Noel Davidson

To: tropical-storms@tstorms.org
Subject: [Tropical-storms] ram&chat rainfall, intensity and dvorak
G'Day Folks,

This is in response to some interesting comments and data provided by Roger E., Masashi Nagata and Johnny C. on Rammasun and Chataan. I
dislike long comments, but there are some related issues here, which are difficult to precis.

1. RAINFALL from CHATAAN

If the rainfall totals ~ 1 Aug were unusually large for Chataan :
Was it because this system was slow-moving, or were the totals large even for a slow-moving storm?
Why was there only a weak response in the central pressure and winds to the apparently-large, latent heating?
Reasons: SSTs? Shear? Size? Some compensation between effects of clouds and environmental processes? Other?

2. INTENSITY and DVORAK

Attached as gifs are estimated and forecast tracks and central pressures from our operational TC system for the same base time,
for Rammasun and Chataan.  ram central pressure,  ram track
                                                chat central pressure,  chat track
Some interesting points are :
(a) Forecasts are quite good except for Chataan's track beyond 48 hours.
Even forecast intensities are quite encouraging. Only one case here, but we have numerous others and some encouraging stats to suggest
that some skill is possible. {So agree with Johnny that hires models are capable of skilful intensity forecasts - but more work yet}

HOWEVER
(b) Both storms were initialised with CP=~995 and max wind of ~ 24 m/s
    Dvorak for a CP=995 has a (JMA) max wind of ~ 22m/s
    ------So far so good!
    Chataan with forecast CP=982 at 30 hours had forecast max wind of ~ 35m/s.
    Ramassun with forecast CP=968 at 72 hours had forecast max wind of ~ 36m/s.
    ie, "balanced" hires model can produce nearly the same max wind  with vastly different CPs. (Note that Ramassan was forecast to be much larger)
    Dvorak for a CP=985 has a max wind of ~ 26m/s
    Dvorak for a CP=965 has a max wind of ~ 35m/s

(c) Forecast rainfall maxima for Chataan and Ramassun at the time of the similar wind maxes were 450 and 300 mm respectively.

Mindful of the limitations but also the attributes of numerical models, my questions are :
*    Are storms with the same max wind and rather different CPs common?
*    Can we tell, and how do we initialise them?
*    On what field (CP or wind) should we INITIALISE and VERIFY - maybe both ??
[I prefer CP since we know even less about inner wind asymmetries.]
*    Are tabulated pressure-wind relations in their current form a near waste of time?
*    Dvorak provides a remarkably good first guess but how do we improve on these structures?
*    Are latitude, size and intensity change important for these purposes?
*    Are rainfall forecasts from hires systems of any value?

3. This is our forecast track (6-hourly data) initialised 48 hours prior to the time for which Dr. Nagata provided data from Miyako-jima.

                           Central Pressure   VMAX at 70m (lowest lvl)
  NAME    DATE    TIME    LAT   LONG   CP      LAT   LONG    MAX
RAMMASUN 20020701 1200   17.5  131.4  974.2   18.3  132.4   41.6
RAMMASUN 20020701 1800   18.9  130.2  969.0   18.3  130.5   41.5
RAMMASUN 20020702    0   19.8  128.8  966.4   19.2  129.3   39.6
RAMMASUN 20020702  600   19.9  127.9  963.9   19.5  128.7   39.3
RAMMASUN 20020702 1200   20.2  127.8  957.6   19.5  128.1   44.4
RAMMASUN 20020702 1800   20.8  127.5  954.1   20.4  127.9   47.1
RAMMASUN 20020703    0   21.3  126.7  952.8   20.7  127.0   44.5
RAMMASUN 20020703  600   21.6  126.3  951.1   21.3  126.9   47.0
>>Miyako-jima
RAMMASUN 20020703 1200   22.0  125.7  948.7   21.7  126.4   46.2
RAMMASUN 20020703 1800   22.5  125.1  950.7   22.2  125.7   45.2
RAMMASUN 20020704    0   23.2  124.5  954.5   22.6  125.2   43.4
RAMMASUN 20020704  600   23.7  124.0  955.8   23.2  124.8   41.8
RAMMASUN 20020704 1200   24.0  123.4  957.7   23.8  124.3   39.8

The forecast track at 48 hours has an error of ~ 250k at the time the "real" storm was near Miyako-jima.
(The 72 hour track forecast is very poor).
The forecast CP is again quite nice.
The FORECAST LOW LEVEL WIND MAX AT THIS TIME IS IN THE SE QUADRANT. Is this verifiable? I would send a graphic but don't wish to inundate people with images.
Dvorak for CP=945 has a max wind of ~ 42m/s -- I can live with the  Dvorak-model comparison here. The times at which RAMMASUN was forecast to have a central pressure of ~957 have somewhat interesting VMAXs.

3. Did you ever contemplate how many ways there are to misspell : RAMASSUN. A<->U, single/double M, Single/Double S, M<->N. I think I have encountered every combination, many of my own making, in trying to run our verification program.

Bye for now. Hope that I haven't annoyed Roger too much? Go the Bullies against the Bombers - even with Hird and Lloyd back. {I wrote this on Friday but didn't have time to send it - the game was a draw!!}

Noel Davidson

Jeff Callaghan

Noel and others
We up here in Queensland are very interested from heavy rain in tropical and sub tropical cyclones as it accounts for most of our weather related
deaths  In all the events we have studied the heaviest rain occurs where  the winds up to 500 hPa back with height (veer NH) through an angle <90
degrees. QG theory would indicate that this is a  region of forced ascent. The EC analysed winds (chosen because we believe they don't insert a bogus
circulation) at 1200UTC 1 July 2002 indicate veering winds with height between 850 and 500 around the centre of Chataan. This veering appears to
be associated with a slope towards the SW of the circulation. Examination  of the SSMI and TRMM rainfall seems to indicated that this was around the
time of the heaviest rain.

Locally this warm air advection pattern is known as snake oil as it appears  to be explain many weather phenomena and as we have discussed with Roger
Smith, how far can you extrapolate  QG theory towards the equator.
Jeff