Not only do the GONG instruments and data reduction system work, but the Global Oscillation Network Group works! The scientific community, organized in a number of teams, has been working together since the conception of GONG to define the scientific program and to prepare the analysis techniques. Thus they were able to quickly apply their efforts to the first look of full-network data. The fruits of their labors were published in a special issue of the journal Science, which appeared on 31 May. In addition to a spiffy--and intriguing--cover, the GONG special issue contains seven articles, covering 29 pages, resulting from the contributions of 71 different authors, and a cast of hundreds more.
These first results are just scratching the surface, in terms of the accuracy and precision that will progressively improve as the data collection, and our understanding of its interpretation, proceeds. Nevertheless, they are extremely heartening in terms of what they already show, and what they promise for the future. It is clear that we are entering a whole new realm of analysis in terms of the care and subtlety required to pursue inference from data with this high frequency resolution and low noise.
The GONG Project was undertaken to test models of stellar structure.
Figures 1a and b show the differences between inferences from GONG
observations and models for the variation of sound speed (squared) 
and 
as functions of depth within the Sun. There are indeed
important differences. First of all, the adiabatic stratification in
the Sun appears to penetrate more deeply than in the model. Part of the
difference in u between Sun and model could be that the model
convection zone may be too shallow. However, the excess u caused by
that property is of lesser magnitude than that in the figure, and
extends more deeply. The small positive value of 
between 0.3 R and
0.6 R might be accounted for in this manner, but the relatively sharp
bump between 0.6 R and 0.7 R cannot. The decrease in u locally may
indicate that immediately beneath the convection zone the accumulation
of helium, which augments the mean molecular mass, has been
overestimated in the reference model. The bump could in principle have
been produced by an opacity error that drops abruptly to zero
immediately beneath the base of the convection zone. However, such a
fortuitous occurrence is unlikely. The discrepancy in the core is the
third prominent feature. Most secure is the negative region of du/u
between about 0.1 R and 0.2 R, which implies that the variation of u
itself is flatter than in the model. Once again this would be a symptom
of there being too steep a composition gradient in the model, which has
been produced here by nuclear reactions. The density inversion (Figure
1b) is consistent with this interpretation: the regions of relatively
steep positive slope in 
in the core and immediately beneath the
convection zone imply that the magnitude of the (negative) gradient of
density is too high in the model.
The area of stellar structure where helioseismology has made a unique
mark is the study of internal rotation. Figures 2a and 2b show two
different analyses of the solar internal rotation rate, based on four
months of GONG data, which are in hearteningly good agreement. Figure
2a utilizes a method known as "Optimized Local Averages," while Figure
2b utilizes another known as "Regularized Least Squares." In the
convection zone above latitude 
30
,
the data show that the rotation
rate at fixed latitude is roughly independent of depth, so that the
variation with latitude is similar to that seen at the surface. Near
the equator, the rotation rate increases just below the surface and
reaches a maximum at roughly 0.95 R. It then gradually decreases with
depth in the convective envelope. At the base of the convection zone
near 0.7 R there is a pronounced transition at all latitudes to nearly
uniform rotation at greater depths. The structure of the transition is
not yet resolved. Thus the GONG data support earlier deductions that
the surface-like differential rotation is smoothed out near the base of
the convection zone and the rotation below appears to become
independent of latitude. The data from GONG currently permit reliable
inferences only to a depth of about 0.4 R, and the use of global modes
yields progressively less information toward the poles. Our confidence
in the inferences made from the nearly continuous GONG data is enhanced
by the improved determination of frequency splittings of individual
modes, because GONG spectra do not suffer from daily sidelobes, which
plague single-site, ground-based observations.
In addition to helioseismic inferences of the solar interior structure,
GONG's measurements of the velocity of the solar surface are of
interest for the measurement of its nearly steady motions--large scale
flows and convection, as well as rotation--all of which are thought to
play important roles in generating the Sun's magnetic field. The
differential rotation stretches meridional magnetic field lines to form
strong toroidal fields; that is, field in longitudinal rings about the
Sun's rotation axis. The meridional circulation transports magnetic
flux and angular momentum across parallels of latitude and from one
radius to another. The nonaxisymmetric convective motions also
redistribute magnetic flux and angular momentum in more complex and
subtle ways. The Sun's differential rotation is accurately determined
from single GONG observations. The rotation profile with latitude
agrees well with previous measures, but also shows a slight north-south
asymmetry. Rotation profiles averaged over 27 day rotations of the Sun
reveal torsional oscillation signal--weak, 5 m/s, jet-like features
associated with the sunspot latitude activity belts. Figure 3 shows the
difference between the measured rotation signal and the spherical
harmonic fit to that signal as a function of latitude, averaged over
three time intervals. The jet-like torsional oscillation signal appears
as the fairly broad (
15
wide)
features that are seen in all three
averages. The latitudes of these features (18
north
and 22
south)
are
slightly poleward of the latitudes where sunspots were found during
this period. The smaller features seen in the shorter averages are
attributed to signal leakage from the nonaxisymmetric cellular flows.
These are just three snippets from the Science articles. For the whole story, and color too, check the original!
John Leibacher