Accuracy of the FG5 absolute-gravimeter

The accuracy is believed to be 1.1 microgals (see below), but environmental and experimental factors limit achievable accuracy to 2.1 microgal, equivalent to a 6 mm change in elevation.

Environmental variables that decrease the achievable accuracy are as follows:
A. water table variations.
B. atmospheric mass variations.
C. errors introduced by measuring the gradient of gravity at the point of measurement.
D. tidal variations (load tides and body tides).
E. operator blunders and instrumental malfunctions.

Errors in A and B can be reduced by measuring the depth to the water table and estimating rock porosity, and by measuring and integrating regional atmospheric pressure variations. Errors in C occur in measurements with a relative-g meter that is operated at different heights above the ground. This error is is negligible if the instrument is always operated at the same point in space and if no mass changes occur nearby (buildings, large objects, trees). If in 30 years another instrument operated at a different elevation were to measure the point, the gravity gradient would be an essential correction. Tidal signals can typically be calculated to high accuracy except where large unknown load tides occur. Operator blunders can usually, but not always, be identified from increased measurement scatter.

Figure shows the perfomance of NSF's FG5 111 during the perod1995-1998 . Each experiment has a duration of 1-3 days, with 60-120 discrete observations of g each hour.

The figure shows 82 measurements taken at various times throughout the period 1995-1998. The blue squares are the total uncertainty, and the other symbols represent its component parts. The yellow circles (grad) are the uncertainty in measured gravity gradient using a LaCost Romberg gravimeter, the blue diamonds (obs) are the standard deviation in observational error of the FG5. The pink line is the estimated absolute uncertainty of the FG5 based on theoretical estimates of errors.

Estimating accuracy Estimates of the accuracy of absolute-g machines have one common difficulty - there is no instrument with higher accuracy against which to calibrate the gravimeter. Tests thus fall into three categories:

1) Tests that compare one gravimeter with another of different design
2) Tests that compare similar gravimeters against each other.
3) Tests of repeatability of one gravimeter against itself over time

Tests of the first kind, combined with theoretical calculations of errors, lead us to suppose that the FG5 gravimeter has an inherent uncertainty of 1.1 microgal in an absolute sense. Tests of the second category where two identical FG5 instruments are alternately operated on two adjacent concrete piers yield a measurement uncertainty of 0.9 microgal. Tests of the third kind lead us to suppose that measurement errors and instrumental uncertainties have a combined uncertainty of the order of 2.1 microgal. For details of these experiments see the bibliography at this web site.