Transcript Frantic About Frances

Frantic About Frances
Richard H. Grumm and John LaCorte
National Weather Service
State College, PA 16803
Introduction
• Significant Hurricane as approached Florida
– Had a big impact on Florida
– Raised issues of impacts as it moved north
– Would the track and timing be better than with…Charlie?
• Big issue was the transition stage
– Heavy rains, tornadoes, and winds
– Frances was a large storm lots of anomalies in wind, moisture and
pressure fields
– And timing of heavy rain, tornadoes, and wind
– Timing was in doubt due to model differences in the speed of
movement
• Heavy rains on north and west side of the storm in
Pennsylvania, New York, and Ohio
– Cold conveyor rains (heavy) in western New York, PA, and
OH.
Frances…the hurricane
…not the mule…
MODUS Image as CAT-3 Storm
3 Sept 2004
4 Sept GOES-12
Track through transition
based on TPC and John LaCorte’s analyses
Florida and beyond
• Big impact of Big Storm Florida’s East Coast
– Winds and rain
– Tornadoes
• Expectations of storm as it moved northward
– Charlie had a bad “track” record (more on this later!)
– Where would it rain and would there be significant
winds etc.
– Tornadoes were a problem too
• Mainly in the feeder bands within the warm tropical air mass.
• Okay, Ivan, terrible as it was…produced more tornadoes.
Tornadoes on the 6th
54 Tornadoes on the 7th
26 Tornadoes on the 8th
Eta 00-hour 850 hPa wind anomalies
0000 UTC 8 September
•Extremely anomalous southerly winds
ahead of the system
•5-6 Standard Deviation above normal
winds
•Strong shear produced tornadoes
•The strong wind anomalies persisted,
though weakened as the storm moved
northward
•Note evolving easterly jet north of the
cyclone…focus for heavy rains
850 Winds 9 September 0000 UTC
•Still had anomalously strong low-level
winds in the warm sector east of the
surface low center
•Last day of tornadoes in MidAtlantic region
•Tornadoes in the feeder bands
and in the warm tropical air
•The easterly jet was now welldeveloped north and west of the main
low
•Cold conveyor belt area of heavy
rainfall
•Well north of the cyclone center
and removed from the tropical air
at low-levels
0000 UTC 9 September
•Low up western slopes
Appalachians
•Anomalous
precipitable water
•Even in the cool air in
western PA and OH
•Strong LLJ into MidAtlantic
•Strong northeasterly
jet
•Cold conveyor
KCCX Storm Total Estimated
Rainfall
•Heaviest rains were well to
the west
•Western PA and Ohio hard hit
•Not nearly as much rainfall in
the warmer air to the east
•The cold conveyor rain effect
dominated in this event.
KPBZ estimated Rainfall
real rainfall shown later
•More reliance on tropical Z-R
•There was some very heavy
rain in eastern Ohio
•Same overall pattern as
KCCX though this extends
farther west
GFS Forecasts For Frances
Eta Forecasts For Frances
Eta taken it slowly?
0000 UTC 09 September
Heavy rains in cold
conveyor belt region
Note temperatures in
the 60s.
L
0000 UTC 09 September
Heavy rains in cold
conveyor belt region
Note temperatures in
the 60s.
L
Tropical air
Cold conveyor rain
• Rain shield heavy rain north and west of
surface low
• Strongly baroclinic
– Northern ageostrophic winds
– Strong low- and mid-level forcing
(Frontogensis)
• Not most common concept heavy rain in
tropical systems but big in Mid-Atlantic
/northeastern US
Frontogenetic Rains
0000 UTC 9 Sept 850 FGEN/OBS/AGE 1000 hPa
ETA 0600 UTC 9 Sep
1200 UTC 9 Sept
RUC FGEN and OBS
0500 UTC RUC
0800 UTC
rain/FGEN anchored in cold airmoving north
1200 UTC 9 Sept
rain about over in PA!
L
1800 UTC 9 September
L
Eta Forecasts For Frances
Eta got faster with time but initially missed Oswego area locaiton!
The Eta and Tropical Storm Tracks
• The Eta was too slow with Frances
– The GFS did a better job and was more
consistent
– SREFS were good too….
– This lead to rainfall forecasts and flood
products talking about rainfall into the day of
the 9th in PA
• The rain ended in most locations before 1200 UTC
• Lets look at the woeful Eta forecasts for
Charlie
Eta Tracks with Charlie
all valid 12Z 15 August initialized different times
GFS Tracks of Charlie
all valid 12Z 15 August initialized different times
Hard to make good rainfall
analyses
Central Appalachian Zoomed in view
power of databases and arcview
Central Pennsylvania
New York
A word about rainfall
• Not very standardized data
• We need better means/methods to share data
• Post analysis is not so easy with such
sporadic data
• We could put together most but not all of the
picture
• We really need to improve sharing
– Snowfall and rainfall data for good case reviews
– This needs to be done in near real-time
Conclusions
• Frances wreaked some havoc
– In Florida as a hurricane
– In the southeast-Mid-Atlantic as
• Rain and tornado machine
– Central PA had mainly a cold rain
• Remained in cool low-level air when the heavy rains arrived
• Track of the storm
– The GFS was faster with storm progression and better than the
track of the Eta
– The Eta has this slow bias with moving tropical storms….what's
up with that?
– Ensembles (not shown) are a better way to go.