OK:  This is the way I see it.

For the past several weeks we have had a large smoke cloud over south eastern Australia, the source of smoke being the ongoing fires in North eastern Victoria and southern New South Wales.  The extent of the smoke cloud has been monitored via surface synoptic observations, mainly of visibility but also of present weather observations of smoke, by visual channel satellite imagery and by MODIS satellite, as documented in our discussion of 5 February.    Listening to the aviation mets at VRO chart discussion in the mornings, it seems this smoke cloud extends from above the nocturnal inversion up to about 8,000 feet.... i.e it is a lower troposphere phenomenon.  Consistent with that, NMOC have been running HYSPLIT for the event with various smoke sources extending from the surface to 3 km.  These have been verified by Alan Wain and have been passed on to the VRO; and generally everyone seems happy with the performance and the verifications.

    So... its a simple story, except we have the Mike Fromm and Andrew Tupper camp producing evidence that, when there is a pyrocumulus about, the smoke/aerosol goes into the updrafts and you end up with a large amount of aerosol being ejected into the upper troposphere at 400 hPa or higher.  They have produced good evidence of pyro-cu cloud tops  extending to 20 km (which as you know is a lot higher than 400 hpa, presumably it is in the stratosphere).

Last week Mike Fromm sent us his raw TOMS aerosol output (see the powerpoint display we have discussed in earlier emails).  Some of the interesting features  include the following:

*    The current Australian fires show up as a huge source of aerosols;

*    The aerosols from the fires  can often be tracked very large distances: the best example being the aerosols from the 18 january blow-up which could be tracked all the way across the Pacific and into the Atlantic Ocean;

*    The large area of high aerosol index that appeared east of New Zealand on the MODIS charts of January 31 were taken by the Tupper-Fromm camp as good evidence that the transport was at upper tropospheric levels.  This was  on the basis of the high wind speeds (minimum 34 m/sec) that would be required to transport smoke that far in about 12-24 hours.
 

On Friday I suggested we may have killed the upper-tropospheric viewpoint as I put up a chart from the Meso_LAPS HYSPLIT output (supplied by Alan Wain).  This chart actually did have a smoke cloud at lower tropospheric levels getting to around the location of new Zealand (which unfortunately was off the edge of the map) in 12 hours.  In this particular situation, Mike Fromm's requirement of 34 m/sec or 60-70kts) to transport so far so fast does actually seem possible.

I have placed links to a couple of charts for 12 GMT 30 January on my web-page under today's date.
GASP 500 hPa hemispheric
LAPS 850 hPa
TLAPS 850 hPA
LAPS 700 hPA
TLAPS  700 hPA

The charts speak for themselves.  there was a trough sitting over southeastern Australia with a heck of a jet at 500 hPa shooting air straight across to New Zealand (wind speeds at 500 hPA over 100 kts all the way from the fires location to New Zealand and beyond).  The jet/highwind   speeds continue down through the lower troposphere (which is a lot of the reason why Jan 30 was a bad fire day); but with diminished magnitudes.  At both 850 and 700 hPa, however, the wind speeds are still between 40 and 60 kts over the Tasman.  Thus addressing Mike Fromm's question sent to me over the weekend:
   "Even if it had a full day to get there from its AUs source, it would have to have travelled 34 m/s sustained.
    That seems to be a condition that is restricted to jet altitudes.  What do you think? "

The answer is not clear.  Looking at this particular situation, it is well possible that the aerosols could have got that far at lower tropospheric levels.  The only way to resolve this is to re-run HYSPLIT from GASP analyses for that situation, which Alan Wain tells me he can do..... I'll send him a note requesting it.

The other note from Mike Fromm over the weekend referred to the actual value of the Aerosol Index itself.  Quoting from him,  " When TOMS AI gets to values above 4 or  so, your getting into the rarified "territory."  WHen it gets up to 7 or 8, it's really noteworthy.  When it gets into double digits, you're talking about an aerosol condition that just doesn't happen at altitudes below (rough guess coming) about 500 mb.  So regardless of its implied speed of advection, those huge, intense blobs EAST of NZ on Jan 31 are almost certainly an indicator of an extreme situation".

I know very little about the interpretation of TOMS (which I'll get back to below), but from my uneducated viewpoint it seems this may be the best evidence in favour of the upper tropospheric transport of material swept up to high altitude in pyro Cb's.
Looking at the output from the TOMS data in the powerpoint display, the display for 31 january shows two enormous blobs east of New Zealand.  I have put MikeF's single image for that date up on web-page for today.

Referring to the two large areas of high index east of New Zealand: when I showed this image to Alan Wain last week, his immediate response was "why are they cut-off at the western edge... why don't the plumes lead all the way back to the source over Australia?
This may actually be the answer:  here is a hypothesis:

The smoke cloud we in the Bureau have been tracking is a lower tropospheric phenomenon and is the advection of plumes  coming directly from the fires emitting smoke in dry-convective plumes up to about the 800 hPa level.  These plumes are being constantly regenerated by the fires which are ongoing.  The residence time of the aerosol/smoke in these plumes is probably only of the order of 24 hours; but the smoke cloud is continuous (as documented in our discussion of 5 February) as the fires themselves are ongoing.  We presumably see the aerosols in these smoke clouds on the TOMS maps but only in the region close to Australia.

The hypothesis is that when we see on the TOMS output aerosols from Australia going large distances, then that is always the aerosols that have reached the upper troposphere in the pyrocu.  An argument in favour of this is that the pyrocu are relatively short-lived of the order of 6 hours) and so the emissions are strongly episodic, which would agree with the observed TOMS output.  A possible argument against this scenario is that the pyrocu have been relatively uncommon during this sustained fires event... we on this discussion list not having noticed them until we spent that energetic day on them on 30 January (see synoptic discussions of 30 January and 1 February).   In favour of the pyro-cu mechanism, though, are the observations of the great heights to which they reach.  Quoting from a couple of emails from our discussion of  1 February:

"from memory we had radar returns up to 12km from the fire storm that hit Canberra on Sat 18th Jan.
During Thu 30 Jan we had radar returns to at least 11km from the fires over NE Victoria, these fires
produced at least one pyro-CB which then started other fires down stream. " (Tony Bannister)

"Tops were also 12km SW of Delegate (confirmed from Sydney) on Jan 30 associated with pyro generated CB. " (Milton Speer).
 

So... all this is pretty interesting. In summary it seems there is evidence for two distinct types of aerosol/smoke cloud: (i) the large slow low-level hazard to aircraft and health that has been drifting around over Victoria and giving us the red sun in recent weeks, and (ii) an upper tropospheric region of aerosols the source of which is detrainment in the upper troposhere/lower stratosphere from the pyro-cu.

The caveat is that I don't know anything about TOMS data.  Despite the fact I announced last week that I  intend retiring from the synoptic-discussion business and going back to doing more normal research, writing papers etc., I'll do some reading.  There are a few references on the NASA TOMS site:
http://toms.gsfc.nasa.gov/
I'll start from there.

cheers

John McB

Mike Fromm

Another one overnight from Mike Fromm with some really interesting background on aerosls from fires and on interpretation of TOMS aerosol data..  I have edited his message down as it contained a level of chat that probably is not of widespread interest.

I have a couple of comments:

1.  He expresses surprise that we hadn't observed pyroCu's much... In response to that, I guess I had overstated the case for their rarity in Oz.  As we all know, they are not uncommon.... what I had never realised is that they are no so much pyrocumulus but more pyrocumulonimbus, and that they extend up well beyond the tropopause.

2.   Given the exciting signal we are seeing here... transports from our fires all the way across the Pacific, it is clearly worth documenting the  event.... So.. anytime anyone notices pyrocu on either radar, satellite image or visually, could you let us know, and save the radar images if possible (they being that much easier to save in real-time)

John Mcb

This whole fire/smoke/convection/strat transport thing started in 1998.  There was an inexplicable increase in aerosol layers over northern midlats at altitudes 3-5 km above the tropopause, observed by two satellites and lidar. We traced these back to Canadian forest fires.  There was a huge camp ou there who scoffed at this. (Some of the most fun I've had in my life has been showing those results and watching the various skeptical reactions, then working to sell the argument.)  We wrote up our results in GRL, May 1, 2000.  Considering your last comment in ytour email, I think you'd get something out of it...we
use TOMS AI data (as well as other data).

[SNIP]

Even with the GRL written there were MANY unanswered questions, so we continued to study the world for other occurrences of big fires, big blowups, and stratospheric smoke.  Then along came the northern summer of 2001.  It was like deja vu all over again:)  Lots of aerosol layers in the strat (up to 450 K theta).  Rene Servranckx and I think we discovered the "smoking gun", and individual pair of thunderstorms that did the work.  We've got a GRL that is within days of acceptance (it's already been reviewed once and revised (just minor revisions)). This work is a case study of that blowup event, and in it we reveal TOMS, AVHRR, and SeaWiFS, and GOES data that pretty much seals the case IMO.

[SNIP]

I think your assessment is pretty accurate.  No doubt the biggest abundance of smoke is from the PBL up to a few km from the sfc.  But episodically high injections are happening too.  It's a little surprising to me that the pyro-Cbs haven't been more readily detected, but I know so little about your land that I don't appreciate the expanse, the radar network, and so on.  As for TOMS, I was told by Jay Herman at GSFC that TOMS can't even see below about 1.5 km, so even an infinite concentration of smoke or dust would not be detectable below that altitude.

Thanks again for your help and interest.

mikef