An intercomparison of LaCoste and Romberg Model‐D gravimeters: results of the International D‐meter Campaign 1983

Abstract

Summary. The LaCoste and Romberg Model‐D gravimeter has found many applications in studies of geophysical processes around the world. To assist in evaluating the results of these studies the performance of 13 Model‐D gravimeters from seven countries was determined during a two‐week observing campaign in 1983 on calibration ranges in the Federal Republic of Germany. Independent analyses of the data were carried out by the Institute of Physical Geodesy (IPG), Federal Republic of Germany, the Earth Physics Branch (EPB), Canada, and the International Latitude Observatory of Mizusawa (LOM), Japan. All instruments were found to have significant non‐linear calibration functions (10–60 μGal corrections) and periodic error terms (2–7 μGal amplitudes). Calibration functions for most gravimeters were expressible in terms of third‐order polynomials but a few instruments exhibited shorter wavelength structure. The calibration results appear to be similar to those obtained by the manufacturer's laboratory method, Cloud‐croft Junior. Redundant observations taken at different reset‐screw positions were used to estimate the uncertainties of the calibration functions. It is shown that the accuracy of a D‐meter can be expressed in terms of the repeatability of readings and the uncertainty of its calibration function. For the average D‐meter the present calibration error exceeds that associated with repeatability when gravity differences are greater than about 26 mGal. Although variations in precision and residual drift of the D‐meters could not be clearly linked to environmental factors, the best results were obtained by D21 which was distinguished from the other gravimeters by having an extra thermostatized enclosure. Instruments transported in the manufacturer's cases without a special suspension system tended to give inferior precision. It is estimated that if two or more D‐meters are employed in a well‐designed network spanning a range of 100mGal, 1 σ errors of 3–4μGal can be achieved with respect to the network mean. Copyright © 1987, Wiley Blackwell. All rights reserved