.
We also notice that most of this increase occurs in the downflows, which
because of their intermittency and turbulent structures are much more sensitive
to the horizontal resolution than are the fairly wellordered upflows.
In this section I present the effect of doubling the horizontal resolution, going from sol50 to sol100.
From Fig. 8 we see that the turbulent pressure increase
appreciable with increasing horizontal resolution, and the increase amounts
to 1.6% when correcting for the 20K difference in .
We also notice that most of this increase occurs in the downflows, which
because of their intermittency and turbulent structures are much more sensitive
to the horizontal resolution than are the fairly wellordered upflows.
Figure 8: The turbulent- to total pressure ratio. Pink curves are for sol50
and black curves for sol100.
I show both the total horizontal average (solid) as well as the
average in the upflow (dashed) and the downflow (dotted).
The upflow is probably mainly affected by conservation constraints.
The superadiabatic gradient displayed in Fig. 9 increases in the
downflow and decrease in the upflow leading to a narrower range of gradients and
a slight decrease in the overall gradient. As 
in the downflow
peaks 126km below the -peak in the upflow, the former
is likely to be smeared out in sol50 because of the small scale of the
downflow structures. With higher resolution, higher temperature contrasts can
be sustained, allowing for a larger gradient in the downflow.
Figure 9: The superadiabatic gradient . Pink curves are for
sol50 and black curves for sol100.
I show both the total horizontal average (solid) as well as the
average in the upflow (dashed) and the downflow (dotted).
The broadening of the -peak in the upflow, might be due to the
more efficient enthalpy transport in the downdrafts, prompting for more flux in
the upflow. This increase of flux in the upflow is supplied both by an increase
in the filling factor, i.e. the fractional area of matter taking part
in the upflow, as well as an increase in in the upflow at the
height of the peak in the downflow .
The change in in the upflow is somewhat counteracted by an
small and almost symmetrical (i.e. opposite) change in the vertical
velocities in the upflow (cf. Fig. 10).
The downflow speed is increased by about 8% from the top of the convection zone and down to a depth of about 2Mm - more than twice as deep as the superadiabatic region. This increase in downdraft speed is probably necessary in order to maintain mass conservation, but also adds to the increase of flux in the downflow.
Figure 10: The vertical RMS velocity 
. Pink curves are for
sol50 and black curves for sol100.
I show both the total horizontal average (solid) as well as the
average in the upflow (dashed) and the downflow (dotted).
The horizontal velocities decrease though, which seems to violate mass conservation as the density stratification is hardly altered. The explanation lies in the fillingfactor. Near and just below the top of the convection zone, the fillingfactor is slightly larger for sun100, which means that a smaller amount of matter has to overturn to the downflow, hence the lower horizontal flow velocities.
Figure 11: The horizontal RMS velocity 
. Pink curves are for
sol50 and black curves for sol100.
I show both the total horizontal average (solid) as well as the
average in the upflow (dashed) and the downflow (dotted).
The rearranging of fillingfactors and velocities, also cause the kinetic
energy flux to increase, whereas the total enthalpy flux is basically unchanged,
causing a net decrease in total flux, hence the 20K decrease in .
Last updated [an error occurred while processing this directive] by: trampedach@pa.msu.edu.