Wind Prediction
NWS/FAA Wind Observations
ASOS Sensors Have Been
Changed from Three-cup to
Acoustic Units
Old ASOS Wind Sensor
• The old ASOS wind sensor, the Belfort 2000, used rotating cups to measure wind speed and a vane to measure wind direction.
• Over a two-minute period ASOS used 24 five-second averages to determine the two-minute average wind speed and direction.
• The highest 5-second wind speed during the previous ten minutes was the gust. Gusts were only reported if there is a variation of 10 knots between peaks and lulls.
• The highest instantaneous wind speed (gust) since the last routine report wss the peak wind.
Problems With Old Wind Sensors
New Acoustic Wind Sensor
The New Wind Sensor
• The new ASOS wind sensor, the Vaisala 425NWS, is a sonic anemometer . It has no moving parts and is designed to operate better in winter weather conditions.
• As with the Belfort sensor, over a two-minute period, ASOS uses 24 five-second averages to determine the two-minute average wind speed and direction. But the highest threesecond running average speed is stored for gust and peak wind processing.
• Most sites have switched to the new sensor.
• The new sensor is responsive to short-term gusts. Can expect to see more gusts and peak winds reported with the new sensor.
Wind Observations
• Major issue is representativeness (e.g., in valley or behind a tree)
• Surface winds are highly variable due to varying surface characteristics and obstacles.
• Wind varies substantially with height and not all sensors are at similar elevations about the ground.
Wind Gusts
Wind Gusts Associated with
Mixing of Higher Momentum
From Aloft
Gust Ratio Depends on Wind
Speed
Do Model Winds Include Gusts?
NO!
Wind Speed at the University of Washingon
(Sampled every 5 seconds: reports 1 minute averages and highest 5 second wind gusts each minute )
MM5
Output
Every Time
Step from the 4-km
Domain:
Much
Smoother!
Wind Speed
MM5 (or WRF) 4-km output every time step appears to have the temporal variability of approximately 15 minute-average winds
Why?
Winds are averaged spatially due to model resolution, grid-box averaging of some terms, and model numeric and explicit diffusion/smoothing.
Thus, for verification we should compare model output to temporally averaged observations .
Gust Guidance for the Forecaster
• Gust ratio depends on wind speed and vertical stability
• In a neutrally stable environment, forecasters often look for the max wind in the PBL as a measure of the max gust
• There is software that does this as well.
How do we do this for WRF?
Danger of Using Model Output Directly
• Lack of resolution..means larger scale models
(e.g., GFS) can’t accurately define and predict local winds forced by mesoscale features…terrain, diurnal circulations. This is getting better.
• Physics problems and particularly PBL parameterization issues. WRF and most other mesoscale models tend to overmix winds in the vertical…particularly under stable conditions-results in excessive winds. Winds generally too geostrophic
• Large scale model errors…from poor initializations and other causes.
Use of Models
• Today’s synoptic and mesoscale models are sufficiently realistic and accurate that is very hard to beat their wind forecasts aloft.
The forecaster’s role is mainly in deciding which model to use and perhaps altering the timing, if phase errors are obvious.
• At the surface, the models are getting better, but there are larger biases and other errors.
Using SL Pressure For Surface
Winds
• Traditionally, an important tool of the wind forecaster was to start with the SLP pressure fields and deduce the surface (10-m) winds from it.
• Lets review these relationships.
Impact of Surface Drag
Variations
• Winds are often 50% to 300% stronger over water than land
Stability and Vertical Momentum
Mixing
Wind Speed Up After WF Passage
The Chanukah Eve Windstorm
Stability Changes and Winds
Momentum mixing: 12/17/06
Orographic barriers can greatly change the pressure and wind fields
• Examples include mesoscale windward ridging and lee troughing.
• Can greatly enhance or weaken the winds.
Mesoscale Pressure and Wind
Perturbations on Mesoscale
Terrain Barriers
• A controlling parameter is the Froude number:
FR = U hN where U is the speed, h is the height of the barrier, and N is stability (Brunt-Vaisalla freq)
• Large FR is associated with flow going up and over terrain (large vertical excursions),
• Small FR with flow being deflected around (quasihorizontal flow)
Strait of Juan de Fuca During the
Coast Field Experiment
Mesoscale Pressure Perturbations: Large
Influence on Winds
Sea Level Pressure
February
13
1979:
The
Hood
Canal
Storm
Very
High
Froude #
Winds over 100 kts destroyed the Hood Canal Bridge
Cost to replace: over 100 million dollars
Stability Influences Topographic
Blocking and Acceleration
Gap Flow
• In gaps through mountainous regions the flow is generally NOT geostrophc, but rather highly ageostrophic and downgradient, moving from high to low pressure.
• Historically, forecasters have developed simple relationships between the wind speeds and pressure differences across the gaps in question.
– Example: SEA office uses UIL-BLI gradient
(westerly winds at exit= 10*delta p)
Gap Flow 101 - Misleading the
Next Generation!
• The Venturi Effect is still used in some introductory texts to explain gap flow!
It turns out that the strongest winds are generally not in the narrowest parts of mesoscale gaps, but in their exit regions
Strait of Juan de Fuca is well known for its easterly gales in the gap exit region.
Troutdale
Columbia Gorge
Most Simple Approach
• 1-D horizontal momentum Equation: du dt
1
p
x
fv
ku
2
u
t
x
u
2
2
1
p
x
fv
ku
2
• Assume steady state, neglect Coriolis and friction and integrate:
• This is simply a form of Bernoulli’s equation. Assuming steady state and no friction:
x
u
2
2
1
p
x
u
1
2 u
0
2
2
P
Gap Flow 101 - Basics
• Provides an upper limit to maximum speed at the end of the gap
– Commonly used in work from the early 1980’s
• E.g. Walter and Overland (1981), Reed (1981)
– Oversimplification.
– Produces winds that are too strong.
• Gap winds are a boundary layer phenomena
– Must account for drag (both surface drag and drag at the inversion)
Gap Flow 101 – Adding Drag
• Reintroduce friction (bulk aerodynamic form)
u
t
x
u
2
2
1
p
x
ku
2 u
1
2
u
2
0
k p
e
2 kx
p k
where k
2 .
8 C
D
H
• Shown to produce a much closer correlation to observed winds
– E.g. Lackmann and Overland (1989), Mass et al (1995), Colle and Mass (1986), Bond and Stabeno (1998)
Hydraulic Effects
• There is another important features of many gap flow situations: hydraulic effects associated with changes in depth of the cold dense air (analogous to water)
Hydraulic Effects Tend to Slow the
Wind at the Entrance and Speed Up at the Exits
Examples of Gap Flow
The Columbia River Gorge
Near Sea Level Gap
On Border of WA and OR
Troutdale
444.4 m grid spacing, Pass Height = 100 m
Portland Troutdale Cascade Locks
Vertical Structure
• Strongest winds near exit
• Hydraulic effects are important
Fraser River NE Gap Flow
Max Winds, 28 Dec. 1990
> 40 ms -1
Summary
• Strongest winds tend to be in exit region because of hydraulic collapse and because of larger scale pressure gradient.
• There can be some venturi acceleration in narrow regions…but that tends to be secondary.
• High-resolution numerical models can do a very good job with fine-enough grid spacing.
Diurnal Winds
• Sea breeze/land breeze
• Upslope and downslope winds
• And combinations of the above.
• Generally larger in the warm season.
Sea Breeze Winds
Average Summer Wind Speeds
18
16
14
12
10
8
6
4
2
Hoquiam
Newport
North Bend
Gold Beach
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour (LDT)
2-minute average July-August winds along the Northwest coast.
Summer
Diurnal
Winds
Picture take in Sequim looking N
Strong Regional Sea Breeze
Winds During the Summer
Land Breezes are Strongest in
Locations with Warm Water
• Weak in cold-water areas such as the northwest.
• One exception: during arctic air outbreaks.
Land Breeze example during a cold period.
Interactions with larger scale flow
• Strong onshore flow results in weak or nonexistent sea breeze
• Weak to moderate offshore flow: strong sea breeze and sea breeze front
• Strong offshore flow: weak sea breeze
• Additive to flow parallel to the coast (e.g., southern Oregon)
Sea Breeze Winds Along the
Southern Oregon Coast: Interaction with synoptic scale flow
• Gusts frequently reach 30-35 knots during the summer during the afternoon.
• Very painful to stay on the beach!
• Strong pressure gradient normal to the coast between the coastal thermal trough and cold upwelling water has a substantial impact.
South
North
Southern Oregon Coast Near Brookings
Fovell (UCLA) Sea Breeze
Simulations
• http://www.atmos.ucla.edu/~fovell/ASother
/mm5/LA_seabreeze.html
Slope Winds
Fig. 7-11, p. 178
Downslope Windstorms: strong winds on the lee side of mountains, generally associated with high amplitude mountain waves
Enumclaw, WA
Mountain Wave 101
Trapped Lee Waves
Lee waves whose energy does not propagate vertically because of strong wind shear or low stability above are said to be "trapped.". These waves are typically at an altitude within a few thousand feet of the mountain ridge crest and turbulence is generally restricted to altitudes below 25,000 feet, particularly in rotors. No tilt and weaken with height aloft.
Vertically Propagating Waves
Vertically-propagating waves occur when waves become more amplified and tilt upwind with height. Tilting, amplified waves can cause aircraft to experience turbulence at very high altitudes.
Clear air turbulence often occurs in the upper troposphere due to vertically-propagating waves.
Such waves have been documented up to 200,000 feet and higher.
Downslope Windstorms
Under the proper circumstances (e.g., a critical level aloft) the wave can amplify and break, resulting in a downslope windstorm
Froude Number and Mountain Waves
The Froude number expresses a ratio between the kinetic energy (wind speed) and the potential energy (stability times mountain height).
• If the Froude number is equal to or slightly greater than 1, then there is the likelihood of mountain wave activity
• If the Froude number is less than one, then the airflow is insufficient to carry the flow over the mountain and the flow is blocked. Lower probability of mountain waves.
• If Froude number is much more than 1, airflow proceeds right over the mountain and down the other side, with no significant oscillations, but there can one high amplitude wave
Trapped vs Vertically
Propagating
• A key parameter controlling the nature of mountain waves is the Scorer Parameter (l) l 2 = N 2 - d 2 U
U 2 U d z 2
•k is the primary wavenumber of the terrain =2*pi/L
, where L is the length scale of the terrain
•k < l : vertically propagating, k > l trapped
Critical Levels
• A critical level occurs when the flow normal to the mountain barrier reverses.
Critical Levels
• Critical levels may be self-induced by wave breaking or result from the overall environmental flow.
• Critical levels do not allow the vertically-propagating energy associated with mountain waves to continue upwards. Instead, that energy is deflected by the critical layer back towards the surface.
• Consequently, critical levels can contribute to the development of, and/or the strengthening of, downslope windstorms.
Stable Layer
• A stable layer near crest level with less stable air above can act like a critical level.
• Happens relatively frequently.
What to look for for in a windstorm forecast
• Strong winds approaching the barrier (and Froude number greater than one so air goes over the mountains). Winds should be within 45 degrees of normal to mountain crest.
• Stable layer near crest level. Lesser stability aloft.
• Critical level above the mountain barrier (to promote wave breaking).
• The existence of weak vertical wind shear or reverse shear
(winds decreasing with height) are more favorable than forward shear (winds increasing with height).
• Strong downslope windstorms are often associated with large cross-barrier pressure gradients, but it is not clear whether those are cause or effect.
Numerical Models and
Downslope Windstorms
• Until recently, criteria such as the above, used subjectively by forecasters, were the only approach.
• During the past decade it is clear that high resolution numerical models (2-10 km grid spacing) are highly effective tools.
• Demonstrated repeatedly here in the NW.
Enumclaw, Washington
“Place of evil spirits”
Enumclaw Windstorm Pressure
Pattern
December 24, 1983
December 24, 1983
High-Resolution MM5 Simulations Do An
Extremely Good Job of Predicting/Diagnosing
Such Gap/Downslope Windstorm Hybrids
High-Resolution MM5 Simulations Do An
Extremely Good Job of Predicting/Diagnosing
Such Gap/Downslope Windstorm Hybrids
December 18, 2010
Extreme Mesoscale Winds
Associated with Strong Cyclones
Over the Ocean
Cyclone Windstorm Issues
• Strongest winds often when low moves north of location.
• Oceanic cyclones often have strongest winds in bent back trough—the “poisionous tale”
• The interaction with coastal mountains depends on stability.
Why strongest winds in Seattle when storm goes to the north of us?
• Maritime Cyclone Winds 101 are needed!
Shapiro-Keyser Model of
Oceanic Cyclones
Strongest Winds With Back-Bent
Warm Front
Warm
Seclusion
Stage
Interactions with our local mountains can amplify or weaken the effects of our great storms
Strong Winds and Terrain
• Ageostrophic acceleration near terrain
• Isallobaric wind contribution.
Why Strongest Winds Near the Central Coast?
Mesoscale Pressure Perturbations: Large
Influence on Winds
Sea Level Pressure
The Most Extreme of the Extreme: The
Columbus Day Windstorm of 12 October 1962
Max Winds
(mph)
Columbus Day
Storm 1962
Columbus Day 1962: At Cape Blanco there were
150 mph with gusts to 179! Strongest winds on bluffs and windward slopes of coastal orography