Re: 天氣概況~2008/06/25~風神過後~
※ 引述《caterlanse (卡特--藍)》之銘言：
: 這次的風神颱風就預報上而言要檢討的地方當然很多
: 像是中高層高壓為何如此持久?連續幾波槽線都撼動不了?
: 颱風都高低分離了~為何中低層引導氣流還佔不了上風~中心反而追著對流雲團跑?..等等
: 問題很複雜~也不是短時間內可以釐清的
: 像我這種外行看熱鬧的就更一頭霧水了..囧
: 不過各單位依照預報結果進行路線預測的過程其實並沒有特別的錯誤
: 畢竟做預報必須要有所本~當然每個專家意見可能不同~
: 也有一開始就篤定颱風不會來的神人
: 不過一般做預報大都是信任模式的結論
: (套句CWB預報員的話~數值模式可是氣象學原理真正實踐的產物~)
: 在主流預報模式的結論約略一致的情況下自然也就會報出類似的結果
: 然後隨著風神不斷的製造意外~模式輸出的結果也跟著不斷修正
: 接著主觀的路線預測也就不斷更改~看起來就顯得手忙腳亂的樣子
: 只是最後的結果~不準就是不準~
: 這種不是YES就是NO的情況都是做天氣預測必須面對
風神的預報誤差,也引起許多國外學者的討論
摘要部份學者原文的討論給大家參考,也興趣的也可上討論區看
http://www.tstorms.org/tropical-storms/
但需要web管理者註冊。以下是熱帶氣象學家Roger Edson的看法
Yes, I watch the forecasts all of the time as well...
and have not seen this kind of systematic error in quite awhile.
I knew JTWC was following the NWP consensus...as was all of the other
forecast agencies in the region, as it is extremely difficult to go with the
1 or 2 extreme outlyers (like the UKMET was for many of the runs).
I've divided the track into 3 parts to help evaluate its motion:
a) East of the PI: As it developed, it had been going WNW for several days.
I don't think the models realized how strong the ridge was to the north, and
as Mark D indicated, the models kept breaking down the ridge and moving the
TC (remnants) north and east of Taiwan (most models could not even hold
onto the circulation, however).
b) Within the PI: Here the TC continued to rapidly intensify inspite of hopping
over the various islands (not unusual for TCs in this area), and then made
almost a 90deg turn to the right (north) as if it was going to finally
follow the recurve forecast. This abrupt turn was also unusual as
I would usually expect it to continue through the islands into the South
China Sea, before recurving. Anyway this was the time (maybe the only
time during its life cycle) when the vertical extent of the TC was fully
developed and the pressure tendency was toward the north and it actually
followed the mean tropospheric steering flow...and the NWP forecasts.
c) In the South China Sea: Thanks to a combination of strong northeasterly
shear and some dry air to the north, the TC came out of the PI...and STAYED
....as a 'half typhoon'---seemed like it was almost fully developed in the
southern half of the circulation, and had hardly any deep convection
(except for some brief periods of time over its center) over its northern
half. Clearly, the various NWP forecasts did not take into consideration
(or physically incorporate) this strong upper level flow (above 400mb).
****This seems to be something the forecasters always sees, especially
in the water vapor imagery, but the NWPs never seem to understand...
(similarily, we see this all of the time with TUTT interaction).
Normally one would not forecast a TC to move directly into the shear!
In addition, when the TC is not fully developed in the vertical,
the models have a problem of identifying the 'apparent' steering flow as
I don't think the vortex and the environmental interaction is accurately
modeled.
Roger Edson
Tropical Meteorologist
--------------------------------------------------------------------------
Johny Chan的回應(from 香港城市大學物理系)
As to the forecasts, I have been following the prognostic reasoning given
in the JTWC forecasts and as everyone has said, the main problem appears
to lie in the prediction of the strength of the subtropical high.
The question is why its continued westward push was not predicted
by most models when we expect global models to have a rather good handle on
the large-scale features such as the subtropical high.
One possibility is the influence of the TC itself.
Is this the reason why the UKMET forecast was better in some cases as
it had a bogus vortex in the model so that such an influence is better
represented or predicted?
Roger's point about the effect of vertical wind shear may also be related
to the modification of the vortex structure and hence a wrong representation
of the latter would result in an erroneous interaction with the large scale
flowm and thus a wrong forecast.
Johnny
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.112.67.72
討論串 (同標題文章)
完整討論串 (本文為第 2 之 2 篇)：