Fog: Its Causes and Forecasting with Special Reference to Eastern and Southern United States (III)* J

Home , Pressure gradient, Wind direction

Fog: Its Causes and Forecasting with Special Reference to Eastern and Southern United States (III)* J. J. GEORGE Chief Meteorologist, Eastern Air Lines, Inc., Hapeville, Ga.

(2) Quantitative Studies of Fog Causss and Forecasting at Certain Airports (continued). JACK SON'VILLE, FLORIDA

T JACKSONVILLE, where the air- (high-pressure ridge) forms down the port is located only 14 miles coastal plain with the center of the A from the shore of the Atlantic HIGH to the northeast. Air is di- ocean, the fog problem is predom- rected over the water close to the inantly one of advection rather than shore as indicated in Figure 6. The of radiation. This is also true of all stations examined which were within the limits of sea-breeze effects. In such situations advection fog forms with a slight or moderate pressure gradient directed from water towards land. With such a flow, the surface wind decreases slowly at night under the combined influences of increasing stability caused by surface cooling over the land and the tendency toward production of a land-breeze which counteracts the general pressure gradient to some extent. The result is a mass of moist, almost stagnant air over the adjacent land surfaces which is easily cooled by nocturnal radiation, producing fog. When the pressure gradient is oriented any direction except from sea to land the land breeze is additive to the surface wind produced by the pressure gradient and usually no fog results because of excessive mixing of the lower layers caused by the abnormally high surface wind.

Advection Types:- FIG. 6. TRAJECTORIES FOR NE STRATUS AT JACKSONVILLE, FLA. (Line of circles is warm "Northeast Stratus" is the name front; cross-section below.) given a peculiar formation which temperature of the cold air is in- forms occasionally in winter along the creased rapidly by passage over the coast from Charleston, S. C., to Vero comparatively warm waters near the Beach, Fla., and which resembles in coast. (Mean surface-water isotherms essentials the stratus of the Califor- are entered in the figure according to nia littoral. A wedge of cold air Slocum). The flow under such a pres-

^Concluded from April and June BULLETINS, sure distribution normally takes place 1940. almost exactly parallel to the north-

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC 286 BULLETIN AMERICAN METEOROLOGICAL SOCIETY [Vol. 21 ern limit of the Gulf Stream. Con- a process of continual change. How- version of cold polar air to warm, ever, if the pressure gradient sud- moist, tropical maritime air takes denly becomes flat along the coast and place rapidly over the warmer waters, allows radiation from the land sur- while the air flowing over the cooler face to act on stagnant air for some shore waters has only a comparatively time, stabilization is rapid and the fog short path over water and is much will remain all night. less affected. This flow rapidly produces frontogenesis which seems to occur almost invariably near the northern limit of the Gulf Stream. In Figure 6, the region of frontogenesis is indicated by means of small circles between the converging lines. The showery type of weather such a flow (cold air over warm water) usually produces is easily transformed to a low-stratus type if a limit to con- vection develops at a low level. The limit in this instance is simply the warm front produced as just de- scribed along the Gulf Stream. This front is oriented north-south with the FIG. 7. FOG GRADIENT-WIND-ROSE FOR JACK- SONVILLE, FLA. cold air to the west and is illustrated in cross section along the line marked Figure 7 shows the distribution of AB in Figure 6. The wind arrows in fog according to the direction of the the cross section do not represent the gradient wind. The types of fog are air flow in the vertical as customary, remarkably consistent with the dia- but horizontal directions assuming gram. Only the stratus types from north at the top of the illustration. the southwest quadrant and occasional The front is usually shallow vertically ground fogs from the NW may be and for that reason is frequently over- considered as pure radiation types, looked in analysis. The stratus usu- and these are a small minority of the ally forms with ample ceilings and total. slowly lowers as the front approaches The large majority of fog occurs in the shore or as radiational cooling winds from some easterly direction takes place from the top of the cloud starting as ground fog, and various stratus. methods of forecasting were found A type of stratus somewhat similar necessary. During the summer months to this one occasionally is observed of June through September fog is al- during the fall and winter months most non-existent. In October it first when the flow from a cold HIGH is begins to form in appreciable quanti- just starting to have a trajectory ties and from October through Decem- across the water, and in this case ap- ber cooling curves seemed to be the parently the extremely stable lapse best available method of forecasting rate of the air mass is sufficient to the formation. Figures 8 and 9 are limit the convective layer. Such a representative composite-curves for fog, of course, can last only a few temperature variation for different hours, usually from dark to midnight, directions and velocities of the gradi- since the flow which produces it is in ent wind. To use them find the inter-

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC hours to time given by intersection. Figure 9: Curve I. Add 3 hours to intersection. Curve II. Add 4 hours to intersection.

FIG. 8. COMPOSITE CURVES OF TEMPERATURE VARIATION WITH WIND AT JACKSONVILLE in Oct.-Nov. section of the temperature curve with the dew point using the appropriate curve for the gradient wind taken from the 5 P.M. pilot-balloon run at sunset; and also the temperature FIG. 9. COMPOSITE CURVES FOR JACKSON curve should be started from its maxi- VILLE, December. mum value. Certain corrections are necessary since fog never forms im- TABLE 3. Time in hours to add to mediately at the intersection of tem- intersection of Temperature and Dew perature and dew point. These correc- Points for curves III and IV in Fig- tions follow:— ure 9. Figure 8: West-wind curve: Add 4 Gradient- hours to the time indicated by the Wind Direction Curve III Curve IV intersection and forecast only NE 4 7 ENE 5 6 light fog, visibility above 1 mile. E 4 4 East-wind curve: For due ESE 3 2 SE 2 3 east gradient winds only, add 5 SSE 3 4 hours to time found at intersection. Visibility will usually re- For the remaining winter months main I to 1 mile for winds from other methods of forecasting proved this direction. more reliable and easier to use. For For all winds NNE to SSE inclu- the winter months of January and sive except due east, add 3 to 4 February, with gradient wind from

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC [Vol. 21 the East to SSE inclusive only, and used for each group of gradient-wind with polar air only (dew points less directions which may be treated as than 59°F) Table 4 is used: possessing fairly homogeneous prop- Separate tables of this type are erties of air.

TABLE 4. Time of formation of moderate or dense fog, E-SSE gradient-wind. Dew-point Depression at Gradient-Wind Velocity (M. P. H.) Sunset (F°) 1 to 5 6 to 15 16 to 30 Over 30 0 to 5 (Use 1900 0000 (none 6 to 10 West- 1930 0139 recorded) 11 to 15 Wind Table, 2230 0230 erratic 16 to 20 Not Repro- 0100 0430 erratic 21 to 30 duced Here.) 0430 (none) erratic

SAN ANT< no, TEXAS San Antonio lies on a steeply slop- during these months is so high that ing coastal plain 140 miles northwest sufficient data to produce quantitative of the coast of the Gulf of Mexico. rules for forecasting were not avail- Its elevation is 580 feet, but the able. The following remarks serve as ground rises to 1000 feet just 16 miles a guide however : to the northwest. The steep general 1. Fog forms with gradient wind be- slope of the terrain and its rather dry tween 12 and 40 mph. from the nature seem to preclude entirely all SSW to E inclusive, provided that normal cases of radiation fog in conti- a direct path of the air from the nental air. Only during January and Gulf has been established. February were cases of this kind 2. It forms with a dew point depres- found and they were few, brief, and sion at 1830 C.S.T. of 0 to 23°. occurred mostly when rain clouds dis- 3. The higher dew point depressions sipated during the night hours. For are always associated with higher this reason, only advection fogs in winds and ceilings. air from the Gulf will be discussed for this locality. 4. A ceiling of 300 feet or lower sel- It is worthwhile dividing the fore- dom or never forms before mid- casting problem into two seasons, win- night. ter and summer, and each of these 5. Ceilings 300 feet or lower are asso- into polar and tropical air masses. ciate with either:— Advection fog Summer, Tropical mari- (a) A rapid decrease in gradient time air:— wind after 1830. June, July, August: The cases in (b) Rapidly divergent isobars which ceilings below 400 feet are at- from coast to San Antonio. tained in these months are so rare as (c) Surface winds at San Antonio to warrant neglecting them from a against gradient flow. forecasting standpoint. Roughly, a (d) Almost entirely with dew gradient flow of from 15 to 25 mph. point depressions less than is required for the formation of the 14° at 1830. stratus. (e) Such ceilings are usually not March, April, May, September, Oc- persistent for longer than 2 tober, November: The average ceiling or 3 hours at a time.

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC Summer Transitional Polar Air:— night. For low ceilings to persist During the summer months fog long the following morning, the wind does not form at this station in tran- velocity must have been rather high sitional polar air. During the early the previous evening and have de- spring and late fall, however, such creased rapidly during the night. fogs do form as ground fog usually Transitional Polar Air:— between the hours of 0500 and 0700 Most important (because of con- and last only for a few hours. sistently low ceilings) and hardest to forecast (because dew point depres- Winter Tropical Air:— (January, February, December). sion means absolutely nothing) are This air mass always forms a stratus advection fogs in old polar air which deck: never ground fog. The time has had a short history over the Gulf of formation is a function of gradient before returning to San Antonio from wind velocity and dew point depres- the southeast. sion and curves for forecasting it If the assumption is valid that even were developed. The ideal range for a short path of the cold air over the the gradient wind is SE 15 to 25 mph. warm waters of the Gulf is sufficient to add enough moisture to form fog The relation between the wind with additional cooling (such an as- velocity and ceiling under the stratus is fairly reliable and is illustrated in sumption appears to be justified since Figure 10. The time of lifting or the addition of moisture to the air under these conditions is very rapid), and if the further assumption is made that during the day, this moisture, which must be concentrated at rather low levels, is dissipated aloft by con- vection as it passes over the heated land surface between the coast and San Antonio (this assumption seems still more valid since the condition is almost invariably attended by clear skies which permit rapid heating), then the conclusion is apparent that only the air which leaves the coast at or near sunset is available to form fog at San Antonio. Accordingly, the method of forecasting this type of fog is by a simple formula which has been found to work remarkably well:— t = 18 + D/V t is the time in hours after the preceeding midnight the fog will form; 18 is the average time of sunset during the winter months, D is the dis- FIG. 10. RELATION BETWEEN GRADIENT WIND tance the air will traverse from the AND CEILING AT SAN ANTONIO, TEXAS. coast to San Antonio in miles, and V breaking the following morning is a the gradient velocity as scaled from function of the thickness of the stra- the isobars in miles per hour. In tum, which is in turn dependent upon using this formula, care must be taken the gradient wind the preceeding that the air leaves the coast some-

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC where in Louisiana, Mississippi, Ala- tially higher than at San Antonio. bama or western Florida, and returns Such fogs form at, or promptly lower to the Texas coast promptly from to, zero ceilings with gradient veloc- about the southeast, and that the dew ities less than 30 mph. The average points along the coast (say at Galves- time of lifting the following morning ton of Corpus Christi) are substan- is at 11 A.M.

NEW ORLEANS , LOUISIANA The airport at New Orleans is built it seems preferable to gauge the on filled land projecting into Lake gradient wind from the isobars rather Pontchartrain, and lies between the than balloon runs in all cases in which lake and the Gulf of Mexico, a short it is involved in fog forecasting. A distance to the south. Almost com- gradient wind which is accelerating pletely surrounded by water, it is no during the evening ordinarily delays surprise that radiation fog is prac- the formation of fog in direct ratio to tically unknown. Another outstand- the amount of acceleration, and the ing feature is the almost total absence reverse, i. e., a decelerating wind of fog of any kind during the summer causes fog to form much earlier than months from the last of May until the normally would be true for any parti- first of October. cular humidity value. The reason for Advection Fog:— this behavior is, of course, quite apparent. An increasing wind tends At this station, as at many seacoast to distribute moisture from the sur- locations, it was found that the time face upwards, while with a decreasing of formation of fog or stratus was wind, the moisture flow is directed controlled more by wind velocity than downward increasing the humidity of by humidity data. In all cases exam- the lower layers. ined, the air had a history over the Gulf of Mexico of at least eight to The proximity of Lake Pontchar- ten hours and arrived with gradient train has two important effects on the winds from ESE to WSW inclusive. formation of fog at the airport. When Forecasting is accomplished Iby the southerly pressure gradient de- plotting each gradient wind direction creases during the night to a value and velocity on a regular dew point below 15 mph by morning, the tem- depression-time of formation graph. perature distribution between the lake, which is shallow and quite cold It can be readily seen that this pro- in winter, and the land is such as to duces a large number of curves and set up a lake breeze quite early. The complicates forecasting considerably. fog condition which formed initially A still more complicating factor is with southerly winds and is intensified that acceleration of gradient wind is over the cold waters of the lake is very important and must be taken now blown back inland with the lake into account also. It has been found breeze, and the airport may report fog at this station that the daytime bal- until early afternoon under such sit- loon runs yield gradient values about uations since solar heating is inef- % under the amount indicated by the fectual in dissipating the fog over the spacing of the isobars, and that night water surface. balloon runs yield values about 50% greater than the isobars. This same The second important lake effect relationship also seems to hold at occurs during the colder winter some other stations, in particular months when a cold front crosses the Houston, Texas, and for this reason station at night. Under these condi-

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC tions, when the dew point of the air five to eight hours. Such a formation, which reaches the lake is above the however, requires a pre-cooling of temperature of the water surface, and several hours after dark before the the gradient wind behind the front is somewhat limited area of the lake is less than 30 mph, dense fog usually able to finish the process of fog forma- attends the passage of the front and tion in the brief time the air is travel- when formed remains for as much as ing over it. ( Concluded)

What is Going On in the Russian Arctic? A recent inauguration of regular Moscow by the Northern Sea Routes mail and passenger service on the Admin., no. 3, 1940. This office issued Moscow — Chutotsk Peninsula 5000- 4297 forecasts during 1939 against mile long airway with its Eastern 3176 in 1938. For the benefit of for- terminal at the very border of the eign shipping 189 were in English. United States, brings out the question In addition to regular forecasts 329 how much do we know about the me- were made for aviation activities and teorological organization, adjacent to 110 for ships on special detail. Official our own, which made this enterprise verification for forecasts of this office possible. Unfortunately very little in- during 1939 was 80%. This figure formation reaches the United States. however is considered below the actual This and the ever growing importance check because formal certification is of more comprehensive data from the based on reports of shore stations Arctic justifies the bringing to atten- while forecasts are made for areas of tion the following fragments of in- craft's locations. formation. In addition to standard forecasts, The scientific activities in the Rus- expressed in words, in cases, such as sian arctic sector during the first ice-reconnaissance flights, covering quarter of 1939 as they are given in a often as much as 3000 km and lasting journal published by the Arctic Insti- 16 hrs without landing, forecast tute at Leningrad, "Problems of the charts were prepared. These charts Arctic" comprised:— are similar to synoptic ones and the 1. Meteorological observations were made at 47 sta 2. Hydrologic—shore observations were made at 47 3. Hydrologic—sea observations were made at 8 4. Pilot-balloon observations were made at 20 5. Radiosonde observations were made at 7 6. Kite ascents observations were made at 5 7. Magnetic observations were made at 5 8. Actinometric observations were made at 4 9. Atmospheric electricity and radio propagation observations were made at 1 10. Weather bureau offices 5 11. Airway weather stations 37 (These data do not include inland Siberian stations). At Tikhaya Bay the highest alti- same symbols are used. The difference tude reached by a pilot balloon was is that the weather conditions repre- 14,000m. Radiosondes were released sented on them refer to the time when at this station daily. Other stations the craft is supposed to reach the released them sporadically. Of 13 locality instead if being synchronous. soundings made at Cape Shmidt the Cloud heights, fronts and their anti- highest reached 21,400 m. Kites were cipated movement are also indicated used at 5 stations, 20 ascents being and notations made giving additional made at Uellen and 19 at Cape Zhela- advice. It is claimed that it takes less nie (Desire) ; highest altitude reached time to prepare a charted forecast was 200 m. Cape Shmidt weather than a worded one and that they are bureau office had the best verification easier to handle and more compre- of forecasts for the period (86%). hensive to pilots. _ A forecaster's review of 1939 work In 1939 three forecasters were as- at the Dickson Island office is found signed to the Dickson office.—Alex. in the "Soviet Arctic", published in Breese, U. S. W. B., San Francisco.

Unauthenticated | Downloaded 10/07/21 11:24 AM UTC