Applications of Absolute Gravity Measurements

Summary of a planning Workshop funded by the National Science Foundation

CIRES, Boulder, Colorado, March 22-23, 1993

Robin Bell Lamont-Doherty Geological Observatory
Roger Bilham University of Colorado
Jim Brune University of Reno, Nevada
Eric Canuteson Scripps Oceanographic Institute
Bill Carter NGS NOAA
Nick Courtier GSC Ottawa
David Dater NOAA Boulder
Ramesh Govind CCAR Boulder
Richard Gross JPL
Susanna Gross CIRES, University of Colorado
Fred Klopping NGS NOAA
Jim Faller JILA Boulder
Jerry Geshwind DMA, Cheyenne
Tom Herring Massachusetts Inst. of Technology
Richard Hilt Colorado College, Colorado Springs
Mike Jackson CIRES Boulder
Marcia McNutt Massachusetts Inst. of Technology
Tim Niebauer Axis Instruments, Boulder
Bishnu Phuyal OSU Dept of Geodesy and Surveying
Tad Pfeffer INSTARR Boulder
Lee Row NOAA/NGDC Boulder
John Rundle Lawrence Livermore National Laboratory
Glenn Sassagawa Earth Physics Laboratory, NOAA
William Updegrove USGS Arizona
Tonie VanDam Massachusetts Inst. of Technology
Lowell Whiteside CIRES Boulder
Mark Zumberge Scripps Oceanographic Institute
written contributions from the following who could not attend the workshop:
John Goodkind, Will Featherstone, Selwyn Sacks, Francis Wu, Rick Williams and Jaques Liard

Introduction

Absolute gravity measurements can be of critical importance in testing models for earth system processes, either in conjunction with deformation measurements such as Global Positioning System (GPS) data, or as a separate data base. In addition, highly precise absolute gravity measurements can be of major importance in monitoring the changes associated with the progress of a variety of natural and man-induced processes. To explore these ideas, a workshop was held at the University of Colorado, Boulder, CO, on March 22-23, 1993. As a result of the discussions conducted during the course of the workshop over the two days of meetings, a number of interesting and important applications of absolute gravity observations were described. It was clear that for many of these applications, absolute gravity measurements, rather than relative gravity measurements, will be essential, and in fact will be required. This is certainly the case for problems involving mantle dynamics, and for measurement of elevation changes induced by active tectonic processes (for example, earthquakes and volcanism) over distances in excess of several hundred kilometers. A particularly important point is the need to combine absolute gravity measurements of elevation change with horizontal and vertical measurements made using GPS or other technologies. Moreover, using absolute gravity measurements along with GPS or leveling measurements of elevation changes is critical in some cases for unraveling the physics of many processes as described below. With these considerations in mind, the workshop participants felt that it is particularly important for the absolute gravity facility to have a close relationship with UNAVCO, so that simultaneous use of GPS receivers and gravity meter(s) are possible.

Discussions at the workshop focussed particular attention on the potential accuracy of the new FG5 instrument, compared to exisiting forms of geodesy and previous generation absolute gravity instruments. Although opinions concerning the theoretical error budget varied somewhat, a consensus was derived indicating an upper and lower limit for FG5 performance.

1. Instrument Accuracy and Precision, and Environmental Noise

There are three distinct issues that must be considered when designing an experiment that makes use of an absolute gravity meter. While the order of importance of the three issues - instrument accuracy, instrument precision or repeatability, and environmental noise - depends on the application being considered, it is highly desirable that experimenters understand the role that each plays.

The new Micro-g Model FG5 absolute gravity meter is of the free fall type. A corner cube reflector is surrounded by a cage that allows it to fall freely through a vacuum path and to be returned to its rest position every few seconds. The position of the accelerating corner cube is monitored by an optical interferometer illuminated by an Iodine stabilized laser. The reference arm of the interferometer consists of a long-period spring arrangement that effectively decouples the instrument from short period seismic accelerationof the ground.

1.1 Precision or Repeatability

Of the three factors that set the limit of the utility of absolute gravity meters, repeatability is the only one that is not subject to a lot of debate; the data speak for themselves. The advent of the FG5 is fairly recent, so a limited number of cases are available in which repeat or side-by-side measurements have been made. Besides the instrument tests made by the manufacturers, a series of inter comparisons has been performed by the NOAA Absolute Gravity group in their Table Mountain lab in Boulder. Figure 1 displays the results acquired with several instruments, including 2 FG5s and one of the earlier generation JILA instruments. The RMS difference between gravity values obtained by different instruments operating more-or-less simultaneously, is 0.9 µGal. The typical RMS variation in the values obtained by an individual instrument at the same site over a period of 30 days is 3-5 µGal. Over 2-3 day periods the RMS variation is of the order of 1µgal.

1.2 Accuracy

A much more difficult parameter to assess is the inherent absolute accuracy of the instrument. In this context, accuracy refers to the certainty with which the measurements can be claimed to be representative of the true (and unknown) value of local gravity. This may be deemed less important when temporal gravity changes are sought, but if the data are to remain useful for a time longer than the life of the instrument that produces them (usually a desirable situation), then the absolute accuracy is of paramount importance. Quite a number of the geophysical signals sought are slowly varying, and may take decades to resolve. These are the cases in which instrument repeatability is not as relevant as accuracy.

Many tests on the accuracy of absolute gravity meters have been performed on previous instruments. The newness of FG5 prohibits its having been tested with the thoroughness applied to previous instruments. Because its design is based on those previous instruments (with improvements being incorporated in those areas where a strong impact on the total accuracy could be made (the laser and the interferometer, for example) earlier investigations of accuracy, along with models of the individual error mechanisms, have allowed a preliminary (and therefore conservative) error budget to be assembled. It needs to be emphasized that more data from the instruments and added experience in their use may result in modifications to the table below. (Starred items are operator dependent.)

Systematic Error Budget estimates

	consensus estimate	low estimate
Differential pressure	0.5 µGal	0.3
Differential temperature	.5	.3
Magnetic field gradients	.5	.3
Electrostatics	.5	.3
Attraction of apparatus	.1	0
Air gap	.6	0
Laser wavelength	.1	.1
Rotation of corner cube	.6	.6
Translation	.6	.3
Floor recoil and tilt	.2	.2
Frequency standard	.1	.1
Glass wedges	.1	.1
Measurement height	.3	.15
Verticality	0.5	1.1
total (RSS)	1.7	1.1

1.3 Environmental Requirements and noise sources

In addition to the inherent instrument uncertainty, the limit in our ability to resolve a change in gravity depends on the sources of noise in gravity produced by the environment during the measurements. The table below provides representative signal sizes from these noise sources, followed by an estimate of our ability currently to make corrections for them.

Source	Typical	Uncertainty in correction
Tides	300 µGal	0.2-0.5 µGal
Periodic Ocean Loading	20 (coastal sites only)	0.2
Non periodic ocean loading	10	2
Atmospheric effects	8	1
Water Table	0-100 site dependent	0-10
Microseisms	7-100 25 typical	0.5
Temperature	10 µgal transient @1 hour	0-1

1.4 Summary: Probable accuracy of FG5 gravimeter in geophysical applications.

Workshop discussions outlined above suggest that instrument accuracies of the order of 2 µgal may currently be available, with site occupational repeatabilities equal or better than 1 µgal. The suppression of seasonal atmospheric and subsurface moisture and water table contributions to the observed value for surface gravity remains an important area for experimental research, since if these are ignored or inadequately modeled substantial errors may be incurred. Assuming that these parameters can be adequately modeled to Å1 µgal (Van Dam and Wahr, 1992,1993)), absolute-g provides important constraints in vertical deformation measurements (Figure 1). In particular, absolute-g, VLBI and GPS have somewhat similar measurement uncertainties for all baselines, and given their very different sources of systematic noise, there was unanimous agreement that measurement combinations of space geodesy and absolute-g will be very rewarding.

Figure 1 Relative accuracy of absolute g, precision leveling and space geodesy (after Bilham 1991). Shaded areas indicate range of vlaues obtained in recent investgations.

2. Applications

The following is a summary of selected scien-tific issues that could be addressed should the abso-lute gravity facility instrument become available.

2.1 Active Tectonics

Measurements of values of the absolute gravity field of the earth can be extremely useful in addressing a series of problems related to active tectonic processes such as earthquakes, volcanism, and nonseismic deformation processes. The preferred measurement of choice for horizontal observations is GPS geodesy or, in some cases, multiwavelength geodimeters together with airborne measurements of temperature and humidity. Vertical measurements of deformation can be carried out by means of precise leveling, GPS, or absolute gravity measurements. The accuracy of each kind of vertical deformation observation can vary with the distance and topography over which the observations are taken, as shown in figure 1. Leveling has the twin disadvantages of being slow and costly. Typically it takes about 1 week or more to carry out observations over 30 km length, and the cost is of the order of $600/km. For these reasons, relative gravity measurements may be used in place of leveling to deteremine changes in elevation, inasmuch as several benchmarks over distances of tens of km can be observed in a single day using several portable gravity meters of the LaCoste-Romberg type. Assuming that the effects of pore waters are minimized by locating benchmarks on crystalline outcrops, and that the proper gravity gradient is known, one can make an estimate of the accuracy in height measurements obtaineable. If the proper gradient is of the order of the free air anomaly, the accuracy in vertical displacements is then of the order of 2-3 cm, of the same order as first order leveling carried out over distances of 400 km. For gravity loops of the order of ~100 km or more in horizontal distance, it is often found that relative gravity measurements degrade due to the effects of random tears in the adjustment spring (nonlinear drift).

An accurate, field worthy absolute gravity meter would have all the advantages of relative gravity meters and few of the disadvantages. It could be transported over arbitrarily long distances with no degradation in the accuracy of the measurements. Moreover, relative gravity observations also suffer from the disadvantage of needing to have calibration runs carried out over elevations larger than those to be encountered during data acquisition. Absolute gravity meters perform at all elevations to the same high accuracy. This attribute permits absolute gravity meters to be used to provide a series of base stations for relative gravity measurements at frequent intervals.

Some problem areas in active tectonics that can be addressed by means of absolute gravity measurements include:

Earthquakes . Measurements of strain accumulation in reverse or normal faulting environments during the interseismic phase; observations of coseismic and postseismic vertical displacements following major earthquakes; calibration of tide gauge observations in coastal regions near subduction zones; and observations of pore fluid migration/diffusion due to stress changes. Examples include subduction zones in Japan, Alaska, South America, Cascadia, and the Himalaya.

Volcanism. Observation of the inflation/deflation episodes for active volcanic edifices both for hazard warning as well as estimates of the mass budget of the volcanic plumbing system; periodic monitoring of dormant volcanoes to establish a baseline for future uplift; and establishment of base stations for relative gravity measurements leading to observations of Bouguer gravity anomalies. Elevation of a magma chamber roof by a gaseous phase would result the observation of surface displacements associated wth the free air gradient. Possible target volcanoes include Mount Saint Helens, Redoubt Volcano, Mt. Augustine, Mammoth Lakes and Yellowstone.

Figure 2 The surface depression and gravitational surface potential changes attending the emplacement of a point mass at different depths in a elastic plate overlying a fluid half space (from Rundle, 1983,88). The displacement signals for the two cases differ by less than 10% but the gravity signal varies substantially.

Continental Extension. Examination of crustal "core complexes" that expose deep crustal material and examination of the mass balance for regions such as Death Valley. It has been proposed that the the Basin and Range in the vicinity of Death Valley is now actively extending with lower crustal material created by this process forming the core complex. Absolute gravity measurements would help to unravel this complex process.

Finally, an important application involves use of absolute gravity measurements to verify and calibrate vertical measurements made by other techniques. These other techniques include leveling, space geodesy measurements (Very Long Baseline Interferometry, Global Positioning System, and Satellite Laser Ranging), and portable relative gravity meters such as LaCoste model G meters.

2.2. Mantle/Lithospheric Dynamics

The Earth's surface is constantly moving vertically in response to changing loads on the surface and at depth. For example, erosion is presently wearing down the elevation of mountain ranges, but other factors such as isostatic rebound, crustal shortening, and flow in the underlying mantle may act to raise the mountains. The challenges which face geophysicists are first to measure these vertical motions and then to separate the signal into the various causative processes.

The availability of an absolute gravimeter at the mgal level may extend our ability to map out regions undergoing uplift and subsidence. Vertical motion of the land surface can only be measured routinely at coastlines with tide gauges, and even this estimate is only a relative measure of vertical motion with respect to eustatic sea level. With the best available precise positioning techniques (e.g., VLBI), it is possible to measure vertical position to the several mm level at points removed from the coastline, but there are only a few places on Earth for which this precision has been achieved. Absolute gravimetry is sensitive to vertical motions as the Earth's surface moves through the free-air gradient. This free-air effect results in a reduction in gravity of -0.3 mGals for each mm of uplift, but other density changes at depth associated with the uplift will, in general, cause the observed gravity change to depart from this theoretical value.

The possibility of simultaneously observing vertical motions using both gravimetry and precise positioning opens up new opportunities for distinguishing among the various processes responsible for the vertical motion. For example, consider the case depicted in the cartoon below intended to represent a continental plateau 1000-km across. Suppose that precise positioning demonstrates that the benchmark in the center of the plateau has moved 20 mm away from the center of the Earth since the last survey. We consider in the Table below 5 possible causes for the uplift.

Figure 3 Cartoon showing various mechanisms for producing uplift of a broad intracontinental plateau. Uplift may be achieved by (a) isostatic rebound in response to erosion, (b) crustal shortening and thickening caused by compression, or (c) dynamic pressure at the base of the lithosphere caused by an upwelling mantle plume.

In Case 1, it is assumed that the uplift is local isostatic rebound in response to erosion of a rock layer 24-mm thick from the surface of the plateau in the vicinity of the benchmark, although the benchmark itself is not eroded away. The columns in the Table show the various contributions to the gravity change that would be associated with this 20 mm uplift. The "Free Air Effect" corresponds to the motion of the surface away from the center of the Earth through the free air gradient, and it is the same for all of the cases considered since the uplift is fixed at 20 mm. The "Bouguer Effect" corresponds to the change in gravity caused by the change in mass of the plateau. In this case, erosion has removed mass from the plateau, and even after isostatic rebound, there is a net reduction in the gravitational attraction of the plateau. The "Isostatic Effect" refers to the change in gravity caused by mass redistribution at depth in order to re-equilibrate the system. In the case of local isostatic rebound, the mass deficit of the plateau's root is reduced, so that this contribution is positive. The Bouguer and isostatic terms nearly cancel because their masses are equal but opposite, but because of the upward attenuation of gravity with elevation, the Bouguer Effect is slightly larger. The net gravity change is a reduction in gravity slightly exceeding the Free Air Effect alone.

Case 2 is similar to Case 1, except that the rebound is regional via flexure of an elastic plate 100-km thick. Because the rebound is response to erosion is spread over a broad area, a rock layer 33.3-mm thick must be removed to produce the same 20 mm of uplift at the station. The Bouguer Effect is larger since more mass has been eroded, but the Isostatic Effect is approximately the same because the reduction in the compensating mass is spread over such a broad area laterally. The net gravity reduction in this case is considerably greater than the Free-Air Effect alone.

For Case 3, the 20-mm uplift is presumed to occur through crustal shortening by 1.3 m. Since mass has been added to the plateau and its compensating root, the Bouguer Effect is positive and the Isostatic Effect is negative. The net gravity is slightly less than the Free-Air Effect alone.

Finally, cases 4 and 5 refer to dynamic uplift of the plateau by the pressure at the base of the lithosphere caused by a 500-km deep rising plume 200°C hotter than the surrounding mantle. If the plume rises through a uniform viscosity mantle, it needs to be only 5.5 m thick to produce 20 mm of uplift. Again, uplift has added mass to the mountain, leading to a positive Bouguer Effect. The Isostatic Effect includes the upward warping of the Moho and core-mantle boundary as well as the gravity reduction due to the mass deficit of the plume itself. The net gravity is again less than the Free-Air Effect, and not too different from case 3. If the plume lithosphere is underlain by a low-viscosity zone, as is assumed in case 5, the pressure transmitted to the base of the lithosphere from the rising plume is substantially less. Therefore, the plume must be 33.3-m-thick to produce the same 20-mm uplift. The Bouguer Effect is the same as in case 4, but the Isostatic Effect is much more negative, largely because the plume itself is larger. The net gravity effect is substantially more negative than the Free-Air Effect alone.

These simple examples demonstrate that neither precise positioning nor gravimetry alone will uniquely determine the density changes that lead to vertical motions, but the combination of the two will provide a powerful tool for learning about the Earth's interior. Ideally, precise positioning and gravity surveys should be carried out over the entire region of study at sufficiently close station spacing to allow computation of the full admittance between gravity changes and elevation changes. Furthermore, given the fact that there may be a time lag between surface elevation changes and the associated subsurface density changes, such geophysical studies should be undertaken with a good knowledge of the geologic history of the region surveyed.

Table 1. Summary of calculation results.

Case	Description	Uplift (mm)	Free Air Effect (mGals)	Bouguer Effect (mGals)	Isostatic Effect (mGals)	Net Gravity (mGals)
1	erosion of 24 mm layer; local isostatic rebound	20	-6	-0.5	0.3	-6.2
2	erosion of 33 mm layer; regional isostatic rebound of 100-km elastic plate	20	-6	-1.5	0.3	-7.2
3	compressive shortening of plateau by 1.3 m	20	-6	2.2	-1.4	-5.2
4	plume of thickness 5.5 m rising through uniform viscosity mantle	20	-6	2.2	-1.6	-5.3
5	plume of thickness 33.3 m rising through an upper mantle low-viscosity zone	20	-6	2.2	-5.	-8.8

2.3. Global Change

Post Glacial Rebound. Rates vary from 1 cm/yr for Fenno Scandia, Barrents and the Hudson Ice Sheets of the Northern Hemisphere and up to 2 cm/yr for the West Antarctic Ice Sheet. The wavelengths of these problems are 1000s of kilometers, and make applications of relative gravity and space based geodetic techniques more difficult. However, weekly means of continuous GPS measurements can now yield accuracies of 8 mm in vertical deformation.. The exception is Iceland where the rebound rates over this this hot crust is 1-2 cm/yr but with very short wavelengths (~100s of km).

Mass Balance Measurements.. Temperate glaciers may be the major contributors to recent global sea level rise. These include the Patagonian Ice Sheet, Alaskan glacier and Central Asia (Meier). Mass balance of these glacier is difficult to determine due to the unknown density structure of the near surface ice column. Absolute gravity combined with height measurements could be used to determine the changes in the absolute mass of these features. This is in contrast to traditional over-ice gravity techniques which have been used simply to estimate the ice thickness. The inherent drift and logistical difficulty of computing loops makes relative gravity measurement unfeasible for this.

Global Sea Level . The current rates of sea level rise are 1.7+.0.7 mm/yr based on 80 years of measurements. The outstanding question is whether the current sea level rise is accelerating. Absolute gravity provides an addition constraint on the vertical motions of tide gauge sites. A 1 micro gal gravity measurement is equivalent to a one centimeter vertical position on the site. Although the accuracy from absolute gravity meters is comparable to space geodesy the systematic errors of the two techniques are different and presumably uncorrelated.

Combining Space Geodesy and Absolute Gravity Measurements . As described in the section on mantle and lithospheric dynamics, absolute gravity measurements must be combined with other precise vertical posittioning techniques to fully constrain the processes involved. If a major ice sheet collapses the load on the lithosphere is decreased and regional gravitational potential is reduced in amplitude. The elastic rebound from this process can be detected both from the reduction in absolute gravity and the associated vertical displacements using space geodesy. Absolute g is sensitive to both phenomena while space geodesy is sensitve only to displacements.

2.4. Environmental Monitoring

A variety of natural and man-induced physical and chemical processes are known to produce both subtle and substantial vertical displacements of the earth's surface. In general, these processes are related to mass or fluid extraction, and to subsurface pore fluid flows. However, other processes that produce vertical displacements include large explosions, landslides and dam failure. Absolute gravity measurements would be invaluable for monitoring many of these processes, because of the sensitivity of the measurements to both vertical displacements and mass extraction. Again , absolute gravity measurements would be particularly helpful because the deformation signals can often occur over large areas (hundreds of km2), and can have either subtle or highly visible effects. Thus, it is critical to have a highly accurate means of measuring gravity over these large areas.

Some problem areas in environmental monitoring that can be addressed by means of absolute gravity measurements include:

Acquifers and Reservoirs. Drawdown of fluids in acquifers, oil and gas fields, and geothermal reservoirs. Examples of acquifer drawdown include areas around Tucson, AZ, and the San Joaquin valley of California and the Geysers geothermal field in California.

Nuclear Waste Repositories. Monitoring of fluid infiltration and water table rise near the Yucca Mountain and other proposed nuclear waste repositories that might endanger the integrity of the repository.

Subsidence due to Mining. Monitoring of subsidence due to mineral extraction, coal mining, and solution mining. Examples include the Strategic Petroleum Reserve in Lousiana, in situ oil shale retorting in northwestern Colorado, coal mining in Pennsylvania and western states such as Wyoming.

Monitoring of Mining Effects. Monitoring the effects ofin situ leaching and pollutant flows from mining operations.

Slope and Earth Fill Dam Stability. Landslides and earth fill dams can fail in both a catastrophic manner and as a more gradual process. In the latter case, absolute gravity measurements can help to monitor the progress of the failure, as gradual cracking and fluid infiltration contribute to the failure process. Examples include the collapse of the Teton and Baldwin Hills earth fill dams, and systematic collapse along coastal California. These processes have large amplitude but very short spatial wavelengths.

On Site Verification/Inspection. Assessment of sites in suspect countries that may have been used for clandestine nuclear tests. With the end of the Cold War, it is increasingly likely that smaller countries will seek to acquire nuclear weapons. If these weapons are tested in a manner calculated to escape detection, it may be possible to explore the site in order to prove/disprove the hypothesis that the suspect country conducted the test. Absolute gravity may play a crucial role here in helping to establish whether cavity collapse has occurred, or that mass redistribution at depth associated with the shot has occurred.

2.5. Atmospheric Effects on Surface Gravity

There are two ways the changing masses of the atmosphere can influence the acceleration due to gravity at the surface of the earth. The air masses exert direct gravitational forces, and the pressure they exert on the surface deforms the solid earth. These effects represent a source of noise for solid earth geophysicists, but the direct effects may possibly be of use to atmospheric scientists. There are several different kinds of direct effects, due to changes in mass, changes in the vertical distribution of mass and lateral variability. The atmospheric pressure at the surface of the earth is due to the weight of the atmosphere above, and is in hydrostatic equilibrium in a local sense because vertical accelerations are much less than g. To estimate this effect, we have:

P = s g

where r(z) is a volumetric density, and s is a weighted surface density. Typical variations in surface pressure are of order 20 millibars for midlattitude winter, so changes in pressure are associated with quite observable changes in g. Using

s = 2000 Pa/g = 204 kg/m2

One finds a perturbation in gravity dg equal to

dg = 2p G s = 8.5 mGal

This simple weight effect is not very interesting to atmospheric scientists because it can be observed more directly and cheaply with barometers.

Other atmospheric effects mean that g is not simply related to P. One of these is variation of scale height of the atmosphere H and consequent variation of the weight of air having the same mass because of the vertical gradient in g. A temperature change of 10 o C would produce a change in scale height dH of roughly:

dH ( 10 o / 250 o K ) 10 km = 400 m

thereby changing the effective g exerted on the atmosphere by 0.12 gals and the surface pressure by 1200 Pa without changing g at the surface at all. The gravity meter is sensitive to the integrated mass of the atmosphere, but the barometer responds to the weight, which is influenced by the temperature structure of the atmosphere. Gravity and surface pressure data together could provide a new kind of constraint upon temperature structure of the atmosphere, but the data are limited in usefulness because of its underdetermined nature.

The atmosphere also has lateral inhomogeneities of the vertical structure which influence surface gravity, just as structures within the solid earth produce gravity anomalies. The boundary between two air masses having a 10oC temperature contrast and a 400 m change in scale height would be associated with a gravity anomaly dg' which can be estimated from a line mass located H/2above the ground and having mass per unit length l = dH s.:

dg' 2G l / H = ( 2G s dH ) / H

Substituting the numbers, we find:

dg' 2 (6.67 x 10 -11 m3/ kg-sec2 10 4 kg/m2 400 m) / 10000 m = 5.3 mGal

The effect could be greater than this if there were pressure contrasts on either side of the front or temperature contrasts near the ground. This last effect is possibly the largest effect the atmosphere has on surface gravity, and may also be the most interesting from the point of view of interpretation for atmospheric scientists. Measurement of the deflection of the vertical would be helpful for distinguishing gravity changes due to lateral heterogeneity from those related to changes in pressure or temperature structure.

Surface measurements of gravity do provide a new and different kind of information about the atmosphere, because g responds to the integrated mass unlike a barometer and also provides information about the lateral structure of the atmosphere. Measurement of g could provide a new constraint on models of mesoscale systems and provide a means of verifying soundings when there was no significant horizontal structure. The short time scales of atmospheric phenomena mean that atmospheric applications could be explored rapidly. Disadvantages are that the cost of the instrument prevents operational uses, and the ambiguity of the data means it would be best used along with more conventional techniques.

References