There have been few systematic studies of thermal structure; the work we
did in the 1970s at Reading University with an insrtumented Falke was
reported at OSTIV Conferences in 78 and 81. The main focus of the research
was heat and water vapour transfer from surface into the troposhere (the
'fuel' that powers the heat engine we call weather). However, it was not
difficult to extract from the data some useful information of thermal
structure. For the purposes of the analysis, a thermal was defined as an
area of positive vertical air motion greater than 1 m/s and more that 50
metres horizontal extent. An important finding is that above about
one-third of the distance to the inversion, there is no significant
temperature difference between the thermal and surrounding air; near the
inversion the temperature is actually lower since the warmer air above the
inversion is being mixed down around the rising air. Humidity is a
significant indicator, H2O molecules being lighter than O2 or N2.
Therefore thermal 'detectors' based on temperature are a waste of time. It
would be nice to have a remote sensor detecting movement of entrained dust
particles, but this would take all the fun out of soaring.

The best thermal indicator remains to be a glider flown by a good pilot
circling tightly and going up fast. (Or a soaring bird).

At 20:14 02 November 2014, krasw wrote:
On Sunday, 2 November 2014 21:03:39 UTC+2, Andy Blackburn wrote:
I'm curious what your running the numbers looks like. I admit I don't
hav=
e a good mental model for how thermals work from a thermodynamic and
aerody=
namic perspective. I thought it had something to do with the fact that
warm=
air was less dense and therefore buoyant. How that buoyancy accelerates
a
=
volume of air until some form of resistance at the edges progressively
resi=
sts the acceleration and a steady rate of upward velocity is reached is
bey=
ond my understanding at a level detailed enough to relate thermal
strength
=
to temperature differences.
=20
As to the temperature gradient across the thermal - I'm not sure it's
lin=
ear from the edge to the center. Imagine a volume of air rising at 500
FPM.=
Presumably you have some mixing at the edges but the rest of the heat
tran=
sfer would mostly be conductive over a period of 10 minutes before the
ther=
mal reaches, say, 5000'. I'm not sure what all the coefficients are, but
i=
t isn't 100% obvious to me that you'd end up with a linear temperature
grad=
ient all the way to the center of the thermal since there is so much new
wa=
rm air being introduced continuously from the bottom, there isn't much
time=
for heat to transfer to the outside air and air isn't that great a heat
co=
nductor in the first place. Have there been studies done?
=20
9B

I can't quote any sources but I bet digging into Google Scholar would
resul=
t studies (boundary layer physics would be good place to start). I know
tha=
t warm air in thermal bubble or column is little bit warmer (numbers
quoted=
earlier are realistic) than surround. That temperature difference is
maint=
ained with altitude (per hydrostatic equation), and thermal is
surprisingly=
"closed" system, mixing at the edge of ascending air is quite small
compar=
ed to thermal volume. It's true that average temperature gradient is not
re=
alistic figure, the gradient is probable decade higher at edge of thermal
t=
han figure I threw out of my sleeve.

I think running constant temp difference analysis would be quite easy
with
=
simple and cheap linux computer, temp signals going through DA conversion
t=
o digital domain. Sensors could be calibrated over time to match each
other=
, so small differences could be readable, even if absolute accuracy is
some=
thing like 0,1 degrees. If this has been done in 80's, technology
probably
=
was analog electronics. Nowadays it would be more software project.