APPLIED CLIMATOLOGY 
in some regions only. We will therefore omit further 
discussion of these factors and concentrate on the other 
five factors, all of which relate directly to the wind. 
The energy of the wind varies with the third power 
of the speed. Hence, speed is the most important single 
consideration. The preliminary climatological investi- 
gation must proceed from the available anemometer 
data, the pilot-balloon information, the known relations 
of gradient winds to pressure distribution, and the 
general change of the wind vector with height. In most 
areas this will yield a first approximation on strength 
and constancy of winds at possible power sites. The 
lack of direct meteorological observations can often be 
supplemented by ecological studies in the area. Strong, 
steady winds will produce deformations in trees, which 
act as suitable mtegrators of the wind vector. 
Factor 2 mentioned above will, after preliminary 
selection of possible sites, have to be supplied by wind- 
tunnel tests on models of the particular terrain. The 
local structure of the wind, such as gustiness, cannot be 
predicted with a great deal of confidence. It depends on 
local topography and roughness. A series of observations 
at various heights in the low layers above the ground 
will be necessary. Such special observations do not have 
to extend over a long period of time. Sets of measure- 
ments under a variety of conditions of stability can give 
enough information. The data thus collected are not 
only of value for the mechanical design of the wind 
turbine and its support, but are also essential for the 
choice of the best height of the windmill above the 
ground at the site. 
The frequency distribution of hourly wind speeds 
must be known. The New England study has revealed 
several points that are universally useful. The speed- 
frequency curves are Pearson Type III curves with 
sharp modes, skewed toward lower speeds. Such a curve 
can be approximated quickly from a number of fixes. 
The following relations hold (at least for the New 
England area): 
1. A small percentage of hours are calm. 
2. The sides of the frequency curve are concave 
upward. 
3. The skewness increases with the mean speed. 
4. The most-frequent speed is lower than the mean 
speed, but increases with higher mean speeds. 
5. The number of hours during which the wind 
blows at the most-frequent speed decreases as the 
mean speed increases. 
Similar general relationships can undoubtedly be de- 
veloped for other environments. 
The mean velocity at a possible site can be established 
from a short record by reduction to a nearby long- 
record station, since the ratios of simultaneous values 
stay essentially constant. The study also established the 
requirements for the length of simultaneous records 
needed in order to establish this ratio of mean speeds 
within specified limits of error. For a 10 per cent error 
limitation, the lengths of record listed in Table VI 
were found. 
Of further importance are the periodic variations 
985 
(seasonal and diurnal changes) and the reliability of 
wind power (maximum positive and negative departures 
from the mean). These will require long records, but a 
nearby station (within about a fifty-mile radius and 
with not too different an exposure) can be used for this 
purpose. This means in essence that short-record sta- 
Tasie VI. Lenera or Recorp NEEDED TO RepucE SHoRT- 
Record Station To LonG-REcorD Station 
Horizontal distance Difference in elevation Length of time 
(mi) (ft) (days) 
38 +2000 90 
12 +2000, 30 
10 +400 10 
Same locality +65 0.6 
tions, operating for about three months, will yield the 
most essential basic data for a climatological analysis 
of this type of case. The low-layer vertical distribution 
at the auxiliary stations should also be established by 
anemometers at three or four levels between the surface 
and two hundred feet elevation. If possible, the three 
months of records should not be continuous but should 
be composed of several periods of two or three weeks 
each, spread over the seasons. 
Single Point—Complex Time Relation—Multiple Cli- 
matic Element 
Instead of establishing classes, the climatic influences 
can often be expressed by a formula. This scheme has 
been employed for determining deterioration of 
materials under weather influences. According to C. 
E. P. Brooks [14], the following elements enter into the 
problem: 
1. Effects due to high temperatures alone. 
a. Mechanical effects. 
b. Speed-up of chemical reactions by heat. 
c. Flowing due to decreased viscosity. 
d. Increased activity of bacteria. 
2. Effects due to high temperature in conjunction 
~ with moisture. 
a. Corrosion of exposed surfaces. 
b. Effect of diurnal eycle—breathing of partly 
sealed packages. 
c. Organic changes—molds and rotting. 
3. Effects due to impurities in the air. 
The combined effects of temperature, humidity, at- 
mospheric impurities, and wind can be stated in a 
general equation of the form 
p 29) 
a K) (1.054) + c1)(1 + 0.0670), 
A = art + 
where A is the rate of deterioration of a fresh sample, 
i the temperature in degrees centigrade at the surface 
of deterioration, H the relative humidity at the tem- 
perature t, J the concentration of effective impurities, 
bacteria, fungi, or spores, and W the effective wind 
speed in mph. The quantities a, x, b, c, K are constants 
which have to be determined in each case. The complex 
