Handbook of Meteorology/The Measurement of Precipitation
THE MEASUREMENT OF PRECIPITATION: RAIN
GAUGES: SNOW MEASUREMENT
All moisture condensed from the air—rain, snow or hail—is classed as “precipitation” and is measured in terms of rain. It represents the depth of water which would accumulate on a level surface without loss by run-off, percolation or evaporation. For convenience in measurement the water of precipitation—rain, or melted snow—is caught in vessels of special construction called rain gauges. These are of various forms, but they have practically the same principle—the exhibition of a depth of rain expressed in inches and hundredths.
The standard Weather Bureau rain gauge is a cylindrical barrel of galvanized iron 26 inches in height over all. The receiver is a funnel with an upright collar of bronze 2 inches high. The edge of the collar is beveled so as to present a sharp cut-water to the rain. The receiver delivers the water to the measuring tube; a sleeve at the lower end of the funnel holds the mouth of the measuring tube in place and prevents the loss of water. The brass measuring tube, accurately calibrated, rests in a seat at the bottom of the barrel.
The receiver is exactly 8 inches in diameter; the section of the measuring tube is exactly one-tenth the area of the receiver. Its diameter is 2.53 inches and it is 20 inches high. An inch of rain therefore measures 10 inches in the tube, and the latter holds 2 inches of rain. Any excess beyond 2 inches overflows into the larger vessel; it is poured into the emptied measuring tube and added to the amount. The measuring stick is graduated to measure inches, tenths, and hundredths, and the depth of the rainfall is indicated by the length wetted when the stick is inserted in the measuring tube. Care must be used to keep the surface of the measuring stick free from grease. If it fails to show the wet-line clearly the surface may be cleaned with alcohol, or—better—rubbed clean with 00 sandpaper.
Recording and Registering Gauges.—The registering and recording gauges are mainly of two classes—float-gauges, in which the increasing depth of water, by lifting a float balanced by a weight translates motion to a pen arm; and the tipping-bucket gauges, in which each tip of a full bucket moves an index hand.
 | ||
| FRONT VIEW | VERTICAL SECTION | HORIZONTAL SECTION, E.F. |
Standard Weather Bureau rain gauge: A, receiver; B, barrel; C, measuring tube.
The Marvin float gauge, used at many stations, is provided with a wind shield of the Nipher type, about 21 inches square. The drum carries a sheet ruled with lines nearly horizontal, but inclined so that they form a continuous spiral. These lines carry the record, one for each day of the week. Vertical lines divide the sheet into ten-minute spaces. The drum, driven by clockwork, makes one revolution in twenty-four hours. A screw thread of the required pitch causes the recording pen to follow the spaces between the spiral lines on the record sheet.
When rain begins to gather in the measuring-tube, the lifting of the float causes the rotation of the cam shaft and this imparts a lateral motion to the pen. The graph made by the pen consistes of a sinuous line which curves back and forth across the day line. A complete revolution of the cam records half an inch of rain. The faster the precipitation, the sharper are the curves.
The Marvin gauge possesses several distinct advantages. Recording begins within a very few minutes after precipitation has commenced—a merit not possessed by tipping-bucket gauges; it likewise records the cessation of rainfall rather more promptly. A very desirable feature is the fact that it also records the rate of rainfall, a matter of great importance. In many instances the total of precipitation is of minor consequence, while the rate per unit of time must determine the discharging capacity of sewers and other run-off systems. The Marvin gauge is not fool-proof, but this detail applies also to other recording instruments.
The tipping-bucket gauge is chiefly used for recording rainfall. The drip from the funnel falls into one or the other of two scoop-shaped buckets placed back to back mounted on trunnion bearings. When 0.01 inch of rain has collected in a bucket the weight causes it to tip, spilling the water into a container and moving a pointer one division on a dial. The tipping of the full bucket swings the empty bucket into a position where it catches the drip.
Where a Friez triple recorder is used the pen which ordinarily records sunshine is also used to record rainfall. This it does with little or no confusion of records, because precipitation rarely occurs in appreciable amount while the sun is shining.
Tipping-bucket rain gauges are constructed so close to exactness of measurement that, when placed side by side with the standard gauges, the difference between the measurements of the two is not much greater than that of two gauges of the same type side by side.
Although the tipping-bucket rain gauges are simple in construction, various conditions may occur that result in erroneous recording. The bucket may rebound on emptying itself, in which case two registrations instead of one are made. With the Friez gauge this will appear as a mark of double length on the record. With the dial gauges, which register but do not record graphically, double registration cannot be discovered except by close watching. It may be suspected, when the catchment of the registering gauge runs uniformly greater than that of the standard Weather Bureau gauge. In the case of the Friez gauge a readjustment of the stop pins is necessary. With the Short and Mason gauge a pressure brake is provided; the tightening of this will prevent back-tipping.
Sometimes it happens that the stick measurement of the catch is considerably in excess of that registered on the dial or recorded on the paper; indeed, this is pretty apt to be the case in heavy summer showers. Usually this discrepancy is due to the fact that a measurable interval of time elapses while the bucket is discharging its load of water, after it has been filled. The water that flows into the bucket during the interval is therefore slightly in excess of the normal 0.01 inch. During a very heavy shower, aggregating 3 inches, the excess of the measured amount over the recorded amount may be as great as 0.15 inch. In such a case stick measurement rather than bucket measurement should be taken as the total precipitation.
The drip aperture of the Short and Mason registering gauge is very small; although protected against clogging by falling leaves, it may become clogged with dust. Under such circumstances the funnel may fill and overflow with no water running into the buckets. If the gauge receives even ordinary care such a condition is not likely to occur. Cleaning the receiver daily is not absolutely essential, but a conscientious observer will see that the gauge is always in order.
The inside of a gauge is a spot most dear to the heart of the spider, and in many a case the accumulation of spider web has tied up the registering mechanism so completely that registration ceased.
Another source of annoyance in registering and recording tipping-bucket gauges has to do with the condition of the buckets. In regions where the air is very dirty, sediment may cling to the surface of the bucket, and, adding to its weight unequally on opposite sides, prevent a true registration. A still greater source of error may result from handling the inner surface of the buckets with greasy fingers. The water will not cohere to, or “wet” a greasy surface; and this may cause a slight but persistent error in registration.
It is not wise to depend wholly upon a registering or a recording gauge; stick measurement is more certain. Nevertheless, a station of any sort should be provided with two gauges, and a registering gauge is a most excellent feature of equipment. H. J. Green makes an indoor dial that may be attached readily to any tilting-bucket gauge. Such a device is very convenient.
Intensity of Precipitation.—The intensity of rainfall may be of greater importance than the gross amount. The Marvin and the Friez gauges record intensity graphically. The observer with the ordinary gauge can find the intensity in one way only—by making measurements at regular intervals. During ordinary rainstorms measurements made at half-hour intervals will suffice, and these need be continued for not more than two hours. During heavy summer downpours, however, the measurements should be made at five-minute intervals.
To the farmer, 2 inches of rain distributed over the greater part of the day means a thorough soaking of the ground; but if concentrated within half an hour it means beaten-down grain and washed-out ditches. Such a rainfall, to the engineer, means washouts all along the track; to the city engineer it means flooded sewers and excavations. The engineer who takes care of drainage must know how to guard against phenomenal rainfalls by building so as to take care of them.
Observers will make their work more helpful by noting not only the fact of excessive rainfall but also its rate at five-minute intervals. In Weather Bureau practise, the term excessive has a technical application, and the tabulation of excessive amounts during such intervals is required. The following table shows the intensity of precipitation that technically is excessive. It is based on the experience of many years.
| [1]Duration in minutes |
Duration in inches |
Duration in minutes |
Duration in inches |
| 5 | 0.25 | 35 | 0.55 |
| 10 | 0.30 | 40 | 0.60 |
| 15 | 0.35 | 45 | 0.65 |
| 20 | 0.40 | 50 | 0.70 |
| 25 | 0.45 | 60 | 0.80 |
| 30 | 0.50 | ·· | ···· |
When excessive rainfall extends beyond a duration of two hours the measurements are recorded at fifty-minute intervals.
The amount of rainfall necessary to insure a specific crop varies with locality and with the character of the crop. More especially it depends on the distribution of the rainfall over the growing season. Roughly, rain must fall during a period which covers three-quarters or more of the growing season for the particular crop. The growing season for wheat is over, in most localities, by the middle of July—in some localities by the middle of June. The growing season for corn extends into September. A rainfall of 12 inches, fortunately distributed, may be all that is required for a specific crop. Unfortunately distributed, a fall two or three times as great may not suffice.
From the nature of the case, the knowledge which concerns crop safety must be gathered locally. Through its Climatological Service, the Weather Bureau is gathering knowledge of this sort, but additional information is very desirable. The observer, whether official or volunteer, can aid in gathering useful information along the following lines:
The length of the growing season—that is, the number of days between late spring frosts and early fall frosts.
The months during which rain is necessary for each specific crop.
The duration of droughts—that is, the number of days during which no rain or only a trace of rain falls—that are hurtful or destructive to specific crops.
The character of soil with respect to rainfall necessary to crop growth.
As a rule the precipitation records of the nearest Weather Bureau station—regular or cooperative—will furnish the necessary information concerning the amount of precipitation. The specific locality sometimes requires its own rain measurements. A rain gauge of the Weather Bureau pattern is useful, but a metal container with straight sides will answer fairly well, and an inch rule will answer the purpose of a measuring stick. The volunteer observer who studies the rainfall of a locality may thus gain the essential information required; namely, the minimum amount of rain, and also the optimum rainfall both as to amount and distribution, for a specific crop.
Mean annual rainfall of the United States: the figures expressed in inches of precipitation.
The Location of the Rain Gauge.—In establishing any sort of a station where the measurement of rainfall is to be recorded, at least two rain gauges are desirable. One of these should be a standard Weather Bureau gauge or one of similar pattern; the second may be any vessel with an 8-inch circular opening in the cover.
In cities which are solidly built the flat roof of a building offers about the only suitable place for a rain gauge. If the edge of the roof is a parapet, so much the better, for the drive of the wind is less apt to blow aside the rain that should fall into the receiver. In a position of this sort the catchment of the two gauges is not likely to differ materially.
In suburban localities and in communities where buildings are 100 feet apart the gauges are better placed in such positions as have the full sweep of the rain-bearing winds. If two places 100 feet or more apart show no material difference in the catch, either location is probably suitable. With gentle rain and still air the two gauges should be in close agreement; if the wind blows in strong gusts there may be a material difference.
The wind is the chief obstacle to accuracy of rainfall measurement and shielding the gauge from the full strength of the wind is the best means to insure an accurate catch. The Nipher shield is a trumpet-shaped metal device about 20 inches across which surrounds the mouth of the gauge. It is surmounted by a rim of copper mesh which prevents insplashing. J. O. Alter, Observer at Salt Lake City, constructed a much simpler shield by fastening a strip of canvas about 9 inches wide, to a metal ring about 30 inches diameter. The screen thus constructed is suspended about the gauge by metal struts. The edge is about 2 inches higher than the mouth of the gauge. The Weather Bureau regards this shield with favor. The author has found a similar shield made of copper mesh, such as is used in window screens, a most excellent device. In the long run, a shielded gauge will catch from 6 per cent to 10 per cent more rain than one unshielded, according to the experience of the Weather Bureau. P. R. Jameson, with measurements covering many years, finds a gain of about 9 per cent in the case of shielded gauges.
The pit-gauge is favorably considered by C. F. Marvin, an authority on precipitation. The pit-gauge is merely a depression in which a standard Weather Bureau gauge is placed so that its mouth is 10 or 12 inches higher than ground surface. It is surrounded by a rim of earth in the form of a ring about 6 feet in diameter. A pit and ring of concrete with a movable cover of wire mesh, coarse enough to permit rain to enter without obstruction and fine enough to keep leaves out makes an ideal position for a rain gauge in an open and fairly level country.
The Ferguson gauge designed for isolated stations by S. P. Ferguson, of the United States Weather Bureau, totalizes a year’s rainfall month by month, or in such measured proportions as may be desired. A film of oil in each of the thirteen receivers prevents loss by evaporation.
The Measurement of Snow.—A reasonably accurate measurement of snowfall is desirable in regions of plentiful rainfall; it is imperative in regions where the irrigation of crops or a knowledge of possible floods, or of droughts is a prerequisite.
In moderately level regions of gentle drainage the measurement of the precipitation derived from snow requires the care and judgment that comes only with experience. In mountainous regions it requires judgment, patience and a lot of hard work.
On the prairies of Indiana, for instance, the measurements made at the Weather Bureau stations give a pretty accurate total of precipitation. If the aggregate error amounted to 12 inches of snow, or even 2 inches of rain, however, the result would not be materially harmful. In California, however, the floods of the Sacramento and San Joaquin River valleys are largely predetermined by the snowfall in the mountain slopes to the eastward; and in many arid regions the crop production which depends on irrigation must be foretold mainly by the total of snowfall.
In level regions where the snow is not disturbed by the wind the measurement is not difficult. The observer uses his measuring stick in a dozen or more places within a radius of 300 feet. Usually the mean depth will become apparent without computation.
To find the equivalent in terms of rainfall requires “puttering and patience.” A very convenient way is to cut a section in the snow with the inverted barrel of the standard rain gauge, thrusting a dust pan or a piece of sheet iron under the mouth in order to hold the section firmly. It is advisable to cut at least three sections. The melted snow may then be measured in the tube, taking one-third of the total. Melting the snow may be expedited by pouring into the barrel containing the 
The Marvin shielded seasonal snow-gauge. snow a measuring tube exactly full of hot water, thereby reducing the snow to a condition sufficiently liquid to be measured. Two inches must be deducted for the water added; one-third of the remainder is the depth of equivalent rainfall.
In mountain regions where the depth of a single fall may be several feet, such a method of reduction is out of the question. Several convenient expedients are employed. A gauge 40 inches high with an interior diameter of 10 inches provided with a Nipher shield, is used at the station where not less than two observations a day are made. The accumulation of snow is weighed from time to time on a spring balance, the dial of which reads hundredths of an inch instead of ounces. A mechanical device lifts the receiver from its support so that it can be readily removed to the swinging arm that carries the balance. This is about the most expeditious method of measuring snowfall yet devised.
In the western slope of the Sierra Nevada Mountains “seasonal gauges,” with collectors large enough to hold the accumulation of snow and rain for several weeks, are employed. The catch is weighed at convenient times.
When deep snowfalls occur the Marvin snow tube has been found a most convenient device. This tube, as improved by Church, is made of galvanized iron and is 2.75 inches in diameter. The upper end is left open; the lower end is reinforced by a piece of tubing forced inside the measuring tube. The lower edge of the tube is serrated with teeth like those of a cross-cut saw in order to facilitate boring through crusted snow and sheets of ice.
Tube and core are weighed by a spring balance that records inches and hundredths. The tube has also an engraved scale to show the depth of snow. Church, working in the Sierra Nevada ranges, used the tube in snow accumulations 30 feet thick.
The Marvin density bucket provides a quick and accurate method of obtaining the rain equivalent. The copper bucket is inverted and pressed lightly into the snow until the top of the snow touches the bottom of the bucket. The bucket, even full of snow, is then weighed on the accompanying scales, which are graduated to per cent of weight of an equal volume of water. Thus, if the net weight—the total weight minus that of the bucket—is 0.12 on the scales, it means that the snow is 12 per cent of an equal volume of water. Assuming that the depth of snow is 20 inches, 0.12 × 20, or 2.40 inches is the equivalent depth of rainfall.
In mountain regions the intensity of precipitation varies with altitude. On the western slopes of the high cordillera of the Pacific coast, McAdie found the greatest intensity of precipitation between 4000 and 5000 feet above the valley floor. On the eastern slope the intensity decreased irregularly with decreasing altitude. It may be incautious to assume this to be true elsewhere, but; as a general truth, the basis of assumption is not unreasonable. It is safe to assume that the measurement of precipitation of the montane part of a watershed must extend from its upper limit to the valley floor.
A multiplicity of snow gauges is not required, but with the combined results of gauges, snow tubes and fixed measuring posts, a fair approximate of the catchment of the basin may be obtained.
In practically all localities where snowfall requires measurement the following difficulties confront the observer: mixed or alternating snow and rain; rapid melting of snow while it is falling; a very light fall, say, less than half an inch; rapidly drifting snow. In the case of the first three, the observer may remove the receiver and tube from the gauge and catch the precipitation in the barrel. In the last case about the only way to overcome the difficulty is to make a considerable number of measurements where no drifting is apparent. One must use care to avoid measuring old snow with a fresh fall.
- ↑ At Porto Bello, Panama, 2.48 inches were reported during a period of five minutes, Nov. 29, 1911, at 2:07 a.m. At Curtea de Arges, Rumania, 8.07 inches fell in twenty minutes, July 7, 1889. These are the heaviest rainfalls of record, but they may have been exceeded by cloudbursts in which measurements were not made.