From Mike Fromm of NRL: I have put the message on my web-page along with a link to his powerpoint file. I have loaded the file as he sent it to me; but I have also gone to the official NASA toms website and and added into the powerpoint a series of the official toms global maps of aerosols for the same sequence. As was pointed out by Mike in his earlier message (see discussion of 5 February) the fire-generated cloud of aerosols that emanated from south east Australia on 18 january can be tracked all the way across to the Atlantic on the 26th. This is indeed impressive.
Mike
has posed some questions about the source of the aerosols observed over
New Zealand on Jan 30-31 also about the possible pyro Cb over Queensland
on Feb 4. If anyone has any info or response to this, please let
him know.
In
this and some of his earlier emails (see my discussion of 5 February) he
put forward the view that some mechanism (e.g intense Cb's) is getting
some aerosols to quite high altitudes, above 400 hPa. At the VRO
chart discussion a few days ago when we were looking at the smoke cloud,
I expressed the opinion that the large scale smoke cloud isn't all that
deep,
extending up to about 8,000 feet. Ward Rooney who was leading the
chart discussion replied "we have had reports where it gets
much higher than that. " Given the issues raised by Mike that the
smoke would have had to get
pretty
high to be transported so far so fast, perhaps Ward would care to elaborate???
John McB
Mike Fromm
Hi John et al.,
I've completed the chronology of TOMS AI maps. It goes from Jan 18 to Feb 4. toms.ppt
I want to make sure the individual who provides me the TOMS data is acknowledged. He is Dave Larko, who works at NASA Goddard Space Flight Center. Dave is also my POC when I have any questions regarding nitty gritty TOMS details, and he always comes through.
At the end of this message is a disclaimer about my mapping methods versus the official NASA version of the AI data and maps on the www (http://toms.gsfc.nasa.gov). Dave, I'm particularly interested in your comments on the disclaimer. Please chime in so I can be sure it meets your standards.
These maps provide lots of clues about changing blowup and transport conditions. Australia is pumping quite a bit of aerosol into the atmosphere from fires and dust in several locations throughout the country. Obviously the SE "corner" is a big (probably the biggest) source. But some aerosols seem to be coming from central regions and still others are spewing out of far northern Queensland. That's a lot of territory to cover and make sense of, and I'm surely incapable of doing it all justice. But I think it's safe to focus much of our attention on NSW and Victoria.
On
certain dates, you'll notice a huge blob of intense AI east of Aus.
Those blobs were no doubt injected into the tmosphere between measurements
(about 1 UTC) and advected away. On some of the dates with one of these
"fresh" blobs, the blob is pretty darn far from Aus.
This
tantalizing clue might well suggest that the "smoking gun" was a particularly
energetic and efficient Cb because the winds necessary to carry it that
great distance in under 24 hours would almost have to be at or near jet
levels. The best example of this might well be Jan 30-31. There
are two huge (in size and intensity) AI blobs near and east of NZ.
They went a long way in a short time if we assume they came from AUS.
I'm guessing convection broke out in a big way after TOMS's measurements
on Jan 30, probably during the later afternoon (prime summertime heating)
hours? Your discussion of Jan 30 and 31
leads me to think that this is reasonable.
One final note. Check out Feb 4. Go to Queensland. Near the coastal town of Karumba. Look at the nice red AI dot there. I checked the MODIS images (I inserted one into the last slide in the .ppt file) and sure as shootin' there's a monster Cb! Looks like that Cb lifted up something other than hydrometeors!
This is more fun than any single meteorologist should be allowed to have:)
Any and all feedback is welcome.
mikef
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
OK,
here starts the disclaimer...
Things
to note on how these maps are made and how to interpret them:
1.
these are pixel values and thus are not spread out regularly as are the
gridded TOMS data on the www. So some of the spaces between dots
is an artifact of this (the density of pixels is greatest at nadir under
each TOMS orbit; TOMs scans to both ides of nadir to make measurements)
2. The TOMS orbit altitude does not allow complete coverage between orbits in mid and low latitudes. So there is an orbit gap. Thee gaps are not marked in any way on my maps. If I ever have the need to make these maps more exhaustively truthful, I will do something to overcome this. But for the present purposes, I don't think it's a big deal.
3.
The fancy AI maps that NASA puts on the www are made from gridded data
and improved by other data-spreading attempts. For instance, there
are special challenges to assign data to the right place and time near
the dateline. To cover over any unsightly gaps the NASA algorithm
does what Dave calls "compositing." In a nutshell, that means filling
in blank
spots
within about 20 deg longitude of the dateline with data from an adjacent
day. My maps don't do anything so sophisticated. I just plot
the pixel at its assigned location on the date it was measured.
4. I plot the enhanced AI pixels after first sorting them in ascending order. I.e the high values are plotted last. I cheat in this way to make sure the most interesting pixels show up when there is overlap.
5. The data I requested from Dave was restricted to the box defined by 10-60S, 120-280E.
6. As you already know, I color code the entire high-end range of AI values. The NASA maps do not distinguish values greater than 4.7
7. The NASA maps ignore any AI pixel with a certain error flag value (5). I think I already mentioned this in an earlier email. So those pixels don't get into the gridding algorithm. The AI maps on the NASA www site would represent any blobs of such pixels as having no aerosol.
Reasons 1-7 are given to address why my maps probably look different than the web-official NASA version of these data. (Obviously the NASA maps have their purpose, and I have my more specific purpose-they're justified in making their compromises and I think I'm justified in making mine.)
Andrew Tupper
John,
I detect a certain note of cynicism in your wording below!
You
might be interested in this, although it's still a bit unrefined.
I've been playing with enhancements on AVHRR/3 & MODIS data designed
to highlight albedo anomalies on the surface of very cold cloud.
Rene and Mike had been looking at similar things for their Canadian
fire cases. I won't go into the details- it needs more work
- but I've put an image at:
http://arist.edu.kagoshima-u.ac.jp/andrewt/events/bushfires/ruangbush.jpg
(local copy here) that shows a wide
area around an opaque volcanic eruption cloud (Ruang, Sangihe Islands,
25 Sept 2002) that reached about 20 km altitude. There's
quite
a bit of very cold cloud in the image, but the bits in red and purple represent
very cold cloud with albedo anomalies (mostly volcanic ash and other noxious
stuff in this case). You can see the location of the eruption.
On the bottom frame is the same enhancement applied to the Canberra bushfires
(both images from Aqua/MODIS, processed on McIDAS). Note that there
is a fair bit of cold cloud in this image as well, but only the CBs over
the intense fires show up. I/we think that
this
might be good supporting evidence for high level aerosol transport for
this event - we're currently looking at more images.....
cheers Andrew
John
McBride
A
bit of discussion, dare I say even a slight argument on whether the aerosols
have to be transported to the upper troposphere to explain the observed
transports.
Item
One: extract of an email form Dale Hess to me:
John,
Thanks for the information.
Alan is
already checking the TOMS data as well as the other
satellite
data as part of his verification work on smoke
forecasting.
Alan is
now running warm start forecasts of the smoke
plumes
and the HYSPLIT model generally does very well.
Item
Two: Reply from me to Dale Hess
>
Alan is now running warm start forecasts of the smoke
>
plumes and the HYSPLIT model generally does very well.
>
Well if that's the case you and Alan will know the answer to the thing we have been "arguing about" on the discussions (all recorded on the synoptic discussion link on my web-page). To be specific, Mike Fromm argues that to account for the very rapid transport observed (e.g his example of the very large blob of high aerosol index that suddenly appears east of new Zealand on January 31 -- see the powerpoint presentation to which there is a link on my discussion of today 7 February), there is a requirement that the aerosols be transported to the upper troposphere, e.g by intense "pyro-Cb's".
For Alan to be producing good verifications of this transport, does he need to place the aerosols/smoke-source up at this high a level? The general large area smoke cloud usually tops at about 700 to 800 hPa.
And Now the answer. (Item Three)
I wandered along the corridor to Alan wain's office to see his verifications of the smoke transports using HYSPLIT. He showed me a few interesting examples from today and yesterday; but I wanted to know the answer to the Mike Fromm/Andrew Tupper question: Can you explain the large blob east of New Zealand on January 31 from transport of aerosols in the lower troposphere.
Well... below is a jpeg of the output from HYSPLIT running off 12.5 km resolution operational MESOLAPS fields. The output shown is run off three vertical line sources extending from the surface to3,000 meters. Coloured contours are some concentration threshold. the output shown is for a 12-hour forecast initialised at 12 UTC 30 January. You'll see there is smoke/high aerosol concentartion way out to the edge of the map, i.e near/over New Zealand.
So... it seems the answer is NO, you do not need to have the smoke transported to high levels by Cb's.....; Over to you Mike/Andrew
John McB
Mike Fromm
Hello all,
First let me say it's great that there is so much interest in this. Interest breeds scrutiny. And I think we need a lot of scrutiny because there is still a lot of mystery to all that is going on.
Andrew, thanks for showing those images. I was sure hoping you'd do that at this point. Andrew's information is unassailable IMO, based on my experience with these studies in the boreal zone, and based on the physical constraints. It's certain the cloud he is showing is in the tropopause region (from the IR) and it's certian that the composition at cloud-top level is distinct from any "normal" Cb cloud-top crystals. Nearby RAOBs constrain the altitude of those distinctive materials to be 12+ km. So SOME aerosols are definitely getting up high. ANother thing about TOMS AI is that, all other things being equal, AI increases with increasing altitude of the plume. And TOMS cannot see "through" opaque clouds, so evidence like the little red dot of very intense AI over the Karumba Cb on Feb 4 also limits the altitude of that aerosol to levels at and/or above the Cb, no doubt pretty high in that case. Which leads me to the Jan 31 discussion...
...the other quality of the AI that I refer to regarding the signal of an extreme event (in addition to the implied great distance travelled) is the value of the AI itself. 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, one that I think involves a significant fraction of the total smoke emission to be resident at unexpectedly high altitude.
But as we say, all that and a quarter will buy you a cup of coffee. What we really need is more and better obs of the downwins smoke. I have contacted a colleague in NZ (he's with NIWA) to inquire about any aerosol profile measurements or even anecdotal evidence of elevated smoke. So far no good data in reply. I've got irons in the fire with other folks with access to other profile data in the southern hemi, but am still waiting to hear back. If anyone has any idea or contact in this regard, that would be a real boon to this study.
Just how high some of these smokes got is still a wide open question. But Andrew's example establishes a very strong datapoint. I'm looking forward to getting more (and providing more) datapoints.
My slow-typing fingers have told me that this will be my last point:) In other cases I've studied or have direct knowledge of, there is rock-solid evidence of dense forest fire smoke plumes observed by eye at near-tropopause levels, plumes traceable to the exciting combination of factors we're seeing down your way. And there's great satellite data showing it gets into the lower strat. Obviously it's too soon to make any such claims in the case of your fires, but lots of pieces still seem to be aligned, IMO.
Thanks!
mikef
Mike
Fromm
Subject:
plume speed.
Hi John,
Just did another little back-of-envelope
calculation. I took the lat/lon of the farthest distant red AI blob
on Jan 31 (I used 51S, 180E)
and computed the distance between that and 35S, 150E). It's 2985
km. 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?
mikef