1050 
exists—leakage through gates and valves and seepage 
through the banks accounting for losses of the same 
order of magnitude as evaporation. In addition, the 
accurate measurement of all inflow to the larger reser- 
voirs is exceedingly difficult. 
In the spring of 1950 the U. S. Navy, Geological 
Survey, Bureau of Reclamation, and Weather Bureau 
jointly undertook a study of evaporation at Lake Hef- 
ner near Oklahoma City, Oklahoma. This reservoir, 
selected after a nationwide survey, is considered to be 
most nearly the ideal experimental site available. In- 
struments have been installed with a view to testing 
the turbulent transfer and energy balance concepts 
and to refine the relationship between evaporation 
pans and lake evaporation. This study should prove 
to be one of the most important steps toward the 
solution of the evaporation problem undertaken in 
recent years. It is to be noted, however, that almost 
any relationship developed will involve an empirical 
coefficient which must be selected largely on the basis 
of judgment before the method can be applied to any 
other lake or reservoir. 
Snow. The snow pack which accumulates each winter 
in the mountains of the western United States con- 
stitutes a reservoir of such proportions as to dwarf 
all man-made lakes in this country. With the coming 
of the spring thaws, this water is delivered to the 
valleys when it is most urgently needed for irrigation. 
Occasionally, under the influence of extremely abnormal 
meteorological conditions, the melt water is released 
so rapidly that damaging floods result. It is natural 
therefore that the problem of ice and snow should 
receive considerable attention from the hydrologist. 
The problem of snow melt is not unlike that of 
evaporation. Melting of snow is the result of heat 
brought to it from several sources—conduction and 
convection from air, radiation, condensation of water 
vapor, rainfall, and the soil. Continued research into 
the mechanics of turbulent transfer will materially aid 
studies of snow melt. However, the snow-melt problem 
involves extensive areas, usually with a wide range of 
elevation and rugged topography. It is doubtful, there- 
fore, that a theoretical approach will find practical 
application in most hydrologic work. It is to be hoped 
that theoretical studies will suggest a more adequate 
empirical approach te the job. 
Design Problems. Another phase of hydrologic work 
in which meteorology can play an important role is 
the field of design. Every structure, large or small, 
designed to control the flow of water, be it a culvert on a 
secondary road or a major reservoir, must be planned 
to withstand some maximum flow of water known as 
the design flood. For the smaller structures, economics 
dictates that the design flood be the probable maximum 
flood to be expected during the estimated useful life 
of the structure. For larger structures, where failure 
may take a large toll of life or cause damage far beyond 
the value of the structure itself, the design may be 
based on the maximum flood which can possibly be 
expected to occur. Cost normally increases with the 
magnitude of the design flood. Hence, overdesign is 
HYDROMETEOROLOGY 
costly in terms of wasted material and labor, while 
underdesign may be equally costly as a result of failure 
and resulting replacement. Any knowledge which helps 
to refine estimates of design-flood magnitude is of 
definite economic value [1]. 
Generally speaking, the design of smaller structures 
is based on frequency analysis, and for structures hav- 
ing a useful life of thirty years or less the design criteria 
may be assumed to be reasonably adequate. For larger 
structures the situation is not so favorable. Reliable 
records of stream flow in excess of thirty years are 
rare. The situation is somewhat better as far as pre- 
cipitation records are concerned but, even for these 
data, records in excess of fifty years are scarce. Some 
writers have proposed the so-called station-year tech- 
nique [3] in which records from a number of stations 
within a homogeneous area are assumed to be inde- 
pendent and can thus be combined to give a record 
equivalent in length to the total length of the individual 
station records. Even if the validity of this approach 
were beyond question, it fails to satisfy the design 
problems for intermediate and larger projects whose 
drainage areas are so large that single-station records 
cannot be considered representative. A study of the 
frequency of occurrence of various average depths of 
rainfall over areas from 100 to 10,000 square miles 
would be extremely useful. Because of the lack of 
processed storm data, such a study would involve an 
immense amount of work. A study of storm morphology 
with particular reference to the relation between maxi- 
mum point and average depths would also be of con- 
siderable value. 
Another problem often encountered involves joint 
frequency, that is, the question, What is the frequency 
of rainfalls of various magnitudes with snow cover on 
the ground? or, What is the probability of a small, 
intense storm which would overload the dramage works 
of a leveed area at the same time that a major flood 
is occurring as the result of general heavy rains several 
days earlier? [16] Probably these problems cannot be 
solved by either meteorologists or hydrologists until 
much longer records exist. However, the considered 
judgment of the meteorologist can be of material aid 
to a hydrologist forced to make a decision and adopt a 
design value, no matter how poor its basis. 
The situation reaches an extreme in the case of those 
major structures whose failure cannot be permitted 
under any condition. Here it is necessary to adopt the 
maximum possible flood as a design criterion. Since 
records are far too short to define such a physical ex- 
treme, it is necessary to synthesize it on a theoretical 
basis. It is therefore logical to turn to the meteorologist 
for an expression of the magnitude of the maximum 
possible storm for the project basin. The jomt venture 
deriving from this need—hydrometeorology—is dis- 
cussed elsewhere in this Compendium.1 
It would be an easy matter for the hydrometeorologist 
to protect himself against an underestimate by using 
1. Consult ‘‘Hydrometeorology in the United States” by 
R. D. Fletcher, pp. 1033-1047. 
