|
<br />only by northerly winds, which are relatively dry.
<br />The moist westerly and southwesterly winds must
<br />pass over the Coast Range or the Tehachapis.
<br />Thus these ranges determine the effective moisture
<br />barrier for the southern Central Valley.
<br />2.3,17 The elevations indicated on tbe map
<br />of figure 2-8 are the lowest elevations to which
<br />wann, moist air with a trajectory directly from
<br />the source region would rise in reaching any par-
<br />ticular point. In some regions moist air can come
<br />from other source regions but would encounter
<br />higher barriers. The arrows indicate the general
<br />directions of the moisture-bearing flow considered
<br />likely to prevail during major stonns in the various
<br />sections of the West and do not show all directions
<br />from which warm, moist air can reach any region.
<br />2.4 Measurement
<br />2.4,1 In dealing with measured precipitation
<br />rates, the methods and errors of measurement
<br />should be considered. In the United States three
<br />types of gages are used in making official measure-
<br />ments of storm precipitation. These gages are:
<br />(1) the tipping-bucket recording rain gage, (2)
<br />the weighing-type recording gage, and (3) the
<br />standard 8- in. nonrecording ga.ge. The last two
<br />will measure any form of precipitation whereas
<br />the first is limited to rainfall. Only the first two
<br />actually measure intensities; the nonrecording
<br />gage measures amounts only. Detailed descrip-
<br />tions of these gages are available in most textbooks
<br />on meteorology or hydrology and will not be given
<br />here. More important in evaluating the represent-
<br />ativeness of maximum observed intensities, which
<br />are of primary interest in this study, is a knowl-
<br />edge of the gage-network density and the errors
<br />of measurement.
<br />2,4,2 The United States, excluding Alaska
<br />and Hawaii, has an area of approximately
<br />3,000,000 sq. mi. In this area there are about 3,500
<br />recording gages, all but about 200 being of the
<br />weighing type, and about 9,500 nonrecording
<br />gages, or a total of about 13,000 gages. The aver-
<br />age network density computes to be about one ga.ge
<br />per 230 sq. mi., but many stations have two gages
<br />so the average station-network density probably
<br />averages about one station per 250 sq. mi. The
<br />countrywide distribution is not uniform, however,
<br />and the average network density in the West is
<br />appreciably less than that for the country as a
<br />whole. Prior to 1940 the network density in all
<br />parts of the country was a great deal less than it
<br />IS nmv.
<br />
<br />12
<br />
<br />2,4.3 The opening through which precipita-
<br />tion enters the standard gage is rougWy
<br />1/80,000,000 of a square mile in area. If all
<br />13,000 gages were concentrated in one group as
<br />close together as possible, the total catchment area
<br />would be no more than 1/6000 of a square mile-
<br />much less than the area of the standard baseball
<br />diamond!
<br />2.4,4 The recor<fing gages are, of course, the
<br />only ga.ges capable of measuring rainfall intensi-
<br />ties with any degree of accuracy, particularly for
<br />durations under 24 hours. The average network
<br />density of the recording gages alone is slightly
<br />over one per 1,000 sq. mi. Their tota.! catchment
<br />a.rea is about 1,500 sq. ft. This lesser network
<br />density, hence fewer data on short-duration rain-
<br />fall intensities, is the reason for basing the PMP
<br />estimates described in chapter 6 primarily on
<br />24-hr. values.
<br />2.4.5 Obviously, the rainfall rates measured
<br />by the existing network-let alone the much
<br />sparser network prior to 1940-are but a small
<br />sample of those that have occurred throughout the
<br />entire country. The sampling is particularly poor
<br />for local cloudbursts, which are restricted to a few
<br />square miles in area. The chance that the most
<br />intense rainfall in a cloudburst would center over
<br />a ga.ge is extremely remote. The more unifonn
<br />rainfall rates in large-area, or general, storms,
<br />often extending over tens of thousands of square
<br />miles, are naturally much better represented by
<br />the ga.ge sam piing.
<br />2.4.6 Supplementary measurements of rain-
<br />fall in severe storms are obtained through field
<br />surveys, colloquially called "bucket surveys".
<br />These surveys are made by meteorologists and
<br />engineers as soon as possible after the ending of
<br />the stonn. The object of the survey is to gather
<br />data on rainfall that may have been collected in
<br />barrels, pails, bottles, etc. If the exposure of the
<br />container is satisfactory and it can be determined
<br />definitely that the conta.iner was either empty or
<br />held a known depth of liquid before the storm, the
<br />stonn catch is then measured. The measurement
<br />is, of course, adjusted if the container does not
<br />have straight vertical sides.
<br />2.5 Errors of measurement
<br />2.5,1 There are several types of errors in gage
<br />measurements. :\lost of these errors are sma.!l and
<br />negligible, especially in connection with measure-
<br />ment of storm precipitation, and will not be dis-
<br />cussed here. There are three types of errors of
<br />
|