JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 98, NO. B3, PAGES 4619-4626, MARCH 10, 1993 Short- and Long-Term Stability of the JILAG-4 Absolute Gravimeter GEORGE PETER, FRED J. KLOPPING,AND GLENN S. SASAGAWA Office of Ocean and Earth Sciences, NOAA, Rockville, Maryland JAMES E. FALLER Joint Institute for Laboratory Astrophysics,University of Colorado, Boulder TIMOTHY M. NIEBAUER AXIS Instruments, Boulder, Colorado Variations of absolute gravity measuredwith the JILAG-4 absolute gravimeter at intervals ranging from 2 hours to 5 years are reviewed to ascertain short- and long-term instrument stability. We find that the standard deviation of the twenty-four 2-hourly drop set means taken during a given station occupationis 1-2/xGal when natural or man-induced microseismic conditionsare low and 3-5/xGal when the microseismicactivity is high. The standarddeviations of the station gravity values obtained by repeated occupationsweeks or years apart are within these same ranges, with lower standard deviations found againat bedrock sites where the microseismic noiseis low. Basedon the repeatability of observations since the beginning of the measurement program in 1987, thereis no indicate!on for drift, gradual deterioration, or aging of the instrument. However, because of the degraded perfor- mance of the lasers used since 1990, the standarddeviation of repeated station occupations increased from 2.27 to 2.87/xGal, and data had to berejected at several sites. Individual station gravity Values in excess of +_3/xGalfrom the stationmean are found mostly at those siteswhere density Variations between reoccupations are expectedon the basisof geological conditions,usually due to groundwater table fluctuations and/or soil moisture changes. INTRODUCTION The potential use of the absolutegravimeters designed and built by the Joint Institute for Laboratory Astrophysics (JILA) for the detection of vertical crustal motion has been recognized immediately after the first test measurements with one of the instruments by Niebauer et al. [ 1986]. In this report we use a 5-year data base to demonstrate the capa- bilities of the JILAG-4 instrument and the implications of high-precision gravimetry in the measurementsof vertical crustal motion. Absolute gravity measurements have been repeated at several sites with the JILAG-4 instrument by the National Geodetic Survey (NGS) since 1987 and by the Office of Ocean and Earth Sciences (OES) since 1991 to establish absolute gravity reference sites and to develop measure- ment, data processing, and modeling techniques. Of the 76 sites that have been visited since the beginning of the program, 28 have been occupied 2 or more times. In the following analysis, repeatability of the measurements (pre- cision) is used to demonstrate the performance stability of the instrument. Repeatability is shown first for the measure- ments taken during a given station occupation and second for the station gravity values obtained at multiple station occupations. The principal environmental and instrumental sources that can adversely influence measurement scatter are highlighted. The operational principles of the JILA instruments have been described by Zumberge [1981], Zumberge et al. [1982], Copyright 1993 by the American Geophysical Union. Paper n!amber 92JB02529. 0148-0227/93/92JB-02529505.00 Niebauer et al. [1986], and Niebauer [1987]. Data collection and correction methodologies used by NGS have been describedby Peter et al. [1989] and Klopping et al. [1991]. In short, individual gravity observations are obtained by mea- suring the acceleration of a dropped object in a vacuum chamber. The gravity observations (individual drops) are taken at approximately 10-s intervals. Because there are differences among singleobservations that range from a few microGals to --• 1000/•Gal depending on local microseismic conditions,the principal absolutegravity measurementdata used for tracking temporal variations are the mean of 250 drops. It takes approximately 43 min to complete 250 drops (called a drop set, whose mean is defined as the gravity determination). In current field procedures, 16 to 24 drop setsare collected in a period of 2 days, with setstaken every 2 hours. The drop setsare weighted by their variance to take into accountthe changing noise conditionsduring the station occupation, and this weighted mean is the gravity value assigned for a given station occupation. To eliminate temporal gravity variations caused by envi- ronmental influences, we apply corrections for lunar-solar attraction, Earth tide, atmospheric pressure variations, ocean loading, polar motion, and groundwater effects. Ex- cept for lunar-solar attraction [Tamura, 1982, 1987], ocean loading (D.C. Agnew, unpublished manuscript, 1988), and site specific corrections for groundwater effects, we use the international standard models summarized by Boedecker, [1988, 1991]. The estimated accuracy to which these correc- tions can be obtained in most locations is better than 1 •Gal; however, the passage of large weather fronts during the measurements, the inadequacies of the oceanloadingmodels at certain coastal sites, and the assumptions involved in 4619