Home | english  | Fakultät | Impressum | KIT

Veranstaltungskalender GPI

How a Superconducting gravimeter can support Absolute gravimetry: Eight years of investigations at Onsala Space Observatory, Sweden
27.06.2017, 09:30 bis 10:30
Geb. 6.42, Raum 001


  1. Absolute measurement of gravity is one key component of research into the post-glacial isostatic rebound of Fennoscandia.
  2. The measurements require reduction for known gravity variations of tidal and atmposheric origin.
  3. A stationary, continuously recording Superconducting Gravimeter can provide improved data for reducing both deterministic and stochastic perturbations.
  4. At a location on the coast of Kattegat, the situation with loading and massattraction of atmosphere driven water mass is a complication but mastered with the aid of the Superconducting Gravimeter.
  5. While the Superconducting Gravimeter is capable to represent site specific perturbations on a time scale above minutes, acceleration readings from a seismometer are used to reduce unmitigated microseismic noise at shorter periods.
  6. A method is presented that adjusts multi-campaign absolute measurements at full bandwidth, from 5 s to 7 yr. At this point we shall have gained insight into repeatablity and reproducability of each instruments's measurement, iincluding the nuisance of an instrumental drift in the Superconducting Gravimeter.


Fennoscandia is still in the process of isostatic adjustment after the Pleistocene glaciation that ended approximately 11,000 years ago. Geodesists in the Nordic countries are cooperating to quantify the range of phenomena entailed. There is uplift in the centre and subsidence in the perifery, horizonal displacement, sea level signatures, and changes in the gravity field. The motion is slow, rates of change peaking at the order of 10 mm/yr and 20 nm/s2/yr.

This presentation will focus on the change of gravity. Solid earth models suggest an average ratio between rates of gravity and rates of uplift of -1.63 nm/s2/mm. The ambition of the cooperative project coordinated by the land survey institutes in Denmark, Finland, Norway, Sweden and partners from the Baltic countries is to map and verify the gravity change, and to investigate for instance into anomalies and systematic non-uniform relations between uplift and gravity change, and the impact of changing sea level. By all means, geoid change impacts precise-levelled topographic height, and land survey agencies are commissioned to provide accurate information as to the expected change of the level surface, not least for infrastructure planning.

The preferred means of measurement of gravity change is Absolute Gravimetry (AG), the instrument a gravimeter employing the free-falling mass principle, and the preferred model the Micro-g FG5 metre. Observation is typically conduced in campaign style, where 25 stations throughout the Nordic and Baltic countries are revisited on an annual basis. Onsala Space Observatory supports these efforts with a stationary, continuously recording Superconducting Gravimeter (GWR ser.no. 054, henceforth: SG). The station was established in June 2009, the location chosen to draw advantage of the space geodetic instrumentation there (VLBI, GNSS) that has been contributing to international geodetic colaborations since 1961 and 1989 and thence to the International Terrestrial Reference Frame. In other words: Position and velocity of Onsala are well known and tightly re-observed.  

The SG supports the AG campaigns at Onsala in a range of ways. The gravimetric laboratory houses the SG and offers two concrete platforms for visiting AG's. In regard of measurement, the SG

  • provides a local set of tide coefficients replacing the standard AG model
  • observes the short-term gravity effect due to the atmosphere, replacing the simple method of using a barometer-to-gravity conversion factor,
  • and atmosphere's secondary effect by way of the redistribution of water mass in Kattegat Basin near-by as a response to pressure and wind forcing,
  • and any other kind of gravity change during an AG campaign the cause of which we actually don't need to understand (yet our inquisitive determination notwithstanding).

AG campaigns consist of an instrument setup (called project) on top of a geodetic monument in a specific orientation north or south, and measuring a series of trajectories of a fre-falling prism, typically one drop every five or ten seconds, thousands of such drops per setup, grouped into sets of e.g. 50 each. In the standard procedure, each drop observation is reduced for the influence of known gravity variations (tides, polar motion, atmosphere), whereupon the sample is screened for  outliers, weighted averages are computed for each set and finally, the equivalent is done on the set sample. Thus, an average mean value of reduced gravity is obtained for each project.  
SG support of AG may thus contribute to reduce AG campaign by campaign. The SG may have a major advantage in its extremely low noise on time scales of minutes and days, but it has an annoying feature in the shape of a drifting mean reading. If the hypothesis of a simple, linear function holds (eventually augmented with a unique, decaying exponential), temporal anomalies of gravity could be identified as departures from this parsimoneously parameterized drift. This would adjust offsets between AG campaigns due to natural causes and possibly identify those of instrumental origin. Can this be done with confidence?

A second approach to reduce AG measurements takes a grip at the very-short time scale, i.e. at drop level. Here, the dominating natural perturbation has it origin in the ambient micro-seismic noise. The AG intrument makers have taken precaution to isolate the dropping chamber for these vibration with an effectively infinite-length spring suspension (called Superspring). However, its noise rejection is less than ideal. At noise levels of 100 nm/s2 or above, I can demonstrate that the superspring overcompensates the vibrations. The SG is not useful in this part of the spectrum owing to measures to record alias-free data at 1 Smp/s interval; a strong low-pass filter causes attenuation and phase lag. Instead I use a broadband seismometer (Guralp CMG T3, 120s) operated by Uppsala University within their SNSN network at a location 800 m west of the gravimetric lab. As it turns out, a bespoke combination of SG and Seismometer acceleration signals affords us a decrease in SG drop noise when micro-seismicity is strong. Since the seismographic correction is citical in terms of precise timing, we can also correct the time-stamp information of the AG-drops to GPS-time precision at ter ±0.1 s.

Once all drops of all the AG campaigns since 2009 are collected and parallelled with the SG+Seismometer and instrument- and setup-bias series, a multi-campaign regression can be performed. The most critical effect (or better: nuisance) in this reduction to account for is a time series of the drift of the SG. The audience of the seminar will be confronted and prompted for their critical comments on this rather challenging, alternatively doubtful, approach.

Dr. Hans-Georg Scherneck
Chalmers Universität, Göteborg, Schweden
Geophysikalisches Institut (GPI)
Hertzstrasse 16
76187 Karlsruhe
E-Mail:webmasterFhe0∂gpi kit edu
Mitarbeiter, Studierende
Neue Veranstaltung für Kalender vorschlagen