, observation outliers, large error bars or ellipses, unacceptable variance factor or other statistical values, etc, then incompatibilities exist between the 2D and 3D observations which must be identified and eliminated. For example:

• Check whether or not horizontal distances have been reduced to ellipsoid, particularly at high terrain elevations (see the Reduce Dists to Spheroid checkbox on the General tab). If the network extends over a large area it is better to not reduce any slope distance and vertical angle observations into horizontal and vertical components but to retain them as valid observations in 3D space
• Did the networks that were computed separately in steps 1 and 2 above appear to be be the same and with matching points in exactly the same positions? If not then there may be observations with points numbered incorrectly
• GNSS baseline vector observations will normally be very reliable but should still be checked by applying loop checks during adjustment (Check loops checkbox) to verify there are no miscloses. Outlier-test errors and excessive corrections or residuals appearing in either the vector observations, or in terrestrial observations connected to the same points, may indicate either incorrectly-numbered points or incorrectly-reduced terrestrial observations (eg Height Difference observations - see next item)
• Have Height Difference or vertical component observations been reduced accurately? While approximate instrument and reflector heights have little effect on reduction of distances to horizontal, the corresponding derived vertical components may be significantly in error. In this case, to include approximate vertical measurements in the network will severely conflict with 3D observations such as GNSS baseline vectors; however their effect can be compensated by modifying the vertical centring standard deviations to more realistic values consistent with their accuracy, but then their usefulness for inclusion in the network is doubtful
• Disabling observation types in turn using the 'Active' checkbox can help to isolate anomolous observations
• Clustering of large error ellipses and error bars (corresponding to excessive corrections, computed standard deviations etc) around certain points in the plotted network provides a quick visual indication of the possible location of a problem

For related information please refer to the following FAQ:

or to these Help topics: