CONVECTION
From the 32nd Liege Colloquium
The front page of the proceedings:

Text and image viewgraphs from the talk
Mostly provided as a service to those who heard the talk in Liege; others may
find the text and figures too sparse and unstructured.
A more complete version of the talk is available as a
preprint (color.
or greyscale).
About this document
o Significance for stellar structure and evolution
- +
- direct, on structure and mixing

- +
- indirect, through activity and mass loss
- +
- without it, stars would be simple & boring
o Realistic supercomputer models

- +
- Quantitative comparisons with simpler models
- +
- Qualitative understanding
- -
- surface cooling, bulk CZ, undershoot
- +
- A lot remains to be understood
- -
- mixing, differential rotation / dynamo, activity, mass loss
Direct Influence
o Convection zones, most stars:
- +
- temperature/entropy jump at the surface
- -
-
for many stars (Sun)
- -
- shallow cellular surface patterns
- -
- does mixing length / C&M have anything to do with it?
- +
- nearly isentropic interiors
- -
- narrow, filamentary, non-stationary downdrafts

- almost isentropic, slowly rising background
- +
- transition to radiative interior
- -
- weakly subadiabatic lower part of CZ (
)
- -
- undershoot region below CZ (
)
- -
- mixing layer, character?
Surface Layers / Granulation
o Rapidly decreasing scale height towards surface
- +
- forces smaller and smaller overturning scales

- +
- up/down mixing by overturning (Böhm-Vitense!)
MLT
o Surface radiation
- +
- scales of the order of
-
-
- +
- extremely rapid energy loss (
)
- +
- optically thin layers subadiabatic; i.e. cooled by convection
- +
- downflows reheating by entrainment / mixing s

- -
- cf. mass fractions
Numerical Modeling
o Required physics for quantitative comparisons
- +
- good continuum opacities (
dominates)
- +
- ionization and dissociation equilibria
- +
- schematic line opacities
line blanketing
- -
- back warming of continua
- -
- line ``cooling'' of higher layers (actually heated)

o Required numerics
- +
- 3D, of course (2D can match basic properties)
- +
-
resolution to match best observations





Shock formation / Turbulent Pressure
o Even in the Sun
- +
- horizontal flow; vertical stand-off shocks
- +
- surface elevated slightly by
o Giant stars
- +
- significantly supersonic
- +
- ``fountain flows'';
free fall
- +
-
in horizontal planes
- +
-
elevates surface several scale heights
Superadiabatic Region
o Helioseismology pinpoints trouble in superadiabatic region
- +
- ML
frequency error at
mHz
- +
- Canuto & Mazitelli better, but
- -
- right result for the wrong reason
o 3D models
- +
- direct analysis
structure is the main effect
- +
- kernel analysis
matches observations
- +
- structure differs qualitatively from C & M
- +
- average T(z) structure as smooth as ML, C & M is steeper
- +
-
pulls high
down, not a steeper T(z) as in C & M
Evidence from the Red Giant Branch
o Giants of increasing luminosity
lower surface acc. of gravity
- +
- increasingly violent and supersonic convection
- +
- yet, mixing length models calibrate to nearly constant

(these plots are courtesy of Raul Jimenez, jimenez@nordita.dk).

Bulk of Convection Zone
o NOTE: Density and pressure ratios
- +
-
!
- +
-
!
o Yet, the only source of cooling is the surface
- +
- a gram cooled at the surface
a ton at the bottom
- +
- is the (flee at the tip of the) tail wagging the dog?
- -
- well, it takes a million
o nearly isentropic interiors
- +
- narrow, filamentary, non-stationary downdrafts
- mixing / entraining, thus reducing amplitudes
- +
- almost isentropic, slowly rising background
Undershoot
o Transition to radiative interior

- +
- weakly subadiabatic lower part of CZ (
)
- undershoot region below CZ (
)
- +
- mixing layer, mostly inefficient wave motions


o Chemical Mixing / Lithium Burning
- +
- Steffen & et al paper on undershoot
Indirect Influence
o Chromosphere
- +
- convection drives magnetic activity
- +
- provides wave heating of non-magnetic chromosphere
- +
- Stein & Carlsson, AjJL 1995 (in press)
-
-
no temperature rise in non-magnetic chromosphere

o Corona
- +
- Galsgaard & Nordlund; CZ shear drives heating and flaring coronae
Current sheets in a magnetically dominated plasma friven from two
boundaries:

A close-up of the region near the bottom:

A transparent view, showing field lines on both sides of the
current sheets:

A low view, showing protrusion of weaker current sheets from
the main one:


Resolution: 136 x 136 x 136 (above), and 88 x 88 x 88 (below).

Red: magnetic field strength isosurface. Green: Electric current
isosurface.
o Winds
- +
- Stellar winds similar to the Solar wind
- +
- Reimers law
- +
- Consequences of Giant's supersonic convection?
o Giant star mass loss
- +
- partly from variability
Conclusions
o Significance for stellar structure and evolution
- +
- direct, on structure and mixing
- +
- indirect, through activity and mass loss
- +
- without it, stars would be simple & boring
o Realistic supercomputer models
- +
- Quantitative comparisons
- -
- with observations
- -
- with simpler models, for calibration
- +
- Qualitative understanding
- -
- surface cooling, bulk CZ, undershoot
- +
- A lot remains to be understood
- -
- mixing, differential rotation / dynamo, activity, mass loss
Aake Nordlund
Wed Jul 5 13:34:40 MET DST 1995