Using a Mesoscale Model to Investigate Local Forecast Problems

1
Using a Mesoscale Model to Investigate Local
Forecast Problems

• Kim Runk - NWS Las Vegas, NV

2
Caviat

• A mesoscale model will not necessarily provide a
  more accurate, trustworthy explicit solution.

3
Initialization Error

• Inadequate observation density
• Inadequate observation frequency
• Objective analysis errors
• Data assimilation problems
• Poorly blended boundaries

4
Intrinsic Model Error

• Approximated equations
• Grid interpolations
• Boundary conditions
• Parameterizations
• Predictability limitations
• Unrepresentative terrain

5
Local Applications

• Operational Guidance
• Case Study Simulations
• Sensitivity Experiments
• Mesoscale Ensembles

6
Four Snapshots from LAS

• Real-time Guidance Tonopah Low
• Case Study Downslope Windstorm
• Sensitivity Study Convergence Zone
• Sensitivity Study Gulf Cal Moisture

7
Frontal Evolution East of the Sierra
8
GOES10 IR 1800 UTC 17 NOV 98
9
17/18Z Nov 98 ETA Surface Analysis
10
17/18Z Nov 98 RAMS Surface Analysis
11
17/18Z Nov 98 RAMS Surface Analysis
12
Downslope Wind Event
Case Study Example
13
Eta 500 mb Hgt/Vort, MSLP VT 3/1200Z Feb 1998
14
MesoEta 700-500mb mean omega
15
Eta Time/Height Section Gabbs, NV
16
MesoEta 850mb Winds 3/1200Z
17
MesoEta Time Height Section Theta/Wind
18
RAMS 12km/4km GRIDS for GABBS Event
19
RAMS Cross Section Theta / Tang.Wind
20
RAMS 2nd sigma level wind 3/1200Z
21
Local Convergence Zone
Sensitivity Study Example
22
Reflectivity Image of LVCZ Event
Local mesonet surface winds superimposed on KESX
WSR-88D composite reflectivity, valid 2039 UTC,
30 July, 1997.
23
RAMS Configuration

• Non-hydrostatic, two-way interactive nest
• 25 vertical layers, 30-sec NCAR terrain
• 30-sec time steps, Schultz microphysics
• Mahrer-Pielke radiation, no convec. parm.
• First set of simulations initialization and
  lateral boundary conditions from RUC
• Subsequent runs horizontally homogeneous
  modified proximity sounding from DRA

24
Composite Sounding from 8 LVCZ casesCAPE 625 J
kg-1 Mean 1-4 km wind 230/06 ms-1
25
18-hour Surface Wind and Convergence
26
18-hour Surface Wind and Relative Vorticity
27
Tilt View of LVCZ Circulation
Surface wind vectors and shaded terrain in
bottom half of image. (LVCZ highlighted) Potenti
al temperature contours and shaded U-component of
wind in upper half of image. (vertical solenoid
on lee-side of Mt Chuck highlighted by arrows)
28
2100 UTC Surface Wind Convergence with Mean
1-4 km Wind 230/12 ms-1
29
2100 UTC Surface Wind Convergence with Mean
1-4 km Wind 310/07 ms-1
30
Key Factors for Classic LVCZ

• Mean wind in the 1-4 km layer AGL
• direction between 200 and 280 degrees
• speed less than 10 ms-1 (optimal 5-7 ms-1)
• Deep, well-mixed unstable boundary layer elevated
  above a shallow, surface-based inversion in the
  early morning hours
• Sufficient heating to initiate solenoidal
  mountain-valley circulation

31
Conceptual Diagram of the LVCZ life cycle.
32
Gulf of California Moisture Surges
33
Synoptic Analysis of Surge Potential
34
Strong Gulf Surge at Guaymas
35
Moderate Gulf Surge at Yuma
36
(No Transcript)
37
Gulf Surge Experiment

• Single sounding initialization
• Uniform south wind at 5 ms-1
• Inner nest 10km grid spacing
• Water vapor no latent heat/precip
• Control run SST August climo
• Sensitivity run Gulf grid pts land

38
30km/10km Grid Domains
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
15h Surface Wind and Dewpoint with Gulf
43
15h Surface Wind and Dewpoint without Gulf
44
950mb Moisture Transport at 15h with Gulf
45
950mb Moisture Transport at 15h without Gulf
46
12h Surface Temperature with Gulf
47
12h Surface Temperature without Gulf
48
12h Mean Sea Level Pressure with Gulf
49
12h Mean Sea Level Pressure without Gulf
50
15h 950mb Wind/Isotachs with Gulf
51
15h 950mb Wind/Isotachs without Gulf
52
Time Height Section of Wind at BLH with Gulf
53
Time Height Section of Wind at BLH without Gulf
54
12h 850mb Wind Relative Vorticity with Gulf
55
12h 850mb Wind Relative Vorticity without Gulf
56
Local Modeling Advantages

• Rerun selected cases at high resolution
• Conduct sensitivity studies of local interest
• Add value to operational NCEP suite

57
What About FF Forecasting?

• Problem Convective QPF is subject to too many
  sources of error to provide definitive Flash
  Flood guidance.
• Problem important factors like antecedent
  precipitation, geometry of drainage basin, amount
  of urbanization not modeled well.
• Solution Use the model to identify the key
  ingredients that favor known FF conditions.

58
Key FF Ingredients

• Rain rate is proportional to magnitude of
  vertical moist flux and precip efficiency.
• Flash flooding is typically favored in situations
  that produce recurring storms within a
  slow-moving system.
• Cell motion largely advective but boundary
  relative flow (thus, propagation) influenced by
  external factors in the mesoscale and storm scale
  environment.

59
Processes on Various Scales

• Synoptic scale general circulation, stability,
  convective suppression during accumulation of
  widespread, deep moisture
• Mesoscale lift to initiate convection,
  processes that regulate propagation, influence
  organization, modify environmental shear and
  buoyancy terrain
• Storm scale interaction of cold pools with new
  updrafts, relationship of local RH/evaporation to
  precip efficiency, cell motion vs. propagation

60
Strengths of a local mesoscale model that apply
to the flash flood problem

• Direct use of explicit grid-scale quantitative
  precipitation forecasts is applicable where
• Synoptic scale processes dominate
• Modulation by terrain is clearly delineated
• Warm rain processes are likely
• Example southern California winter rains
• Otherwise, an ingredients-based approach is more
  advisable.

61
Strengths of a local mesoscale model that apply
to the flash flood problem

• Real time, high resolution, non-hydrostatic
  solutions offer valuable operational insight.
• Historical case studies increase understanding of
  important physical mechanisms and assess a
  models ability to accurately reproduce various
  aspects of a specific event.
• Sensitivity studies assist us in isolating the
  most important processes and determine which
  parameters the model solution is sensitive to.

62
(No Transcript)
63
(No Transcript)
64
(No Transcript)
65
(No Transcript)