Transcript MM5 Simulations of an Intense Mesoscale Snowband

High-Resolution Simulations of the 25 December
2002 Banded Snowstorm using Eta, MM5, and WRF
David Novak
NOAA/ NWS Eastern Region Headquarters, Scientific Services Division, Bohemia, New York
Brian Colle
University at Stony Brook, State University of New York, Stony Brook, New York
Daniel Keyser
University at Albany, State University of New York, Albany, New York
25 December 2002 Snowstorm
Conceptual Models
Synoptic Environment
Frontal Environment
Novak et al. 2004
Moore (2003)
Motivation
• Can high-resolution models simulate the physical
processes responsible for mesoscale banding?
• What is the simulated relationship between the modeled
band and frontogenesis, weak moist symmetric stability
(MSS), moisture, and ascent?
• Does precipitation drift influence the forecast band
location?
• What are the predictability limits of mesoscale banded
features?
Case Study Design
• 25 December 2002 Banded Snowstorm:
– Compare Eta, MM5, and WRF forecasts to observations
– Models initialized at 0000 UTC 25 Dec 2002
Model
Resolution
FSL RUC
20 km
Convection
Boundary
Layer
Microphysics
Grell–Devenyi
Burk–
Thompson
Simple Ice
NCEP Eta 12 km
BMJ
Mellor–Yamada Ferrier
(MY)
12 km
NCAR
PSU MM5
12 km
NCAR
WRF v2.0
Grell
MRF
Simple Ice
Grell–Devenyi
YSU
Simple Ice
25 December 2002 Snowstorm
Low Track Comparison
25 December 2002 Snowstorm
MSLP Time Series
1000
MSLP (mb)
995
990
Obs
985
Eta
980
MM5
975
WRF
970
965
960
12Z
15Z
18Z
21Z
Time (UTC)
00Z
03Z
06Z
QPF Performance
1200 UTC 25 Dec – 1200 UTC 26 Dec QPF
Obs
Eta
Max = 3.00 in.
Max = 1.50 in.
MM5
WRF
Max = 1.85 in.
Max = 2.25 in.
RUC Analyses
•
•
•
Radar Reflectivity
700-hPa Miller 2D Frontogenesis [thin solid; ºC (100 km)–1 (3 h)–1 ]
700-hPa Geopotential Height (thick solid)
RUC Cross Section
• Frontogenesis (shaded)
• Өes (solid)
• EPVf* = 0 PVU (green)
• Relative Humidity (shaded)
• Vertical Velocity (dashed; µb s-1)
Eta Forecast
•
•
•
3-h Accumulated Precip (shaded; hundreths in.)
700-hPa Miller 2D Frontogenesis [thin solid; ºC (100 km)–1 (3 h)–1 ]
700-hPa Geopotential Height (thick solid)
Eta Cross Section
• Frontogenesis (shaded)
• Өes (solid)
• EPVf* = 0 PVU (green)
• Relative Humidity (shaded)
• Vertical Velocity (dashed; µb s-1)
MM5 Simulation
Reflectivity
Frontogenesis
Ө
Circulation
EPVf* < 0
Өes
Circulation
WRF Simulation
Reflectivity
Frontogenesis
Ө
Circulation
EPVf* < 0
Өes
Circulation
Relationship between the Band,
Frontogenesis, weak MSS, and Moisture
MM5
WRF
Precipitation Drift?
2000 UTC
2100
2200
2300
0000
0100
0200
0300
ENX
After 2200 UTC
•30 dBZ core of band tilted towards SE
•height of 30 dBZ core steadily descends
ENX
Precipitation Drift?
Using SCH winds at 2200
UTC, assuming crystal fall
speed of ~1 m s-1 and
steady-state conditions:
Crystal drifts 46.3 km @
222°
Schenectady (SCH) Wind Profiler
Findings
• Eta/MM5/WRF capable of simulating mesoscale band
formation and dissipation
• All three model forecasts showed a narrow sloping ascent
maximum associated with strong midlevel frontogenesis
and weak moist symmetric stability
• Model QPFs were underdone by as much as 50% in the
banded region, and exhibited a southeast position error.
• Slantwise ascent and precipitation drift may account for
slight displacement (20–40 km) between frontogenesis
maximum and band
• Despite different surface cyclone intensities, all three
models predicted band formation, suggesting some
predictability in this case
Future Work
• Explore why all three model forecasts exhibited a
southeast position error and dissipated the band too early
• Diagnose 3D hydrometeor trajectories across the band
• Perform higher-resolution (4 km / 1.33 km) model runs
• Use high-resolution ensembles to explore mesoscale
banding predictability in this case
Ensembles
MM5 Eta init
Grell / MRF
Simple Ice
MM5 GFS init
Grell / MRF
Simple Ice
MM5 Eta init
Grell / GS
Simple Ice
MM5 Eta init
KF / MRF
Simple Ice
MM5 Eta init
Grell / MRF
Reisner
WRF Eta Init
Grell–Devenyi / YSU
Simple Ice
WRF Eta Init
KF / YSU
Simple Ice