874 
described, the organized lines of cumulonimbus cloud, 
accompanied by appropriate wind changes, which they 
identified as fronts. The existence of the lines has been 
confirmed, but it now appears that the frontal explana- 
tion must be rejected. The following empirical facts may 
be regarded as established: 
1. If an adequate streamline analysis of the lower 
wind field in which the cumulonimbus line is embedded 
is carried out, using the isogon method of V. Bjerknes 
and his collaborators [9], the line of cumulonimbus 
coincides in position with an asymptote of convergence 
in the streamline field. But not all asymptotes in the 
TROPICAL METEOROLOGY 
mogeneity. Orographic lines are rarely more than 200 
miles long. 
3. Except on such mountainous islands, the lines 
rarely show temperature contrast in the lowest layers 
but may do so in the upper air, as a reflection of dif- 
ferent lapse rates in the air on either side of the line. 
The contrast may intensify, vanish, or reverse its sign 
both im space and in time but without, as far as can be 
discovered, affecting the location, movement, or inten- 
sity of the cloud line. 
4. By far the greatest number of the lines in the 
equatorial Pacific, and in the Caribbean in summer, 
PILOT DE-BRIEFING 
re, 0745 OBS: ENTERED FRONT 
Zw LIME OF hEAWY CY 1 Siz 
tw 
ou 0815 OBS: PRECIPITATION BEGAN 30 MINUTES BEFORE 
Ow REACHING POSITION 
=o LINE OF HEAVY CU TO SE 
0845 OBS: PRECIPITATION ENDED 10 MINUTES BEFORE 
KWAJ. 1330 1300 1230 REACHING POSITION 
LINE OF HEAVY Cu TO NW 
30 z WIND ANALYSIS 
N 1500 FT. 11008 
MAY 26,1946 
25 
wy 
ir} 20) 
iL 
6 15 
(2) 71030" 
2 10F 
<q 
7m) 
=) 
oS § 
ae 
[= Y 
Sil THA SAbS 
TIME 0915 0815 O715 KWAg. 
WEATHER oo he 
TEMP. 24 24 24 21 21 24 
RH 88 80 72 96 96 86 160° EAST 165° 170° 175° 
Fie. 1.—Streamline and isovel analysis at 1500 ft of an equatorial wave passing the Marshall Islands, 1100 Bikini Time, May 
26, 1946, with two cloud sections observed by weather reconnaissance aircraft. Speed in knots. (Reprinted from the 7th quarterly 
report of the Tropical Pacific Project, U.C.L.A., a research supported by funds provided by the Geophysical Research Directorate, 
Air Materiel Command.) 
field are accompanied by a “front”; there is an organ- 
ized cloud system only if, at the same time, the total 
horizontal velocity divergence is negative in the neigh- 
borhood of the line. Aloft, say at 20,000 ft, the cloud 
line may correspond to an asymptote of positive diver- 
gence in the horizontal wind field. We shall henceforth 
call organized lines of cumulonimbus (or tall cumulus, 
on occasion) convergence lines, not fronts. We ought to 
point out that regions in which there is negative hori- 
zontal velocity divergence in the lowest layers are not 
always the seat of lines of convergence; these occur 
only if there is also convergence in the streamline along 
an asymptote. Figures 1 and 2 illustrate some of the 
properties of convergence lines in the Marshall Islands. 
2. Lines of convergence form readily as a result of 
diurnal and orographic disturbances of the wind field 
in the neighborhood of large mountainous islands. We 
may then get temperature contrasts across the line, 
but these contrasts are confined to the lowest layers of 
the atmosphere. Aloft there is usually horizontal ho- 
show no temperature contrast whatever below 20,000 
ft. Above this level, temperature gradients across the 
line may have either sign or be absent. 
5. In low latitudes, the vertical component of the 
vorticity along the line may be positive, negative, or 
zero. Cyclonic vorticity is not a necessary condition for 
the existence of the line. 
6. The convergence line may or may not be accom- 
panied by a trough in the pressure field. 
7. Convergence lines may develop within an air mass 
which is, within the limits of observational error, com- 
pletely homogeneous. 
8. Convergence lines do not move with the wind, 
nor are pressure changes useful in forecasting the move- 
ment. 
We conclude: Density contrasts and cyclonic shear 
are conditions which are neither necessary nor suf- 
ficient for the formation of convergence lines and hence 
these lines are not fronts in the sense that that concept 
is employed in the frontal and air-mass theory. The 
