SHORT-RANGE WEATHER FORECASTING 
understood and the forecasting of this phenomenon is 
still unsatisfactory. Verification of thunderstorms even 
in the relatively short-range airway forecasts (12 hr) is 
occasionally less than 20 per cent. 
This summertime type of precipitation is of major 
economic consequence over many of the most important 
agricultural areas in the entire world and yet there is 
no comprehensive and satisfactory description of any 
reliable technique for forecasting it. Quantitative pre- 
cipitation forecasts are needed by river engineers, hy- 
droelectric power companies, and many other interests 
in commerce and agriculture. A rather primitive pro- 
cedure for a forecast of this type has been devised by 
Showalter [50] based on the amount of moisture which 
will enter a given region and the unprecipitated amount 
likely to leave it. Verification is unsatisfactory since no 
precise methods exist for a rapid quantitative calcula- 
tion of effective convergence over specific areas. The 
convergence of flow over mountain ranges is more 
susceptible to computation. 
Klein [31] has correlated precipitation anomalies and 
the five-day mean 700-mb chart with considerable suc- 
cess but verification dropped from about 66 per cent 
on past weather charts to 25 per cent when his tech- 
niques were used on prognostic 700-mb charts. Although 
the technique was derived from five-day mean charts, it 
appears equally applicable to daily forecasts. 
Precipitation results from the lifting of saturated air 
and is therefore associated with horizontal mass con- 
vergence in lower levels and horizontal mass divergence 
at higher levels, of which some of the associated factors 
are: 
1. Frontal lifting, upslope or upglide. 
2. Surface friction. 
3. Topography. 
4. Thermal instability. 
V.J.and M. B. Oliver [40] have collected a number of 
principles of practical use in forecasting, which include: 
1. Cloudiness and precipitation are prevalent under 
eyclonically curved isobars aloft (about 10,000 ft), 
regardless of the presence or absence of fronts on the 
surface chart. 
2. In a cold air mass, the instability showers, cumu- 
lus, and stratocumulus clouds will be found only in 
that portion of the air moving in a cyclonically curved 
path. 
3. In a warm air mass moving with a component 
from the south, cloudiness and precipitation will be 
very abundant under a current turning cyclonically or 
even moving in a straight line. (The importance of tra- 
jectories with increasing cyclonic curvature in producing 
convergence and heavy precipitation is stressed by all 
writers on this subject.) 
4. Hlongated V-shaped troughs will have cloudiness 
and precipitation in the southerly current in advance 
of the trough with clearing at the trough line and be- 
hind it. 
5. Cold fronts will produce no weather when the 
component of the wind perpendicular to the front 
increases with height through the front. 
6. If the 10,000-ft flow is parallel to the cold front, 
761 
the front will be active and cloudiness and precipitation 
will extend as far behind the front as this condition 
obtains. 
The applicability of all these rules is subject to the 
normal variations of moisture and stability. 
Fog. The physics of fog formation have been dis- 
cussed by Petterssen [43]. Fog is the subject of a sep- 
arate article by J. J. George in this Compendium.! 
Techniques developed during the past two decades by 
George [25], based largely on analogues and statistics, 
provide the most practical methods of forecasting fog 
in middle latitudes. 
Temperature. The main factors governing local tem- 
perature change in the free atmosphere are horizontal 
advection and vertical motion. The mimimum tempera- 
ture may be predicted by determining the location of 
the air expected over the station the following night 
and forecasting any indicated modification. The dew 
point of the expected air mass is a helpful indicator. 
Consideration must be given to modifications resulting 
from emergence of air from snow-covered to bare ground, 
trajectories over water, and such factors as wind and 
clouds. With clear skies and calms, the dew point of an 
air mass falls somewhat during the night as moisture 
is extracted from the air through the formation of dew. 
The maximum temperature may be predicted from 
the sounding in the air mass expected to be over the 
station during the day. Depending upon the season, 
the lowest 800-1400 m will be warmed by heat trans- 
ferred to the air from surface insolation. Although more 
refined methods are available, a satisfactory approx1- 
mation of the maximum temperature may be obtained 
by following the dry adiabatic from a point on the 
sounding at the top of the area expected to be warmed 
down to the surface and picking off the temperature. 
Corrections must be applied for any indicated cloudi- 
ness and precipitation. In winter, diurnal imsolation 
may have difficulty in overcoming even very shallow 
cold air inversions. 
Wind. The gradient wind may be determined from 
the prognostic surface pressure pattern. The proportion 
of the gradient wind actually realized on the surface 
varies from 40 to 80 per cent or more over water with 
the variation dependent upon a number of factors, of 
which the stability of the air is most important. Over 
land areas the frictional effect is greater. 
Synoptic situations, mostly in spring, favorable to 
tornado formation may be recognized by the forecaster, 
but any precise definition of the areas where tornadoes 
may occur has, so far, been impracticable. Recently 
E. J. Fawbush and R. C. Miller, in an unpublished 
paper,® have claimed some progress in defining the 
geographical areas in Oklahoma where tornadoes may 
occur by locating the intersection of areas of greatest 
instability, increasing dew point, narrow band of high 
4, Consult ‘‘Fog”’ by J. J. George, pp. 1179-1189. 
5. (Added in press.) These results have now been published 
in an article by Fawbush, E. J., Miller, R. C., and Starrett, 
L. G., ‘An Empirical Method of Forecasting Tornado Develop- 
ment.’’ Bull. Amer. meteor. Soc., 32: 1-9 (1951). 
