NASA, ESA, and the European member countries have long term goals that require a better understanding of the Sun, Solar variability on multiple scales, and the "Space Weather" that impacts the terrestrial environment. These goals are described in a number of long range plans and need not be further described here. The present question is to see how SOHO is an asset that can contribute to these goals as well as expand and complement the goals that provided the imperative to develop SOHO in the first place.
SOHO as a whole was developed to address a number of interrelated fundamental questions about the nature of solar physical processes. A number of these phenomena are most simply observed at or near activity minimum. The Sun has been quite obliging in providing a nice quiet minimum during the first year of SOHO operations. Some of the results and studies in progress are described below. A key fact often under-emphasized is that SOHO is working phenomenally well. The spacecraft systems, instruments, and operations teams and facilities are all working as well or better than could be expected. This fact not only enables but provides an imperative for the discussion of how to best use these assets in the coming Solar maximum years.
The MDI instrument is providing data that often exceeds our expectations. While after only eight months of normal operations the science investigation is by necessity still in the early stages of development, it is not too early to consider the opportunities for progress that would be enabled by continued operation to and through the coming activity maximum. The Solar Oscillations Investigation using the MDI instrument has a number of science goals. Some of these require only short duration observations while others need very long duration observing sequences.
Some of the primary SOI science objectives require several years of continuous observations to achieve even the original goals. This is due to the nature of making exceedingly accurate measurements of oscillation mode frequencies. Accuracies of a few tens of nano-Hz are required to make reliable inferences about the structure and rotation of the solar interior and that simply takes several years. While a mission of only two years is justifiable in its expected gains, longer duration time series will certainly provide better limits to the structure of the star. The region of the Solar interior that most needs long-baseline observations is the energy generating core where the present limits of accuracy provide tantalizing hints of mixing.
Other of the primary objectives need continuous sequences of days to months. These are the goals that can be addressed by the new science of Local Helioseismology. These techniques allow imaging of the motions in the interior of the star. Solar minimum is the optimum time to begin this study since the Sun provides its phenomena in the simplest configurations. Early progress has convinced us that the developing techniques will indeed work and we will be able to watch the development of the magnetic cycle from the inside. MDI provides a unique tool for this study. It is the only tool available with the required spatial and temporal resolution to study the upper convection zone where activity evolves as the cycle progresses. If, as we believe, local helioseismology provides a tool that can allow accurate prediction of active region growth, continued MDI operation through to Solar maximum can provide the observational foundation to enable the development of a predictive capability.
In addition to helioseismology observations, MDI can and does provide observations essential for complementary view of solar processes. We have demonstrated the ability to follow the generation and evolution of magnetic fields on the granulation to global scale with a spatial and time resolution sufficient to, for the first time, follow the life-cycle of magnetic fields. MDI also provides measurements of relative brightness of active and quiet regions to enable the better understanding of the radiative flux balance. It also provides direct measurements of surface motions from meso-granulation through to global scales. These studies continued through maximum would provide a unique view of the process that links interior dynamics to the phenomena that affect the EarthÕs space environment.
SOHO as a whole, combining the unique capabilities of all the instruments, has the capability to provide the answers to many of the fundamental questions of solar variability. It would be unconscionable to not use this resource so long as it is functioning.
The MDI instrument is presently operating as designed. There are no known life-limiting elements. There are several areas we are watching. These include the front window temperature which affects the ability to keep the high resolution field in focus. We have no further adjustment range in the direction needed to correct for increased temperature. Front window contamination could cause an increase in temperature which would reduce the quality of the high resolution data. We have seen no evidence of such a trend. The total MDI throughput has not changed. The aging drift of Michelson tuning has slowed from many m/s per day to about 2 m/s per day. Since this drift appears to be limited to the Michelsons only it can continue to be handled by adjusting the tuning every several months with no degradation of science performance. There has been a measurable but negligible increase in CCD dark level. At the present rate, the CCD will outlive the SOHO fuel supply.
The primary MDI science data collection began between 24 April and 23 May depending on the observable. For the primary helioseismology datasets we have now examined the first 5 months (end May through end October) and we have completed preliminary quality checks of the 2-month continuous run. Initial results for the first 60 days were reported at the IAU Symposium 181 in Nice in October (20 papers). Additional results for the first 150 days were reported at the December AGU meeting in San Francisco (8 papers). Additionally special observations obtained during the commissioning phase of the mission (January through mid April) have also resulted in presentations and publications. First result papers have also been submitted to Solar Physics for inclusion in the SOHO First Results issues. All together MDI observations have been reported in at least 50 presentations with 23 papers submitted or in press, and 4 articles in less technical journals. Some of these results are described here.
The science productivity of SOI/MDI can be discussed in four parts: Global Helioseismology, Local Helioseismology, Associated Objectives, Collaborative Science. The global and local helioseismology together address the primary SOI science objectives but are addressed separately due to the differing data requirements. In the discussion below the names of the individual investigators have been omitted for brevity. The work has been accomplished by members of the SOI investigators team which includes people at many institutions in the US and Europe. Please see the attached lists of publications and presentations for the names of the particular investigators responsible for the results described below.
Global helioseismology results are just beginning to be available from MDI data. Normal helioseismology operations began in early May 1996. As of mid December we have obtained and calibrated 60-day, a 144-day, and a 150-day datasets, with the duration depending on the particular data quantity. Since definitive results are expected to require much longer time series, only preliminary results can be expected at this time. Nevertheless we have achieved some useful results.
We have completed analysis of the first long series of medium-l observations. These observations are similar to the resolution obtainable from the ground with the GONG project. We have allocated some MDI telemetry to this resolution data primarily to allow cross-calibration and comparisons with the longer duration ground based observations. We have been pleasantly surprised by how much cleaner the signals are in the MDI data. It is clear that the spectral noise level is lower in the space-observed data. We have been able to measure the p-mode spectral line asymmetries at mid frequencies and the amplitude and spacing of the acoustic mode ridges at high frequency. The high frequency ridges are well above the acoustic cutoff frequency and are formed by the interference of waves that are reflected only once in the interior with waves from the same sources that proceed directly through the photosphere. Analysis of the high frequency ridges and the mid frequency asymmetries gives information about the nature and location of the sources of the acoustic waves. They are seen to originate about 50 km below the photosphere.
Figure 2 shows vertical slices at l = 100, 200, and 300 from the data in Figure 1. Note
the low noise levels and the clear power peaks up to the Nyquist frequency. Also note
the asymmetry at low frequency. Models of the upper reflection layer show similar
asymmetry. the spacing of the peaks at high frequency indicates the source of the acoustic
modes is about 50 km below the photosphere.
The first inversions of this medium-l data have enabled us to determine the
likely location of the dynamo responsible for the Solar cycle.
Determination of the sound speed profile with depth shows a sharply
defined discrepancy between observations and the best current models just
beneath the base of the convection zone. The excess sound speed in this
region is consistent with additional turbulence not accounted for in the
models. Parallel analysis of interior rotation shows that the shear layer
that separates the differentially rotating convection zone and the more
rigidly rotating radiative zone is also localized just beneath the convection
zone. The combination of these two results, both more sharply
determined with 60 days MDI data than from years of ground based data,
is the evidence for the location of the dynamo. These results were
presented at the IAU symposium in Nice and at the AGU in San Francisco
in December. Further medium-l results must await longer time series.
Figure 3 shows the difference between observed and model sound speed as a function of
depth in the Sun. The "bump" just below 0.7 R indicates the possible location of
excess turbulence.
Figure 4 shows the inferred rotation rate as a function of depth and latitude. Evidently
the convection zone rotates uniformly along a radius with all depths showing the
differential rotation seen at the surface. Below the convection zone is a layer of shear
below which the radiative interior seems to rotate rigidly. This shear zone which
coincides with the sound speed excess could be the region where the solar cycle dynamo
operates.
The analysis of the high-l observations is now underway. These datasets
are quite large, about two hundred thousand megabytes per observable.
We have completed calibration of the data and are processing it through to
spherical harmonic amplitudes and mode frequency determinations.
Results are expected by mid-winter. Preliminary results show a sharp
increase of the line width of the f-mode presumably due to scattering of the
surface gravity waves by turbulence.
Figure 5 shows an l-nu diagram from an 8-hour campaign of high resolution Doppler
images. The ridges are clearly visible to very high l.
Analysis of the MDI data has provided substantially improved
frequencies for the f-mode at moderate degree. Analysis of the f-mode
frequencies provides help understanding the physics of the upper part of
the convection zone. The uncertainty in this region affects our
understanding of all mode frequencies which affects our confidence in
inferences of the structure of the solar interior.
The very high quality of even the comparatively short sequence of data
already analyzed has provided a very significant improvement in our
ability to discern subtle aspects of the thermodynamic properties of the
solar plasma. This confirms that, with the expected improvement from
longer observations, we can realize our hopes of using the Sun as a
laboratory for the study of the physics of dense and hot plasmas.
We have just completed the first intercomparisons between MDI, GOLF,
and VIRGO. We have confirmed that time series and spectra of low degree
modes are nearly identical for each type of observable independent of
instrument. This is a very important demonstration of performance of
each instrument in the domain of overlap in observational capabilities. It
means that joint and independent efforts can proceed with confidence that
the spectral detail is due to the nature of stochasticly excited damped
oscillations rather than observation noise. The comparisons of mode
frequencies also shows that low degree mode frequency determinations are
systematically different for simultaneous observations of brightness and
velocity. This strongly suggests the line asymmetries and differing heights
of formation lead to systematic errors in mode frequency determinations.
The differences are very small but significant. Development of better line
fitting algorithms and assessment of the effect on structure inversions is
underway. These results are described in a paper in preparation for the
second volume of Solar Physics first SOHO results.
The search for g-modes continues. The MDI limb data has a sensitivity
equivalent to a few mm/s. When a long-enough time series is available we
will compare limb, velocity, and brightness spectra and cross compare
with GOLF and VIRGO to further the g-mode search.
Local helioseismology is a new field that is developing primarily with
MDI observations. We are studying three techniques to deduce convective
motions. The time-distance technique has developed along with MDI and
provides the best resolution with the most computational effort. The
ÒringÓ analysis provides a lower resolution view but with a significantly
lower computational effort. We are continuing to examine other Òphase-
advectionÓ methods which might prove as useful as time-distance methods
but with a more straightforward link between the analysis and solar
motions. At this point only the first two methods have produced useful
results.
Data from the very first high-resolution test observations made in January
1996 demonstrated the potential of the time-distance approach. The
method allows imaging of the bulk motions on the meso-granulation to
super-granulation scale well into the convection zone. The first results
showed supergranules to have a shallow structure with a depth to width
ratio of 1:10. The first-ever maps of horizontal and vertical flows in the
upper convection zone were presented at the AAS meeting in June 1996
and have since been featured in several popular magazines. Further
analysis of this data has also shown that the rotation shear in the top 1%
of depth inferred from global helioseismology is detectable in local
averages. Analysis of full disk data using time distance methods has
shown that differential rotation differs in the north and south hemispheres,
confirming inferences from magnetic field pattern rotation. The method
has been validated by inference of motions that are so shallow that they
can be directly observed in the Doppler signal - the patterns are nearly
identical.
Figure 6 is a comparison between the absolute value of the magnetic field (in color) and
the simulated surface velocity from the time-distance pictures (arrows). The longest
arrow is for 1 km/s velocity.
Figure 7 shows a map of horizontal motions inside the convection zone. The colored
shading indicates temperature variations.
Figure 8 shows a vertical slice into the Sun. The supergranule flow evidently extends
only 3-4 Mm into the convection zone. This is the latest analysis of this dataset with
improvements to the tomographic inversion code used to convert time-distance
measurements into velocity and temperature maps.
Figure 9 shows the first analysis of motions in the upper part of the convection zone.
The observations for this type of analysis are unique to the SOHO-MDI instrument. The
approximate location of the region is shown on a schematic diagram of the Sun for scale.
We have recently shown that 15 degree averages of horizontal motions
inferred from time-distance analysis agree very well with flows inferred
from the earlier ÒringÓ analysis. The latter method gives lower resolution
results with much less computation. Now that we have shown we can link
direct Doppler surface measurements to time-distance inferences and to
ring analysis inferences we can proceed to map the convection zone with
confidence.
Figure 10 shows a verification of the time-distance method of inferring motions. The
panel on the left shows a direct photospheric Doppler image while the image on the right
was determined for very near surface depths using time-distance methods.
The potential of local helioseismology to unravel the complex relationships
between interior dynamics and the generation and evolution of solar
magnetic activity is just beginning to be exploited. We eagerly anticipate
applying these techniques to long sequences of several rotations in each of
several years as the cycle progresses
Many of the SOI science objectives are referred to as ÒAssociated
ObjectivesÓ to indicate that they are not addressable by helioseismology
techniques but are nevertheless central to the SOI goals of developing a
more complete understanding of the relationships between solar interior
dynamics and the generation and development of magnetic activity. These
objectives are being addressed in a number of diverse techniques including
correlation tracking of granulation, detailed tracking of magnetic flux,
analysis of the distribution of radiative flux, observations of the shape of
the limb, and direct measurements of photospheric motion by Doppler
shift.
Initial correlation tracking analyses have enabled detailed mapping of the
network and allowed studies of flux migration to and along the network.
The network evolution can now be examined continuously for intervals
long enough to see the complex relationship between the emergence and
dispersion of magnetic elements and the more stable flow patterns. It is
now clear that any given snapshot of the network shows a partial
population of flux elements along the boundary, but a longer view shows
the full network traced by field elements moving first to the network then
along it.
Figure 11 shows an MDI high resolution magnetogram overlaid with lines of
convergence of the horizontal flow and with green dots showing the convergence points.
The measured flow is shown as colored arrows, red for inferred downflow and blue for
inferred upflow. The field is shown light grey for positive fields and dark for negative
fields. Only field above the background noise is shown.
Sophisticated magnetic flux identification algorithms have been developed
that have allowed tracking the evolution of individual magnetic flux
concentrations. It has been possible to measure the mean lifetime of
magnetic flux concentrations as a function of total flux in the concentration
and the rates of merging and fragmentations of concentrations. From a
chemical kinetic model and the measured rates it is possible to predict the
distribution functions for the number of flux concentrations of each
polarity given the total flux and the net flux in a region. These predictions
have been verified from low total flux regions, where the average total flux
4 to 8 Gauss, to dense plage, where the average flux is 100 Gauss. The
tracking results also allow the generation of a number of statistical
properties of the flux. These include the RMS velocity as a function of
the size of the flux concentration.
Software has been developed to identify and track supergranulation cells.
Initial results show that half the cells are lost in 20 hours. However, the
other half remain unchanged. We are currently constructing high
resolution data sets of 90 hours duration and will shortly have an initial
estimate of the mechanisms for the evolution of supergranulation.
High cadence multi-day series of magnetograms are leading to a new
appreciation for the dynamics of magnetic field patterns. A movie has
been prepared with the 96-minute cadence of full disk magnetograms from
April to November. It clearly shows that the solar magnetic field is
continuously changing even at Solar minimum.
Figure 12 shows a typical full disk MDI magnetogram. The movie shows one
magnetogram each 96 minutes for the end of April through September 1996. Full
resolution views are also provided for the south pole, north pole, and equatorial regions.
Figure 13 shows more detail in the magnetogram. All of the light and dark structure is
field. The background texture is leakage from granulation and is not magnetic field. The
field is seen to be mixed polarity within and on the boundaries of the supergranule
network.
An analysis of high resolution series has also been completed. A
previously under-appreciated mechanism for renewal of the network field
has emerged. Small magnetic bipolar flux element pairs are continually
emerging at seemingly random locations within supergranules. The
elements are rapidly swept to the cell boundaries, the two polarities
moving independently to different sections of the network where they
both cancel and replace the existing dominant polarity. This provides a
mechanism to refresh the dominant polarity while at the same time
changing the photospheric location of flux elements at a speed comparable
to the horizontal flow speed rather than the much slower random walk
time. The entire network flux is rearranged in less than a day by this
mechanism. There are profound implications for coronal heating on the
top side and questions of local field generation on the lower side of the
photosphere. We now see that at least 90% of the flux at minimum has an
origin in local structures. It is now clear that this is not simply due to
recycling of existing flux. There is sufficient flux emerging on the scale of
ephemeral regions that there must be some mechanism for local generation.
The rate of generation is such that all of the flux is replaced in 40 to 60
hours. It is not presently clear if the continual rapid rearrangement of
photospheric fields on supergranular scales is directly connected with the
continuous stream of tiny CMEs seen by LASCO.
Initial analyses of the shape of the Solar limb have also been surprising.
The expected accuracy of limb shape determination has been realized.
Results from the first SOHO roll observations have produced the most
precise determination of solar oblateness. There is no excess oblateness.
The p-mode ridges are clearly visible in the limb data with individual mode
amplitudes measurable in micro-arc-seconds (corresponding to a few
meters). The data also show evidence for a larger new component of the
Solar figure. There is evidence for ÒmountainsÓ on the sun with horizontal
scales several times supergranule scales and height scales of hundreds of
meters. These results were announced at the December AGU meeting in a
special helioseismology session. Simultaneously we have seen evidence for
similar scale structures in both direct Doppler data and in line-depth
observations. We have not yet combined the three types of observations
to see if this new component of structure is the tail of the supergranular
spatial and temporal spectrum or if it is an indication of a previously
unobserved phenomenon.
In addition to the SOI primary and associated science objectives, MDI
data can be an exceptionally useful component in multi-instrument
collaborative investigations. MDI has participated in many of the SOHO
test JOPS during the commissioning phase and many more in the 5 months
of normal operations since the end of MDI Dynamics. Calibrated data
with proper headers is now available to collaborators. Preliminary analyses
of early JOPS is very encouraging. For instance, the study of plumes at the
Solar south pole has demonstrated that there is a "flux tube" at the base of
each plume. We anticipate significant new understanding of dynamical
processes in the corona when we combine MDI magnetograms, EIT, and
LASCO observations.
Figure 14 shows the list of "Campaign" observation sequences obtained with MDI
during the initial 8-months of normal operations. Calibrated datasets from most of these
campaigns will be available for science analysis within a few weeks.
The MDI instrument can be operated in several modes depending on telemetry opportunities. These modes provide observations suitable for differing science objectives. These modes, in order of their priority for SOI are:
The mix of modes actually to be used during an extended mission would depend on the funding available for operations and for science analysis and on available DSN support for VC2/VC3. The optimum science return can be achieved by continuing operations in the same mode as at present. If the extended mission will be supported at nearer half the present funding level than the present level (as has been suggested by NASA) the operations of MDI would be severely restricted. The operations requirements for both scenarios are described below.
MDI Operations requirements for optimum extended mission
The optimum use of the SOI resources (MDI, the SOI data facilities and the SOI operations team) would be to continue to operate MDI nearly as it is now operated. We expect the software development efforts for science operations (including data processing through to calibrated datasets) to have ceased long before the extended mission interval would begin. While continued maintenance will be necessary there will be some significant cost reductions in science operations even while retaining full use of MDI. Some of the savings will come from freezing the development of new ÒobservationsÓ with the associated testing and processing code development. We are presently evaluating the efforts involved with taking VC2/VC3 data directly from JPL which would add data reformatting and data capture tasks but perhaps simplify the bookkeeping tasks. The costs of extended operations in a full-use of MDI scenario will be studied over the next 6 months.
Under normal circumstances, the bulk of MDI commanding can be accomplished in 30 to 60 minutes of NRT commanding per day. Software reloads and calibrations typically take 2 to 4 hours to accomplish, but can be scheduled well in advance of the planned activity.
For nominal operations, much of the MDI commanding could be accomplished by Delayed Commanding. This change would require several months of development and testing by the MDI operations team to insure that all activities can be adequately supported. In addition, this approach would severely decrease the MDI operational flexibility.
Continue as now.
A minimum of two offices are required; one for Julia Saba and one for Craig DeForest. An additional office would be useful for visitors.
The nominal data recovery rate for VC0/VC1 telemetry should remain better than 99.5%, or the MDI helioseismology studies will suffer. The data recovery for the MDI VC2/VC3 telemetry should be better than 95%; the limiting factor has been recovery of all the data collected by the DSN sites.
It depends which requirements are being considered. At this time we are investigating whether the MDI high rate telemetry (160kbps channel) can be delivered directly from JPL to Stanford. This has certain implications on the data recovery and the effort at Stanford to insure that data is received in a timely fashion. The need for PACOR processing of the VC2 data might be eliminated. The need for Real Time data during contact times would remain.
a) Routine: The present mode with c. 10 minutes VC2 on short passes and 1 hour during long passes would continue. The remainder of the long pass coverage would be VC3.
b) Campaign: For continuous contact times, i.e., yearly 2-3 month intervals and monthly 3-day intervals, the VC2 requirement could reduce to 10-20 minutes twice per day.
MDI Operations requirements for a severely restricted extended mission
As described above, the SOI primary science objectives would have priority to the exclusion of associated objectives. At present about 40% of the SOI support covers instrument operations and data processing through to archived calibrated datasets. If the present full-use mode of operations were to continue but the total funds were only half that available now the funds for science analysis would be about 1/6 of the present level. It would be pointless to operate the instrument at the full level but do only 1/6 of the science analysis, so in the underfunded case operations would be scaled back to allow science analysis for the primary science objectives to continue at least at a reduced level. In the reduced mode there would be no support of JOPS unless they required simply the primary SOI data. There would be no further development of observing sequences. This would allow shutdown of the MDI simulator presently used for verification of command sequences as well as simplified operations. In terms of the MDI operating modes described above, 1 through 3 above would have priority with modes 4 and 5 likely abandoned. The MDI instrument can operate nearly unattended in mode 2. We would support operations from the EOF only during intervals of continuous multi-day contacts. The MDI "science operations" would be handled from Palo Alto so there would be no scientist presence at the EOF. In this mode the science output of MDI would be a tiny fraction of the potential but at least the most unique helioseismology capabilities would continue to be utilized.
Much of the MDI commanding would be accomplished by Delayed Commanding. NRT commanding would be required only to deal with anomalous situations.
NRT commanding would only be used in anomalous situations.
Remote operation from home institutes.
EOF presence would be only to handle severe anomalies. We would leave a workstation but would have no normal presence.
There would be no SOI presence at the EAF.
The nominal data recovery rate for VC0/VC1 telemetry should remain better than 99.5%, or the MDI helioseismology studies will suffer. VC2/VC3 recovery during continuous contact times should be better than 95% as it is now. There would be no VC2/VC3 except during continuous contact times.
See note in normal operations section, but note the VC2 data volume would be limited to 10 minutes each pass.
a) Routine: VC2 would be required for c. 10 minutes on each pass. There would be no MDI VC3 use on long passes.
b) Campaign: For continuous contact times, i.e., yearly 2-3 month intervals and monthly 3-day intervals, the VC2 requirement could reduce to 10-20 minutes twice per day.
SOHO has been observing the Sun continuously in unprecedented detail since December 1995. SOHO images of the Sun and its atmosphere are available on the world wide web almost as soon as they are transmitted to Earth. In particular, images from the MDI show changes in the Sun's magnetic field, surface motions, and brightness. These and related SOHO data are ideally suited for development of a web-based educational outreach program.
We are developing graphical and video material and a set of web-based activities to interest, inform, and involve people outside the project in the study of the Sun. The target audiences include elementary and secondary school students as well as the general public.
TOP of page
