The following list covers the major forecast problems facing Utah forecasters during the cool-season.

     I.    Snow Amount Forecasting: The Basics
     II.   Snow Forecasting for the Urban Wasatch Front
     III.  Heavy Snow Forecasting for the Mountains
     IV.   Pre-frontal High Winds
     V.    Post-frontal High Winds
     VI.   Canyon Winds, and Other Downslope Windstorms
     VII.  Heavy Rain and Rain-on-Snow Flooding
     VIII. Dense Fog
     IX.   Freezing Rain and Drizzle 

I. SNOW AMOUNT FORECASTING

This is, by far, the forecast problem that gets the most attention
in Utah.  It is one of the most difficult tasks faced by
forecasters.  Most of the knowledge and skill associated with snow
forecasting in Utah pertains to the urban corridor along the
Wasatch Front, and the adjacent Wasatch Mountains.   

A) Moisture: Upward vertical motion (UVV) is the key ingredient required to produce accumulating snow in Utah during the winter. Because of the long time-scales associated with cool-season vertical motion compared to vertical motion associated with warm season convection, the amount of moisture in the air is not the limiting factor in snowfall forecasting. There are two exceptions to the previous statement. 1) A deep arctic airmass with precipitable water values of less than 0.10" will generally not produce much snow even if lifted. Of course these airmasses are usually quite stable and are not easily lifted. 2) Unusually warm-moist airmasses with precipitable water values of more than 0.75" may produce unusually large amounts of precipitation if lifted, and given the warm-moist nature of these airmasses, they are often quite unstable and prone to summer-like convection. In most situations over Utah, there is generally enough moisture in the airmass to produce significant snowfall if there is enough UVV.

B) Stability: Vertical motion in Utah can be divided into two categories conceptually. These are dynamic UVV and orographic UVV. In both cases it is useful to think of "forcing" and the "response" of the atmosphere. To a great extent, the "response" and hence the strength of the vertical motion is determined by the stability of the airmass. The less stability, the greater the response to forcing. Thus, a little forcing will produce a lot of UVV if the atmosphere is nearly moist adiabatic, while very strong forcing will not produce much, if any, lift if the atmosphere is stable. This is true for both dynamic and orographic forcing. An obvious example of the importance of stability is when there is strong northwest flow into the Wasatch, but no precipitation forms. It generally isn't the lack of moisture that inhibits precipitation, but rather the stability of the airmass prevents the "response" to the orographic lift. These situations result in lenticular clouds rather than precipitating stratus or cumulus.

     Tools for diagnosing stability:

     1. lapse rate progs in NTRANS: This represents the temperature
     difference between 850-500 mb divided by the thickness, and is
     expressed in units of degrees/km.  A lapse rate of 6 or
     greater is relatively unstable during the cool-season, with 8
     being a number to get excited about.  Beware of large diurnal
     changes in lapse rate in fall and spring, particularly in the
     AVN.  The lapse rate progs are part of the div-Q and
     differential vorticity advection graphics.

     2. theta-e time/height sections in NTRANS: This represents the
     "convective" instability.  If theta-e is decreasing with
     height, then the atmosphere is convectively unstable, meaning
     if the layer were lifted, the resulting lapse rate would be
     greater than moist adiabatic.  We see this all the time in the
     summer, but it is much less common in the cool-season.  More
     often in unstable situations we see vertical lines of theta-e,
     which show theta-e is constant with height. If theta-e is
     constant with height or decreases with height through a
     significant depth of the atmosphere, a significant UVV will
     result from any forcing.

     3. model soundings in NTRANS: This is fairly self-explanatory
     with regards to stability, with steeper lapse rates
     representing unstable conditions.  The additional information
     provided by the soundings over other methods is that if they
     are accurate, they can show where capping inversions exist,
     and may therefore provide a better indication of where cloud
     tops will be.  In most situations, deeper clouds will produce
     more snow.

     4. lifted index in NTRANS and in FRH(T)73: The LI is a gross
     indication of stability in that it represents the difference
     in degrees between a parcel lifted from the LCL to 500 mb and
     the environmental temperature at 500 mb.  If the lifted parcel
     is warmer than the environment, the lifted index is negative
     and indicates unstable conditions.  The FRH data provides the
     LI at SLC and CDC, while the NTRANS graphics in the
     "convective" meta-files provide a spacial depiction of LI.
     The parcel to be lifted is selected in a variety of ways and
     this is the reason the LI's may be different for the FRH data
     when compared to the NTRANS graphic.

In many ways the more important forcing for vertical motion is the dynamic portion. If dynamic subsidence is present, a subsidence inversion will quickly form and the resulting stability will kill off orographic lift or convection. Conversely, UVV will produce adiabatic cooling, which, will result in steeper lapse rates. Thus, dynamic lift can strongly influence the stability and indirectly enhance or negate the vertical motion "response" to forcing.

Diagnosing dynamic UVV is one of the BIGGEST topics in meteorology. The numerical models are very good at it, but the omega displayed in the models includes the orographic portion that the models are very bad at. Additionally, prior to access to gridded model output, the only vertical motion data in AFOS was a plan view graphic for 700 mb. Thus, forecasters have traditionally used QG theory to try to subjectively evaluate dynamic vertical motion and then add in their own "knowledge" of other processes such as orographic lift, mesoscale convergence zones, lake effect, and any non-hydrostatic processes such as convection. QG theory is based on some simplifying assumptions and therefore does not include all the terms of the Primitive Equations that are solved by the numerical models. An example of what's missing in QG theory are the advections by the ageostrophic wind. QG theory also is based on very large scales and thus is not appropriate for features with short wavelengths or sharp curvature.

The basic technique for diagnosing QG vertical motion is to apply the QG omega equation. Areas with positive vorticity advection increasing with height and areas with a positive Laplacian of temperature advection (usually approximated as areas with warm advection) are associated with UVV and vice versa. The two terms often cancel out (cold advection in areas with PVA increasing with height), making application of this method difficult. The stability term in the denominator is usually ignored, although as previously noted, stability is extremely important. Finally, there is a Laplacian operator on the left-hand-side of the QG omega equation that is also usually ignored. Another technique for the QG forcing is to use Q-vectors. Without going into the gory details, forcing for upward motion is associated with convergence of the Q-vectors and vice versa. The problems with stability and the left-side operator are still present, however, the cancellation problem is resolved. A complication of the Q-vector approach is the issue of which levels or layers to use, and the problem of what to do when the divergence of the Q-vector changes sign in the vertical. The Trenberth method of advecting the vorticity of a layer by the thermal wind of that layer is nearly (but not exactly) identical to the Q-vector approach.

As the horizontal grid spacing of the numerical output has dropped down to 80 km or below, the display of QG fields such as divergence of the Q-vector, or differential vorticity advection has become very noisy and somewhat difficult to interpret. Many times the QG fields are so noisy in the vicinity of the jet that an interpretation of either upward or downward motion could be argued. One possibility is to look at highly derived QG fields on AWIPS using the National area grids (190 km). These are smoother than the 80 km grids. However, its worth noting that a side by side comparison of the National and Regional (95km) grids on AWIPS will at times show a reversal of sign for particular regions. Thus, the use of QG diagnostics to infer vertical motion is increasingly becoming difficult to apply as model resolution improves.

Of course the models don't use any of this to come up with omega. Since they are hydrostatic models, there is no prognostic equation for omega. But instead of using a diagnostic equation for omega (like the QG omega equation), the models integrate the continuity equation through the depth of the model to come up with omega. Mesoscale dynamic processes such as secondary circulations around fronts are not captured by the models. In fact, fronts are not captured by the models (there is a lot of debate these days about the nature and even actual existence of fronts).

If the model's precipitation forecasts were good, we wouldn't need to do any of this "forcing/response" stuff. Verification of the FRH73 data for the ETA (the model with the best precipitation verification scores) shows the model predicts light precipitation much more than it is observed, and consistently underforecasts the heavy events. Embedded convection during the cool-season is virtually non-existent over Utah in the ETA. In general, the model QPF's exhibit the lowest skill in the mountainous regions of the West.

     Tools for diagnosing vertical motion:

     1. omega time/height sections in NTRANS: These graphics
     provide the total omega predicted by the models through the
     entire depth of the troposphere at specified points throughout
     Utah.  This represents a solution of the Primitive Equations
     and does not suffer from the limitations of QG theory.  It
     also implicitly incorporates the models prediction of
     stability.  The model's depiction of orographic forcing is
     also included.

     2. differential vorticity advection in NTRANS: These graphics
     are quite noisy and difficult to interpret.  Vorticity
     advection between 700-300 mb is displayed.  This graphic
     approximates one term of the QG omega equation.  Lapse rates
     from 850-500 mb are also shown on this graphic.

     3. vorticity advection time/height in NTRANS: These graphics
     show the vertical distribution of vorticity advection at
     specified points around Utah.  A visual estimate of the
     differential vorticity advection term of the QG omega equation
     can be made.  Comparison of this graphic with the total omega
     time/height graphic can often help a forecaster understand
     which process is responsible for the model's omega prediction.

     4. temperature advection time/height in NTRANS: These graphics
     approximate the temperature advection term of the QG omega
     equation and can be used in conjunction with the model's omega
     forecast to infer which processes are dominant. This graphic
     can also show how differential temperature advection will
     change the stability.

     5. divergence of the Q-vectors in NTRANS: These graphics are
     also quite noisy and difficult to interpret.  Only areas of
     convergence of Q (forcing for UVV) are shaded to help improve
     clarity of the graphic.  The calculation is for the 700-300 mb
     layer. This graphic approximates both terms of the QG omega
     equation, and lapse rates are also shown on this graphic. Its
     worth noting that the effect of cold advection below 700 mb
     will not be captured in this graphic.

     6. winds, isotachs, ageostrophic winds at 250 mb in NTRANS:
     These graphics (in the West and Main meta-files) show the
     location of the jet at 250 mb.  The along-stream change in
     wind speed in jet entrance and exit regions and the associated
     ageostrophic winds in these regions can be qualitatively
     associated with the secondary circulations that extend through
     the troposphere and result in vertical motions.  This approach
     is similar to QG in some ways, but is not limited by the QG
     assumptions.  However, its application is very  subjective,
     and there are many complications with associating jet-level
     winds with the location and strength of lower-level vertical
     motions.

     7. graphics that show vorticity and temperature advection:
     These graphics represent these processes at an individual
     level and are found on AFOS, NTRANS and AWIPS.  They have a
     long history of use in the era prior to the availability of
     gridded model output, and each represents an attempt to
     approximate one of the terms of the QG omega equation.  Their
     use in this way involves the application of many dangerous
     assumptions...so be careful if you use these graphics for
     vertical motion.

The Big/Super snowstorm studies for SLC indicate the most important factor for heavy snow at lower elevations of the Wasatch Front is the presence of strong dynamical forcing. A well-defined trough is over the western U.S. in virtually all cases. The primary difference between the Big (6-10") and Super (>10") storms is duration. The Super storms last longer. The Big/Super snowstorm study has recently been redone, and a more detailed classification of the patterns has been found. These results will be passed on to the staff this fall. One new result of note is: A stratification of the Super storms shows that winter super storms consist of a broad long-wave trough with embedded short-waves moving through Utah, as was found in the original study. However, spring Super storms consist of a closed upper-level low over the northern Rockies. A closed 700 mb low is over Wyoming, providing mid-level cyclonic northwest flow, while the 500 mb trough axis or closed low is still west of SLC providing strong dynamic forcing. We are working on getting all the results of the Big/Super storm study on-line for viewing via Netscape.

Characteristics: Some characteristics of snowstorms over northwest Utah's valleys include the following: Model omega forecasts in the time/height sections typically show UVV through the entire troposphere (sometimes referred to as "dashed-line dynamics"). Theta-e time/height sections usually show instability through a deep layer of the atmosphere behind the wind-shift/front. Most snowstorms at lower elevations of northwest Utah are associated with a well-defined frontal passage at sometime during the event. Obviously there are exceptions to this last statement, but systems without a well-defined thermal contrast near the surface typically don't produce heavy snow in northwest Utah. There is usually a sea-level pressure low over the Intermountain west in these cases, although the exact location of the low varies from case to case. One classic pattern is when surface cyclogenesis takes place in Nevada. If the sea-level pressure low then tracks directly through SLC or just north of SLC it is very favorable for heavy snow. This track of the low results in low-level northwest flow behind the front with upper-level southwest flow. In this situation, strong dynamical UVV combines with low-level orographic forcing. If a closed low from the surface to 700 mb tracks south of SLC, downslope flow from the Wasatch will be in opposition to the dynamical forcing and heavy snow is much less likely.

It is worth noting that conditions at upstream sites can be very misleading when attempting to forecast snow for the Wasatch Front. Many upstream locations are strongly influenced by their own topography. For example, northwest flow at Boise represents downslope and this will often result in little or no precipitation at Boise in a situation that is quite favorable for the Wasatch Front. Conversely, Boise and parts of northern Nevada tend to receive considerable precipitation in situations that resemble classic warm fronts, while these events are quite rare in northern Utah. Don't be mislead by events at upstream locations.

Snow amounts: The Big/Super storm studies point out perhaps the classic "truth" about snowfall at SLC. The main difference between the average storm and the storms that produce more than 10" of snow is the duration of snowfall. Snowfall intensity, or the "amount of moisture" do not play as large a role as duration. Intense, but brief periods of snow are common in many different types of events, but duration is almost always more important than intensity for prediction of total snowfall. Fast moving, low amplitude systems followed by strong subsidence tend to produce less than 4 inches of snow. Well developed upper-level troughs with surface cyclogenesis and strong dynamics produce about 4-9 inches of snow. This is typical of the "Big" storms where snowfall lasts for 8-12 hours followed by strong subsidence. A more prolonged period of snowfall will usually result in a "Super" storm of greater than 10 inches. Examples that prolong the period of snowfall into the 1-2 day range include: 1) a series of disturbances following the initial snow event that result in continued snowfall or just small breaks, 2) particularly slow moving systems such as in the case of a cutoff low or a stalled front, or 3) significant post-frontal lake-enhanced banding.

Rain versus Snow: A 700 mb temperature of -7C in mid-winter is usually cold enough for snow to reach the valley floor (4200') with snow on the benches (5000') at -6C. Often in the spring and fall, the atmosphere will be unstable and the lapse rate near the surface will be greater than moist adiabatic. In these situations a 700 mb temperature of -9 or -10 is needed to ensure snow will reach the valley floor. But, these situations often also produce convection, which can temporarily lower the snow level. Trapped cold air can result in snow with warmer temperatures aloft, but in most significant snow events in northern Utah, the atmosphere is well mixed, so it is unusual for this situation to occur. A variety of thicknesses can be used for the rain/snow problem, but may not be quite as reliable since the same thickness can be produced from a variety of temperature stratifications. A study by Glahn and Bocchieri showed the 50% rain versus snow 1000-500 mb thickness for SLC was 5430 meters and 850 mb temperatures of +4. The effects of evaporation and melting in the sub-cloud layer can result in snow with even warmer temperatures, and there are some rare conditions where snow will not form even when the entire sounding is below freezing (see the freezing rain discussion). If the atmosphere is moist adiabatic from the surface to above 700 mb either thickness or 700 mb temperature will work well for the rain/snow issue. There is a chart in the operations area that shows the elevation of the rain-snow line for various 1000-500 mb thicknesses.

Southwest flow: Low-level northwest flow is a favorable condition for snow in the urban corridor, but it is not required. Pre-frontal southwesterly flow can and does produce significant snow if the dynamical forcing is strong. Much of the pre-frontal precipitation may be rain if it isn't cold enough. Another situation with heavy snow in southwest flow is in overrunning. However, warm-advection precipitation events in northern Utah are rarely associated with a surface warm front in the classic sense. In fact, it is rare to even be able to analyze a surface warm front in the Great Basin. One set-up for this overrunning that can occur anywhere in the state is when a cold front drops south into the state and then becomes stationary. A wave develops along the front in response to another upper-level short-wave trough. This happened in January 1996 and resulted in over 30 hours of snow at SLC with no northwest flow. Don't assume that heavy snow won't occur just because the flow is not from the northwest.

Another type of precipitation event that occurs a few times every winter in northwest Utah is when moist northwest flow aloft occurs in a pattern of warm advection. The low-level flow is typically west or southwest, while at 700 mb and above the flow is from the northwest. The sounding in these cases will exhibit significant veering of the wind with height, which indicates warm advection. Inevitably, Utah is downstream from an upper-level ridge axis in these situations. Because of the warm-advection, the atmosphere is stable in these events. The dynamics are supplied by the warm advection and there will typically be little or no differential positive vorticity advection. The absence of PVA is perhaps the main reason that these forecasts are often missed. Precipitation is very likely and the duration can be relatively long (6-12 hours) in this pattern, but valley amounts are generally light, with snowfall amounts of 2" or less the most common occurrence. Amounts in the mountains are frequently 4-10" in these events. Again there may be no low-level northwest flow during these events.

Warm seclusions: Yet another somewhat oddball pattern for snow in northwest Utah is the so-called "warm-seclusion" pattern. In these cases a closed low with a well defined cold pocket at 700 mb moves eastward through southern Utah. A tongue of warmer air wraps around the east and north side of the low through southern Idaho. The term "warm-seclusion" has been used because the isotherm pattern in these cases resembles (weakly) the isolated warm pocket documented in some oceanic cyclogenesis cases that have been called warm seclusions. Northwest Utah will be near the comma-head in the satellite imagery with southerly or southeasterly flow early in the event. Valley precipitation is generally light at this time with a major snowstorm in progress over southern Utah. When the 700 mb trough axis passes and the flow switches to northwest, heavy snow begins in northwest Utah. However, because of the warmer air to the north-northwest of the state, this northwest flow represents warm-advection. In these events, the northwest flow occurs in a stable airmass, and orographic lift does not occur. Heavy snow (6-10") can fall in the valleys, and will generally equal or exceed the amounts in the mountains.

Mesoscale banding: Mesoscale banding is common in many snowstorms over northwest Utah. The most famous of these are the Dreaded Lake Effect (DLE) bands. Recent research suggests the surrounding mountains may be interacting with the synoptic-scale flow to produce low-level convergence zones that help to organize the lake-effect bands. This concept is not far enough along to use operationally, but suffice to say there is much we don't understand about the post-frontal bands that frequently produce significant snowfall.

Lake effect: The most favorable conditions for lake-effect snow are when cold air passes over the lake in unstable post-frontal conditions. Stability is very important. Without instability, very cold air can pass over the lake without producing any snow. The depth of the unstable layer is also a key. If a capping inversion exists at 700 mb or below, the cloud depth will be insufficient to produce heavy snow. In the heaviest lake-effect snowstorms there is typically no capping inversion below 500 mb and at times the unstable layer will extend to 400 mb. Model soundings are the best place to look for instability and the height of the capping inversion when considering lake-effect snow. Research in the Great Lakes indicates vertical wind shear may dictate whether there are multiple bands or just a single band of snow. This is unplowed ground for the Great Salt Lake. Most lake-effect events effect Salt Lake and Tooele counties, but Davis and Weber can also be hit in westerly flow. To date, the location of the bands is based on seat-of-the-pants estimates of the mean wind between the surface and cloud top. We will be trying to use a program from WSFO Buffalo this year to make this a more objective process.

CSI: Another type of banding that can produce heavy snow in Utah is from Conditional Symmetric Instability (CSI - also called slantwise convection). We had at least two CSI snow events in northwest Utah during the 1995-96 winter. The conditions for CSI include:

     1. strong vertical speed shear
     2. stability with respect to upright convection
     3. saturation
     4. inertial neutrality/instability 

III. Heavy snow forecasting for the mountains

The basic theory behind orographic snow forecasting is simple enough. Strong, unstable, moist flow oriented nearly perpendicular to the axis of the mountains will produce very large snowfalls in the mountains. In areas where we get adequate observations, the reality is a bit more complicated. Unfortunately the only mountainous area of Utah where we get enough observational data to go beyond the basics is the northern Wasatch Mountains where there are a number of ski resorts. In the rest of the state, daily snotel observations and other observations are insufficient to go much beyond the basics.

In the northern Wasatch, the orientation of the terrain and as-yet unknown mesoscale/microscale processes combine to produce very heavy snowfall in Little Cottonwood Canyon when the 700 mb flow is between 290-330 degrees. Two or three times as much snow can fall at Alta as nearby Brighton in these situations. Brighton, Park City and the Wasatch mountains near Ogden also show a peak in snowfall in northwest flow, but there is another snowfall peak in southwest flow (220 degrees at 700 mb). Moist, unstable pre-frontal flow often produces more snow at all the resorts except in Little Cottonwood, while they are the clear champion in post-frontal northwest flow.

The mountains generally receive the most snow in events with both strong dynamics and strong orographics. But, unlike the valley locations, a well-defined surface front is not as important for mountain snowfall. One study that compared Alta to SLC showed that Alta averaged 5 times as much snow as SLC when there was a well-defined cold front, but this increased to 8 times as much snow for systems without a well-defined front.

In addition to the wind speed and direction, the key element for post-frontal orographic snow is the degree and depth of the atmospheric instability. The best tools in the office to diagnose this quantity are the time/height sections of theta-e. If theta-e is decreasing with height or constant with height through at least the height of the terrain, orographic snow is likely. The deeper the layer of instability, the greater potential. However, there are exceptions to this last statement. On January 4-6, 1994, Alta received 55.5 inches of snow in 24 hours and 69 inches in 48 hours (>4 inches of water) in post-frontal orographic conditions with very low topped clouds. The clouds tops were so low they barely showed up at all in infrared imagery. No snow was reported in the valley during most of this event.

Forecasting the end of snow in the mountains can be quite difficult. Instability is the key. The time/height sections of theta-e are probably the best tool available for this task. When the theta-e begins to increase with height, or the depth of the convectively unstable/neutral layer becomes very shallow, orographic snowfall will greatly diminish or cease.

IV. PRE-FRONTAL HIGH WINDS

Strong southerly winds are common over much of Utah during the cool-season. High wind warnings and wind advisories are often required to cover these events. The conditions for these events are nearly always associated with pre-frontal situations with low pressure to the west or northwest of Utah. The strong winds are most widespread in the western valleys of the state, but strong winds in the east may just not be detected due to the lack of observations. The new automated sites in eastern Utah may change our view of the high wind climatology.

There are three physical mechanisms involved in the pre-frontal high wind events. The most obvious is the establishment of a surface pressure gradient across the north-south oriented valleys of western Utah. Lower pressure at the north end of the valleys leads to southerly winds. The terrain produces channelling and friction such that the winds are unable to become geostrophically balanced, and thus, continue to accelerate towards lower pressure. This situation sets up whenever lower pressure is to the northwest of the state. If lower pressure is to the southwest, with isobars oriented from southeast to northwest, higher pressure will be at the north ends of the valleys and strong winds are unlikely.

A second physical mechanism is the mixing of strong winds aloft down to the surface. It is not uncommon at all for a well-mixed, dry adiabatic layer to form in the pre-frontal environment of western Utah. With these steep lapse rates, strong winds aloft can reach the surface even in the absence of impressive pressure gradients. There are times when trapped cold air near the surface will delay the onset of strong winds. Strong winds may form at mid-elevations and gradually erode the cold dome until the winds reach the surface.

A third mechanism, that is yet undocumented in Utah is the creation of a barrier jet on the west side of the Wasatch Plateau. The models frequently show a low-level jet at 850 and 700 mb just west of the high terrain that bisects the state. Whenever statically stable air is pushed up into a terrain barrier, a deflection to the left occurs. A jet will form at a level below the top of the terrain. Barrier jets in southwest flow have been documented along the west slopes of the Sierra Nevada foothills by aircraft during scientific field experiments. A similar configuration is present along the west slopes of the Wasatch Plateau, and the models seem to indicate the formation of this jet-like feature as troughs approach Utah. These winds could also eventually mix down to the surface if lapse rates steepen.

     Tools for pre-frontal high winds:

     1. 700 and 850 mb winds in NTRANS: These winds are found in a
     number of different graphics, but the best ones to look at are
     the wind/isotach graphics in the convect meta files since
     every grid point is plotted, and isotachs are shaded.  This is
     the graphic that frequently shows a jet forming in western
     Utah.

     2. 850 mb heights:  This field most closely approximates the
     surface pressure gradient.  Look carefully at the orientation
     of the isohytes to ensure that lower pressure will be found at
     the north end of the valleys.

     3. sea-level pressure:  This field has the longest tradition
     of use, and is generally accurate.  However, due to reduction
     to sea-level it can be misleading.  Remember the reduction to
     sea-level uses temperatures from 12 hours earlier, so
     locations that were unusually warm will have lower pressures
     and vice versa.  Remember, this field shows the sea-level
     pressure gradient, which may not be the same as the surface
     pressure gradient.

Strong winds of greater than 50 knots immediately behind cold fronts occur 3-6 times per year. It is most commonly observed in the northwest desert across the Salt Flats and at times in the urban corridor along the Wasatch Front although these events can occur anywhere in the state. The events are most likely in the fall and spring. Blowing snow, and more often blowing salt on Interstate-80 are associated with the events. The "fishing" fleet harvesting brine-shrimp-eggs in the Great Salt Lake during their winter season also is affected by these wind events. Last year more than 8 boats were sunk during one of these events.

Physically, these events are associated with tightening of the post frontal pressure gradient. Sometimes a storm system will propagate eastward with an unusually strong pressure gradient just behind the surface front. But other times, the pressure gradient will develop rapidly in response to diabatic processes.

In the first case, examination of predicted and observed 850 height and sea-level pressure gradients in the post frontal region will often depict conditions for high winds. Close monitoring of surface observations and isallobaric tendencies as the front moves into the Great Basin will usually identify the events with enough lead time to issue warnings and advisories.

In the second case, diabatic processes such as evaporation and melting behind the front will cool the air and raise the pressure, while ahead of the front solar heating can add to an already well-mixed, warm airmass to lower the pressure. These processes can act together to locally increase the pressure gradient in just a matter of hours. An afternoon or early evening frontal passage is ideal to produce the pre-frontal warming and pressure falls. The June 5, 1995 frontal passage over northwest Utah is probably the extreme example of this type of an event. Post frontal winds exceeded 100 mph in this case and there was not a history of strong winds with the front upstream over Nevada and Idaho. This type of event is sometimes referred to as a "density current". Tools for diagnosing this type of wind event are pretty much the same as those for the pre-frontal winds, but with three additions.

     Tools for post-frontal high winds:

     1. the same tools as for pre-frontal winds

     2. isallobaric tendencies:  Look at pressure tendencies at
     least every 3 hours, and every hour is best.

     3. mesonet graphic: Since most of these events begin or are
     maximized in the western deserts, the mesonet graphics, with
     their 15 minute temporal resolution, will often be the first
     indication that an event is beginning.

     4. Doppler velocity:  Because these events are post-frontal,
     there are often many reflectors for the radar.  The velocity
     data can provide information about wind speeds.  Two cautions
     of note are: 1) the KMTX beam is at least 2500 feet above the
     surface, and generally higher, so strong surface winds in a
     diabatically driven event may be hard to see, and 2) northwest
     winds will be perpendicular to the beam over much of the west
     desert, so the strong winds may not show up until the front
     passes a point due south of the radar.  In the June 5, 1995
     event the strong winds could be seen in the velocity data only
     after the front reached the SLC airport, even though mesonet
     observations showed strong winds much earlier.

High winds associated with terrain are quite common in Utah during the cool-season. There are three physical mechanisms responsible for these winds. Sometimes these mechanisms act alone, but more often high winds are a result of a combination of these mechanisms.

1. Gap flow:
Strong gap flow results when a mountain barrier effectively separates two airmasses of different densities. The result is a strong surface pressure gradient across the mountains. Air moves through gaps in the terrain from high to low pressure, and since the terrain channels the winds, they accelerate along the pressure gradient instead of becoming geostrophic.

2. Momentum mixing downward:
In this situation, strong winds aloft blowing perpendicular to the terrain barrier reach the surface in a well-mixed environment. The airmass must be relatively unstable or the strong winds will remain aloft. In the vicinity of terrain, the winds will be strongest at higher elevations and strong winds near the surface will be a function of the mixing depth of the atmosphere.

3. Mountain waves:
High winds due to mountain wave activity is a bit more difficult to picture physically, and to some extent, the actual process is not completely understood. Essentially, flow across a mountain barrier results in a high amplitude wave on the downstream side of the mountains. There are a number of different theories on how this happens, but the end result is that winds at the surface in the lee of the mountains can often reach very high speeds, and are frequently stronger than the winds aloft. Forecasters in locations to the east of major mountain ranges, such as Denver and Reno are big believers in mountain waves, while forecasters on the west side of mountain ranges, such as SLC and Seattle tend to view most events from a gap wind perspective. Since the winds aloft are more frequently from the west, mountain waves are much more common on the east side of mountains, but there is clear evidence from a number of west-side locations of mountain wave activity.

The conditions for mountain waves are as follows:

     1. a cross barrier wind component near ridgecrest
     2. some stability or an inversion near or above ridgecrest
     3. a critical level at higher levels in the atmosphere (a
     level where the wind component perpendicular to the axis of
     the mountain range either reverses or goes to zero)

Wasatch Front "Canyon" winds:

The name implies that these are primarily gap winds. This turns out to be a highly controversial assumption. There are times when the strong winds are confined to the mouths of canyons where a substantial break in the mountains occurs. This is most commonly observed at the mouths of Weber, Parleys, Emigration, Provo, and Sardine canyons along the Wasatch Front, and Logan Canyon in the Cache valley. Ogden canyon represents another significant gap, but observation of high winds near this canyon mouth is uncommon. Either they don't occur, or there isn't a reliable observer. Big Cottonwood, Little Cottonwood, Millcreek, and American Fork canyons extend into the mountains with terrain that reaches at or above 700 mb. Gap winds are rarely, if ever, observed at the mouths of these canyons.

The most severe high winds are reported in Davis county in the Farmington/Centerville area. Farmington Canyon doesn't represent much of a gap through the Wasatch, and in general, the terrain in this area is more like an unbroken steep escarpment. These winds would seem less like gap winds and more like mountain wave events or well mixed events.

Even in what would appear to be classic gap wind events, field observations by University of Utah students and faculty has shown repeatedly that during "canyon" winds, the strongest winds are along the edges of the canyon mouths and the winds actually decrease in the canyons themselves, where true gap winds would be expected to be higher.

There are certainly events that have the characteristics of gap winds, but most of the really significant events along the Wasatch Front and Cache valley appear to have characteristics of both gap winds and mountain wave type events. And the biggest events, such as the April 4-5, 1983 windstorm appear to be more mountain wave in character. In this event the highest anemometer wind was 104 mph at Hill AFB, near the mouth of Weber Canyon, but damage indicates many gusts from Ogden to central Davis county were likely near 120 mph. UP&L reported damage or total destruction from 54 major transmission towers, 12 flatbed railroad cars with loaded trailers were overturned, 15-20 semis were blown over, and nearly every glass window in downtown Ogden was destroyed. Both deep easterly flow aloft and a pressure gradient across the Wasatch were present. Unlike many Wasatch Front windstorms that tend to peak late at night and early in the morning, this event did not show the diurnal tendency.

     Tools for diagnosing Wasatch Front "Canyon" winds:

     1. the pressure gradient: The sea-level pressure gradient is
     used most frequently because surface observations are
     available every hour.  However, reduction to sea-level can
     lead to misleading patterns, and the 850 mb heights offer a
     better indication of the surface pressure gradient.  Gridded
     model output from the Meso-eta offers 850 height forecasts
     with 3 hour temporal resolution.  The most favorable
     configuration for the isobars/isohytes is when they are
     parallel to the terrain and closely spaced near the Wasatch,
     indicating the mountains are separating two different
     airmasses.  If the isobars/isohytes are oriented more east-
     west, this is less favorable, and may indicate more northerly
     winds.  If the isobars/isohytes are packed together east of
     Evanston, WY, then this may indicate the Wasatch are not the
     point of separation and again, high winds are less likely.  A
     sea-level pressure difference of 8 mb between LND and SLC has
     tradition on its side, but hard and fast rules are not
     recommended.

     2. temperature gradient: Coincident with the pressure gradient
     is the thermal contrast across the barrier.  Colder air should
     be on the east side of the mountains for high winds along the
     Wasatch Front.  If there is little temperature contrast, or if
     the air at 850 mb and 700 mb is colder in the Great Basin than
     in Wyoming, then high winds are less likely, even in the
     presence of other factors.

     3. winds aloft: Winds at 700 mb should have some easterly
     component for strong events, although 10-20 knots is often
     sufficient.  If the easterlies are strong, a widespread event
     is more likely.  A critical level higher in the atmosphere is
     often present and is typically associated with northerly or
     westerly winds at 500 mb.

     4. cross-sections: Cross-sections from LND to SLC with
     isentropes and winds displayed can help identify good mountain
     wave events and to some extent can also help with gap winds.
     The isentropes should slope downward toward SLC with the
     greatest slope near the Wasatch.  This indicates a deep cold
     airmass on the east side of the mountains, and the winds aloft
     will also tend to move along the sloping isentropic surfaces.
     The cross-sections can also clearly show critical levels in
     the winds higher in the atmosphere.  These cross-sections were
     produced via pcgrids in the past, but will be available as
     ntrans meta-files this year.

These high wind events appear to be a combination of gap winds and momentum mixing downward. Mountain waves are possible in these events, but study of the events has been limited so far. A surface pressure gradient and north to northeast winds aloft are the main ingredients for these events. The terrain drops off sharply south of Cedar City towards St. George. Although there are not "gaps" through the terrain, there are north-south oriented canyons that extend northward into the plateau. The strongest reported winds tend to be near the mouths of these canyons near the towns of Hurricane and La Verkin.

The pressure gradient should be primarily across southwest Utah, with the isobars/isohytes oriented mainly east-west. The 850 mb and 700 mb winds in the model gridded data should be from the north-northeast. Meso-eta lowest level winds are typically in the 15-25 knot range when these events occur, with peak gusts of 50-70 mph often reported. Synoptically, these events are post-cold frontal with moderate to strong cold advection in progress.

Downslope windstorms east of the Wastach Plateau:

The area east of the Wasatch Plateau has the physiographic characteristics necessary for downslope windstorms in westerly or northwest flow. The towns from Price south to Castle Dale, including Huntington and Orangeville are all located east of a major mountain barrier with a relatively steep escarpment on the east side. Mountain wave events and downward momentum type events should be common in this area. Observations are scarce so it is likely that many (most) events in this area go unreported, but high wind events do occasionally get reported from this region. We plan to install some anemometers in this area in the next year, so these events may become more noticeable.

Strong flow perpendicular to the Wasatch Plateau, which should be represented by 700 mb winds of 40 knots or greater is the key. Mountain wave events would ideally have an inversion just above 700 mb, although a critical level is less likely in westerly flow. However, as previously mentioned, some research indicates wave breaking in westerly flow can produce its own critical level. Probably more common than mountain wave events in this area are windstorms with well mixed environments, particularly associated with deep cold advection in post-frontal situations. Look for steep lapse rates and large pressure rises in isallobaric analyses to go along with strong winds aloft.

VII. HEAVY RAIN AND RAIN-ON-SNOW FLOODING

Heavy rain and associated flooding during the cool-season are most common in the Virgin River Basin of southwest Utah. It can also be a problem in other parts of Utah during the snowmelt that occurs in May and early June, although these events are more "warm-season" in character. In February 1986 a rare mid-winter flood occurred on the Weber River in Morgan county (it was declared a disaster area) when heavy rain fell at high elevations and combined with the rapid melt of lower and mid-elevation snow. Other than this rare event, most mid-winter flooding in the last 20 years has been confined to southwest Utah.

St. George is just under 3000 feet above sea-level and the Pine Valley mountains just north of town rise to over 10,000 feet. This relief is equal to that found in southeast Salt Lake county. Little is known about mesoscale circulations in this area, but most flood-producing events result from a combination of dynamic and orographic forcing. Flooding in Washington county takes place when warm, moist (often tropical origin) air associated with a trough in the sub-tropical jetstream moves through the area. Prolonged south to southwest flow ahead of the trough impinges on the Pine Valley mountains to produce heavy rain. Similar flooding is often in progress over California and/or Arizona when these events take place. Snow levels are often near or above 9000 feet. It is worth noting that most of the flooding is from the rain falling into saturated conditions, and the actual contribution to the runoff from melting snow is usually small.

The most common flooding takes place on the Santa Clara river and other smaller tributaries. Flooding can also take place on the mainstem of the Virgin in these events. The tools for forecasting these events are essentially the same as those for forecasting heavy snow, but also include precipitable water in the "convect" meta file.

VIII. DENSE FOG

Dense fog (visibility of 1/4 mile or less) is very common in Utah during the cool-season. These events can be divided into two types. The first are episodic in that the dense fog forms immediately after a snowfall when the skies clear and radiational cooling leads to shallow, but dense fog. Light surface winds are required for this type of event. The second type of dense fog event evolves more slowly over a period of days as a strong inversion works its way downward toward a snow covered (or partially snow covered) valley in extremely quiescent conditions with little or no wind through a deep layer of the atmosphere.

The first type of dense fog event is most common in the Cache Valley, the high valleys east of the northern Wasatch crest, the southwest deserts near Delta, Milford, and Cedar City, and in the southeast near Blanding and Monticello. Anytime snowfall is followed by a clear night with light winds, dense fog is a possibility. Although fog does form on clear nights immediately after snowfall along the Wasatch Front, it is generally not widespread and usually does not lower visibility to 1/4 mile.

The second type of event is most common along the Wasatch Front, Uinta Basin, Cache Valley and the northwestern deserts. These events are nearly always confined to the months of December, January, and February. Once established, these events can persist for days and sometimes weeks, although visibility rarely stays below 1/4 mile all the time. These events often start with patchy fog and the formation of a stratus layer. Often this will not occur until about 3 days after the last snow event. It usually takes at least 3 days for a strong subsidence inversion to lower to near the surface. As long as the subsidence inversion is more than two thousand feet above the surface, the mixing layer is sufficiently deep to prevent dense fog formation. The elements that lead to these events are:

     1. snow covered ground, patchy snow cover is sufficient, if
     there is no snow, these events are very unlikely

     2. strong ridge aloft

     3. light winds through at least 700 mb

     4. clear skies, any cloud cover will usually prevent dense fog
     formation

     5. a subsidence inversion near the surface resulting in a very
     shallow mixed layer

By the end of February, there is enough solar insolation, that the presence of the above conditions will not result in dense fog if an event is not already in progress. If dense fog and stratus are already present, they can persist into early March.

IX. FREEZING RAIN AND DRIZZLE

Freezing precipitation occurs when super-cooled liquid strikes a sub-freezing surface. There are two ways for this to happen.

The classic freezing rain event takes place when snow falls through a layer of above-freezing air that melts the crystals, which then fall into a sub-freezing layer in contact with the ground. The warm layer is usually about 100 mb thick. These events are generally rare in Utah because most precipitation events are associated with enough mixing that a sub-freezing layer does not remain trapped near the ground.

A more common event in Utah, although still relatively rare, is freezing drizzle. In these events the entire sounding is below freezing, but the saturated layer is shallow. Research indicates that with cloud tops warmer than -10C and the entire sounding below freezing, ice nuclei will not be present. Super-cooled cloud droplets will gradually collide and coalesce until their size results in fall velocities sufficient to reach the ground. This drizzle forms ice upon contact with the sub-freezing surface. This process is essentially the same as the "warm-rain" process where collision and coalescence dominate over the more common process where ice nuclei combine with super-cooled liquid and water vapor to produce snow. If there are ice crystals present, freezing drizzle will not form.

In Utah, a stratus layer associated with stagnant conditions will often be in place during freezing drizzle events. The sounding will show nearly saturated but sub-freezing conditions from the surface to the base of the inversion. This will often appear as a nearly isothermal layer of -5C or warmer. The frequency of these events is approximately one or two per year, and the physical mechanism that triggers the production of precipitation is unknown. Model soundings are probably not accurate enough to help with this forecast problem. Don't discount a report of freezing drizzle (or rain) just because the sounding is completely below freezing.