Title: 
==============
Non-Axisymmetric Large-Scale Photospheric Motions

Lead Investigator:
=======
R. S. Bogart

Team Members:
==================
T. Bai (if agreeable)
P. H. Scherrer (if agreeable)

Local Coordinator:
==================
R. S. Bogart

Lead Programmer:
=====
J. Aloise

Technical Summary:
==================
Averaged MDI Dopplergrams show unmistakeable evidence for large-scale
velocity features localized to narrow latitude bands.  These bands may
be evidence of steady zonal flows (torsional oscillation), localized
meridional flows (convective rolls), radial flows or lineshifts.  There
is evidence that they are not continuous across the entire disc, however,
indicating that they may be large-scale, but non-axisymmetric flows.
They could also be associated with supergranule motions, as they are
of comparable scale.  I propose to investigate these structures with
a view to (a) establishing their nature, i.e. the vector velocity fields
or other phenomena responsible for them, (b) intially characterizing their
size scale and lifetimes or stability, and (c) their probable effect on
estimates of truly axisymmetric velocity structures.

Proposal Text:
=======================
The Dopplergrams taken during the first long uninterrupted set of MDI
observations March 27-31, 1996, when averaged together, even
before calibration, show clear alternating zonal bands of apparently
enhanced and diminished velocity superimposed on the general pattern
of global rotation (Fig. 1).  The same structures are also plainly
visible in only 6 hours of high-resolution Dopplergrams averaged together
(Fig. 2).  These zonal bands appear to be continuous over a sizeable
portion of the globe (30 degrees in longitude), but *not* over the entire
visible disc.  Their width is about 3 degrees in latitude, with spacings
of roughly 10 degrees in latitude.  They show up clearly only at latitudes
higher than about 30 degrees, but do not appear to be otherwise strongly
affected by disc position.  Their amplitude is of order 40-80 m/s.

There are several possible explanations for these structures, and the
immediate evidence for and against any of them is not obvious.  They
could be what they look like, bands of enhanced and diminished zonal
velocity, similar to the torsional oscillation flows, but of sub-global
extent.  Their apparent continuity across the central meridian, or
equivalently their lack of modulation with meridian distance, would argue
against this.  They could also be zones of meridional flow, or of radial
flow, or of some more complicated velocity.  The component vectors of the
equivalent velocity field, if that is what the structures represent, can
be determined by their symmetry properties with respect to the angle of
observation, provided that the structures are sufficiently long-lived,
and they appear to be.  An alternative explanation is that the structures
are artifacts caused by smaller stable velocity or line-shift structures
rotating across the field of view, especially supergranules.  Arguments
against this are that the structures are not symmetric about the disc
center (they do not appear at low latitudes even near the limb), and
the length of the coherent bands seems to be independent of the length
of the averaging interval.  Nonetheless it should be possible to test
this hypothesis simply by attempting to identify the endpoints of the
zonal bands with corresponding supergranules.

The method I propose to use to isolate and analyze the zonal bands is
to first form a very accurate observational map of the limb shift
(including the anisotropic component critical to measuring mean
meridional flows) by forming long-term averages as simple functions
of radius and azimuth as the apparent disc size changes over its
full normal range; this will require 6 months of data, but reasonable
range can be achieved with a much shorter interval if the radius is
changing rapidly.  Once this phenomenological limb shift is removed
from the calibrated Dopplergrams, they can be mapped to appropriate
heliographic grids as necessary and low-pass filtered to remove
oscillations and granulation noise.  In general remapping will not
be necessary or desirable, as the distinctions among the various
hypothetical sources depend on viewing angle.

Whether or not the zonal velocity structures can be unambiguously explained,
their mere characterization should suffice to estimate their effect on
estimates of axisymmetric flows, provided that they can be followed for
one or more complete rotations.

Required Observations / Data:

The data required for this project are simply full-disk Dopplergrams
(high-res Dopplergrams may also be useful), suitably sampled and filtered
to suppress oscillations and to be able to identify super-granulation.
Dynamics Program data are ideal; any other long runs of Doppler data
exceeding about 4 hours' duration, even data sampled as little as a few
times an hour, may be useful.  Certainly no special observations are
required.  However, one very useful, even critical observation for determining
the phenomenological limb shift is a series of at least several hours of
Dopplergrams made with a non-zero (and preferably large) position angle.
I propose that the desirability of making such observations be considered
when observing sequences are scheduled for the rare opportunities afforded
by extended spacecraft roll maneuvers.