SALT LAKE VALLEY DENSE FOG INITIATION STUDY

Predicting the onset of inversion-related dense fog episodes (1/4 mile visibility or less) in the Salt Lake Valley has always been a challenge to forecasters. A climatology of 30 years yields an average of 12 dense fog days at the Salt Lake City airport every year. Over the 3-year baseline covered by this study, about 15 percent of the dense fog days were precipitation related and 75 percent were inversion related. The remaining 10 percent had both precip and inversion-related characteristics.

Most of these dense fog events occur several days after a storm as strong high pressure builds over the region. As high pressure builds, the atmosphere warms from the top down due to a subsidence inversion. This subsidence inversion acts like a cap on the cold air. Owing to mountain-valley topography, the lack of surface heating during winter months, and atmospheric stagnation, cold air often becomes trapped in many of the valleys of northern Uah. Moisture and other particulates build up in these trapped pools of cold air under the subsidence inversion. When conditions become right dense fog and stratus can become widespread across the valleys of the Wasatch Front, lasting for weeks at a time during some years. Such episodes if nothing more than dreary and depressing for most can have serious negative impacts on air quality, airport operations, and highway safety (i.e. multiple car pile-ups).

STUDY OBJECTIVES

Despite local studies and forecaster rules of thumb, prediction of the onset of dense fog episodes remains difficult. The intent of this study is two-fold:

1. check the validity of commonly used rules-of-thumb, and

2. investigate the use of inversion height as a possible predictor of onset of inversion-related dense fog episodes.

These rules-of-thumb include the following:

1. The "3-day rule": the idea that it takes 3 days after a storm before inversions intensify to the point where dense fog forms

2. The "snow cover rule": the idea that snow cover is necessary as both a moisture source and inversion enhancer in the development of dense fog

3. The "broken snow cover rule": the idea that broken snow cover is preferable over widepsread snow cover for the formation of dense inversion-related fog

4. The "precursor visibility restrictions rule": the idea that inversion-related dense fog rarely forms without precursor haze or light fog the afternoon before

5. The "near or above freezing temperatures rule": the idea that temperatures sufficiently warm for snowmelt are necessary the afternoon before the event in order to juice up the near surface atmosphere.

1. The "3-day rule": the idea that it takes 3 days after a storm before inversions intensify to the point where dense fog forms.

During the 3-year period of study, 44 days with dense fog were reported. Twelve of these events were directly associated with precipitation and the remaining 32 were inversion-related. Of the 32 days of inversion-related dense fog, 21 were repeat days of dense fog following onset. For the purposes of checking the rules-of-thumb as predictors of dense fog onset, we will only consider the remaining 11 days on which inversion-related dense fog began. Once dense fog commenced, recurrence was likely in the absence of weather disturbances.

Following is an event-by-event list of the number of days following a storm when dense fog began (for the purposes of this study, "storm" has been defined as a measurable precip event. This does not include measurable precip events due to airport cloud seeding):

1983-1984 2 days, 2 days, 3 days, 5 days
1987-1988 6 days, 11 days, 5 days
1988-1989 6 days, 5 days, 3 days, 5 days
1989-1990 21 days, 5 days, 9 days
1990-1991 14 days, 2 days, 4 days, 5 days
1991-1992 9 days, 3 days, 4 days
1992-1993 3 days, 3 days, 8 days
1993-1994 NO strictly inversion-related dense fog days
1994-1995 4 days
1995-1996 NO strictly inversion-related dense fog days

2 . The "snow cover rule": the idea that snow cover is necessary as both a moisture source and inversion enhancer in the development of dense fog (for results, see rule-of-thumb 3)

3. The "broken snow cover rule": the idea that broken snow cover is preferable over widepsread snow cover for the formation of dense inversion-related fog.

Rules-of-thumb 2 and 3 will be considered together.

1983-1984 6", 7", 5", 3"
1987-1988 3", 2", 1"
1988-1989 1", 5", 4", 6"
1989-1990 0", 0", 0"
1990-1991 0", 3", 3", 2"
1991-1992 T, 4", 3"
1992-1993 3", 4", 1"
1993-1994 No Events
1994-1995 3"
1995-1996 No Events

CONCLUSION: The "snow cover rule" is a fairly reliable rule-of-thumb (19 out of 25 times). Although this rule works in most cases, one may well ask the question "Isn't snow cover pretty much a fact of life in SLC for most of the winter anyway?" Yes, but qualitatively speaking, it makes meteorological sense that snow cover will do a couple of things that enhance the inversion and the chances of inversion-related dense fog. Snow on the ground slows daytime surface heating and enhances nighttime radiational cooling, both of which serve to strengthen the inversion.

It is important to note, however, that the absence of snow cover does not preclude the formation of dense fog. Such was the case for the winter of 1989-1990.

CONCLUSION:As for the "broken snow cover rule", aside from being hard to define (what is "broken snow cover" anyway?) it appears to be unfounded. In most cases, snow depth is at least a couple inches. A couple of inches at the airport usually means significantly more at other places in the valley. One could argue that in the cases of no snow at the airport, there cold well have been snow elsewhere in the valley. A review of climate records indicates that in the five cases above, where zero or trace snow amounts were reported at the airport, there had not been any more than a trace at the airport in the entire preceeding week. This makes it unlikely that other parts of the valley had much if any snow on the ground.

4. The "precursor visibility restrictions rule": the idea that inversion-related dense fog rarely forms without haze or light fog reducing the visibility to 6 miles or less the day before (haze and fog being an indicator of significant atmospheric stagnation).

1983-1984 YES, YES, YES, YES
1987-1988 YES, YES, YES
1988-1989 YES, YES, YES,YES
1989-1990 YES, NO, YES
1990-1991 YES, YES, YES, YES
1991-1992 YES, YES, YES
1992-1993 NO, YES, YES
1993-1994 No Events
1994-1995 YES
1995-1996 No Events

CONCLUSION: This rule-of-thumb worked well; in 23 out of the 25 cases over the 10-year period precursor restrictions to visibility due to haze or light fog were exhibited the preceeding day.

5. The "near or above freezing temperatures rule": the idea that temperatures sufficiently warm for snowmelt are necessary the afternoon before the event in order to juice up the near surface atmosphere.

1983-1984 37, 33, 36, 45
1987-1988 31, 36, 34
1988-1989 39, 35, 35, 37
1989-1990 NA, NA, NA
1990-1991 NA, 29, 25, 36
1991-1992 31, 32, 40
1992-1993 38, 29, 36
1993-1994 No Events
1994-1995 37
1995-1996 No Events

CONCLUSION: With most (18/21) of the dense fog onset events (excluding, of course, cases with no snow available for melting, marked "NA") temperatures the day preceeding had been in the 30s or warmer (near or above freezing). The thing to remember here is that cold temperatures hinder snowmelt reducing the chances of dense fog formation.

Procedures: for a period of 3 winters, all 12Z soundings were examined to determine the height and strength of any inversions (done using SHARP program). All months where stagnation appeared to be any kind of a problem were considered (generally November through February, but in some years as early as October or as late as March).

Inversion strength and height were noted for each day there was an inversion. Inversion strength was broken down into the following categories: 1. "isothermal" (0 degrees C/1000 feet +/- half a degree), 2. "weak" (0.5 to 1.5 C/1000 feet), 3. "strong" 1.5 to 4.0 C/1000 feet, and 4. "very strong" (greater than 4.0 C/1000 feet). During this detailed inspection of all 12z soundings for the 3-winter period, it became clear that a "weak" inversion or "isothermal" lapse rate was never sufficient for the onset of dense fog. Only "strong" or "very strong" inversions (about 1.5 C/1000 feet or greater) lead to the onset of dense fog.

Inversion height was defined as the point at which temperature began to decrease with height (the inflection point between warming with height and cooling with height). The inversion height defines the depth of the stagnant, moist layer. Click here for examples of weak, strong, and very strong inversions.

NOTE: During the 3-year period of study, 44 days with dense fog were reported. Twelve of these events were directly associated with precipitation and the remaining 32 were inversion-related. Of the 32 days of inversion-related dense fog, 21 were repeat days of dense fog following onset. This leaves us with 11 inversion-related dense fog onset events. Once dense fog commenced, recurrence was likely even when some variables for dense fog were no longer favorable.

The above examination yielded an inversion threshold height of about 3700 feet above ground (7900 feet msl); the height of the inversion at or below which dense fog frequently formed and above which it rarely formed. The following table summarizes how well this inversion threshold height worked. cellpadding=5 border=5 cellspacing=5


TABLE 1: INVERSION THRESHOLD HEIGHT AND ONSET OF DENSE FOG

	inversion <= 3700 ft		inversion > 3700 ft
YEAR	dense fog (positive cases)	NO dense fog (negative cases)	WITH dense fog (exceptions)
1983-84	4	5	1
1988-89	3	11	1
1989-90	2	8	0
TOTALS	9	22	2

Note: see footnote for explanation of events not included in this tally.

Tables 2, 3, and 4 following summarize the key meteorological elements for each of the 3 groupings from Table 1 above (positive cases, negative cases, and exceptions). By scanning these tables you can see the complexity of this forecast problem, i.e. how numerous variables need to come together for dense inversion- related fog to form in the Salt Lake Valley. Although the "inversion height threshold" is only one of several important criteria, it helps with the forecast problem.



TABLE 2: THE 9 POSITIVE CASES: Dense Fog Forms/Initiates with Inversion Height AOB 3700' (KEY FACTORS UNDERLINED)

DATE	INVERSION LAPSE RATE	INVERSION HEIGHT	PRECURSOR FOG OR HAZE	SNOW ON GROUND	WIND SPEEDS 4 LOWEST GATES OF SOUNDING	MOISTURE ABOVE INVERSION
JAN 28, 1984	2.7 C/KFT	2800 FT	YES	6"	4/2/5/10	CLEAR
FEB 3, 1984	2.7 C/KFT	2800 FT	YES	6"	4/4/5/12	NEGLIBLE
FEB 20, 1984	1.4 C/KFT	3500 FT	YES	7"	0/2/1/4	CLEAR
MAR 8, 1984	2.0 C/KFT	2700 FT	YES	3"	2/4/3/4	NEGLIBLE
DEC 5, 1988	2.0 C/KFT	3300 FT	YES	1"	4/3/6/2	CLEAR
DEC 31, 1988	2.8 C/KFT	3200 FT	YES	5"	2/26/33/40	DEEP SATURATION AND MEGA WINDS (MAJOR ANOMALY)
JAN 18, 1989	3.6 C/KFT	2200 FT	YES	6"	4/3/3/4	NEGLIGIBLE
NOV 11, 1989	1.8 C/KFT	3700 FT	YES	NONE	4/3/5/9	CLEAR
JAN 18, 1989	3.3 C/KFT	2700 FT	YES	NONE	4/0/0/0	CLEAR

CONCLUSIONS for the POSITIVE CASES:
The Inversion Height Threshold works well assuming other key ingredients for dense fog are in place. Aside from having a STRONG INVERSION AOB 3700 FEET, note the following:

1. each case had precursor haze or fog,
2. all (except the one major anomoly) had essentially clear skies above the inversion,
3. all (except the anomoly) had light winds (less than 10 kts) in the lowest 3 gates of the sounding,
4. most of them had snow on the ground.

1. in 20 out of the 24 cases, at least one key element was lacking for the initation/formation of dense fog.
2. in 4 of the cases (those labelled as "exceptions"), virtually everything looked primed for the formation of dense fog, although none formed, attesting to the complexity of this forecast problem
3. a substantial number of the cases (9) had no precursor haze or fog (in my mind this is the most reliable atmospheric predictor)
4. half of the cases had no snow on the ground (a much higher percentage than the cases where dense fog did form).
5. 9 of the cases had significant cloud cover
6. 10 of the cases had significant low-level winds
7. 3 of the cases were on consecutive days very early in the cool season (October)...climatologically too early for the formation of inversion-related dense fog.

1. strong inversion
2. precursor haze or fog
3. snow on the ground
4. light winds in the lowest four gates of the sounding
5. little or no moisture above the inversion (essentially clear skies)

PRIMARY CONCLUSIONS: Dense inversion-related fog rarely forms/initiates

1. when the depth of the inversion is greater than 3700 feet (only two such cases during the 3-year period) and

2. when the lapse rate is less than 1.5 C/1000 feet.

1. Significant moisture/clouds above the inversion layer and strong winds in the inversion layer hinder the formation of dense fog (no big surprise). Note that both of these elements can be related to approaching storminess with the break down of high pressure.

2. The "snow cover" and "precursor haze/fog" rules were reaffirmed.

A. The "precursor haze/fog rule" worked better than any other rule of thumb (23 out of 25).

B. The "snow cover rule" worked well, but there were definite exceptions (19 out of 25).

C. The "near of above freezing temperatures" rule worked well (18 out of 21).

D. The "3-day rule" did not work well (5 out of 25).

E. The "broken snow cover rule" was unfounded.

A. Definition of Inversion Height Threshold:
the height above which dense inversion-related fog rarely formed...and below which it frequently formed.

B. Inversion Strength: it was also found that inversion-related dense fog only initiates when the inversion lapse rate is greater than about 1.5 C/1000 feet.

C. examples of weak, strong, and very strong inversions.

D. Additional Observations: in the process of researching the inversion height threshold, the following additional observations were made made:

1) significant moisture/clouds above the inversion layer and strong winds in the inversion layer hinder the formation of dense fog (no big surprise).

2) the "snow cover" and "precursor haze/fog" rules-of-thumb were reaffirmed.

1. cases/days following initial onset of dense fog for the same stagnation episode...the inversion may lift a bit and dense fog persist...or vice versa.

2. precip-related episodes (i.e. measurable precip...trace precip is common-place with dense fog episodes...measurable precip is not out of the question with subsidence inversion dense fog episodes, so these rare cases were carefully weeded out)...light precip falling into an already moist stagnant/inverted layer can lead to dense fog...heavy precip in a non-inverted atmoshpere can lead to a report of "dense fog" when fog is reported in conjuction with the precip...dense fog can form immediately on the heals of precip if winds go calm and/or skies clear regardless of whether the atmosphere is inverted or not. These are not the type of dense fog events we're trying to isolate and study. We are trying to get a handle on dense fog episodes related to fair weather, high pressure, subsidence, etc.




APPENDIX: ADDITIONAL DETAIL ON THE SYNOPTIC PATTERN FOR THE NEGATIVE CASES (obtained from the following sources: soundings, ntrans archive program (to view charts from the day in question; 500 mb 700 mb, 250 mb, and surface), and surface observations)

very strong inversion aob 3700 feet, but no dense fog (24 cases)

PH/F = PRECURSOR FOG/HAZE
SOG = SNOW ON GROUND
PDM = PREVIOUS DAY'S MAX TEMP

10/28/83  PH/F=YES SOG=0 PDM=66 low level moisture lacking a
          little, although not completely dry.  Also, way too early
          in the winter season for the typical dense fog inversions
          to form.

10/29/83  PH/F=YES SOG=0 PDM=66 way too dry (and no snow on
ground
          to help juice things up) despite strong high pressure
          surface and aloft.  Too early in the winter season,
          etc....

10/30/83  PH/F=NO SOG=0 PDM=70 ditto 

           HAZE...FIRST 3 DAYS BEING PRECIP-RELATED.  too much
          short-wave energy (H5 shortwave went through between 00
          and 12z, strong 500 mb ridge over Oregon with Utah in lee
          of ridge in cyclonic nw flow at 500 mb).  Anticyclonic NW
          WAA flow at 700 mb.

           HAZE...FIRST 3 BEING PRECIP RELATED.  too much
short-wave
          energy (500 mb ridge over E. Pacific with moderate-strong
          cyclonic NW flow over Utah, vort lobe upstream centered
          over Washington; moist, cyclonic, WAA NW flow at 700 mb).

          Otherwise, sounding looked very favorable  for F+
          formation...strong inversion, moist inversion
layer.

12/03/88  PH/F=YES SOG=1 PDM=40 BIG EXCEPTION: everything
looked
          primed for dense fog; can see no reason why it did not
          form.  Precursor vis restrictions, snow on ground, strong
          surface high, strong ridging at 500 and 700 mb.  moist
          inversion layer, light inversion layer winds.

           a bit suspect)  NOTE: dense fog DID finally form
the next
          day

12/13/88  PH/F=NO SOG=0 PDM=49 a little too dry, too windy,
and too
          much short-wave energy.  No precursor vis restrictions
          and no snow on the ground.  on the fringes of strong
          zonal flow at 500 mb.  500 mb shortwave upstream over
          Oregon/Washington.  extremely thin layer of moisture at
          surface.  winds in lowest four gates of sounding 6, 8,
          15, 12kts.  Tail of surface trof/cold front on doorstep
          (Salt Flats)

           evidenced by sounding and absence of precursor vis
          restrictions (and no snow on the ground)...otherwise
          everything looked pretty favorable for dense fog
          formation.  light winds in the inversion layer, strong
          ridging at 500 & 700 mb, and strong surface high
          pressure.

01/16/89  PH/F=NO SOG=7 PDM=27 too dry and windy.  despite
strong
          surface high, moderate straight NW flow at 500 mb and
          moderate WAA moist anticylonic NW flow at 700 mb.
          extremely thin layer of moisture at surface (note no pre-
          cursor restrictions to vis).  too windy...lowest four
          gates of sounding 10, 14, 14, 16kts.

01/17/89  PH/F=YES SOG=7 PDM=34 breezy with some shortwave
          energy...winds apparently dried-out/cleared airmass
          overnight...12z sounding shows inversion layer very dry.
          moderate cyclonic nw flow at 500 mb with vort
          lobe/shortwave centered over E. Oregon.  Surface high
          weakens, but still fairly strong.  Winds in four lowest
          gates 10, 10, 10, 8kts.

          NOTE: Dense fog DID settle in on the next day (the
18th)
          when winds went light and high pressure began building
          over the region.

02/16/89  PH/F=YES SOG=9 PDM=35 In general, this case looked
pretty
          favorable for the formation of dense fog...strong surface
          high pressure, precursor haze/fog, snow on the ground,
          moist near-surface layer.  Classic dense fog
          sounding...very strong inversion, dry aloft, moist in the
          lower portion of the inversion.  Possible explanations
          for the lack of dense fog formation would be fairly
          strong winds near the top of the inversion and weak
          shortwave energy aloft.  Winds in the four lowest gates
          of the sounding were 4, 9, 18, and 20kts.  There was also
          a weak 500 mb shortwave embedded in moderate westerly
          flow.  Moderate westerly surface gradient.

02/17/89  PH/F=YES SOG=9" PDM=30 similar to the 16th above,
but the
          main hindrance to dense fog formation may have been lack
          of moisture instead of winds.  This seems a bit hard to
          understand, considering the previous afternoon had
          restrictions to visibility.  Winds in the four lowest
          gates were light (2,3,8,7), but the 12z sounding
          indicated only a very shallow layer of moisture at the
          surface.

02/18/89  PH/F=YES SOG=9" PDM=34 similar to the 16th and
17th...a
          little too windy and too dry still as well as a little
          bit of shortwave energy.  Near surface winds of
          4,7,16,15kts.  Low-level sounding moisture still weak.
          Flat 500 mb ridge with strong shortwave on Cal coast.
          WAA sw flow at 700 mb.  Ridge decaying....third of an
          inch of precip following day.  NOTE: .32" precip next day
          with precip-related dense fog.  Precursor vis
restr...but
          moisture lacking??

11/20/89  PH/F=YES SOG=0" PDM=53 everything looks favorable
for
          dense fog onset...light winds, strong high pressure
          surface and aloft, abundant near surface moisture,
          precursor fog/haze.  No explanation for the absence of
          dense fog formation.  At least dense fog DID eventually
          form the next day.

12/04/89  PH/F=YES SOG=0" PDM=41 moisture lacking.  Everything
else
          looked great (strong high pressure surface and aloft,
          light winds, pre-cursor haze, etc.).  The sounding
          indicated, though, that near-surface moisture was
          lacking.

12/05/89  PH/F=YES SOG=0" PDM=44 same as the 4th, above (low-
level
          moisture lacking), but in addition high pressure
          beginning to weaken.  weak shortwave energy begins to
          flatten the 500 mb ridge, winds aloft begin to increase.
          Note that there was no snow on the ground to help in
          juicing things up.  NOTE: precip-related dense fog DID
          form the next day.

12/24/89  PH/F=NO SOG=0" PDM=47 moisture lacking as indicated
by
          sounding and lack of precursor haze/fog.  Otherwise,
          looked favorable with strong high pressure surface and
          aloft and light winds.  Note that there was no snow on
          the ground to help in juicing things up, but dense fog
          DID form the next day.

01/01/90  PH/F=NO SOG=0" PDM=42 too dry, too windy, too much
          shortwave energy.  Sounding indicates very dry lowest
          levels and winds fairly breezy (4,12,13,17).  Weak
          shortwaves in wsw flow at 500 mb.  surface low/cold front
          just upstream in N Nevada.  High pressure weakening in
          advance of approaching shortwave (.15 precip next
day).

01/07/90  PH/F=NO SOG=1" PDM=38 too windy and surface moisture
a
          bit weak.  Sounding winds in 4 lowest gates 6,14,13,15.
          Sounding moisture weak and no pre-cursor haze/fog.  Flat
          500 mb ridge and moderate southerly surface gradient.
          Strong zonal flow in E. Pacific.  High pressure weakening
          ahead of approaching storm (.04 precip next day).

01/11/90  PH/F=NO SOG=0" PDM=47 everything looks totally ripe
for
          dense fog (light winds (4,2,5,11) strong high pressure
          surface and aloft)...except lack of moisture as indicated
          by sounding and lack of precursor haze/fog.  Once again,
          there was no snow on the ground to help in juicing things
          up.