MONTHLY WEATHER REVIEW ALFRED J. HENBY, Edltor.
-
CLoam APRIL 3, 1922
ISSUED MAY 29, 1922 MARCH, 1982. VOE 49, N.o. 3.
W. B. No. 769.
.. .. . _. . .. - - - - . . . . .. - . .. - . . _. .- - . .. . . .. ._ -. . . . . . - .
THE NEW PRECIPITATION SECTION OF THE ATLAS OF AMERICAN AGRICULTURE.
By ROBERT DE i!. WARD.
IHarvard University, March 17,1922.1
Zn,troduction.-The preparation of a section on Climate for the new Atbs of American Aqricultwt! marks IL very important advance in the accurate chartino and dis- cussion of many of the essential features of tee climates of the United States. A s yet, but one part of this section has been issued in its. final form, viz, that on
Fr0st.l The Precipitation section has been rinted but
have, however, been sent to n limited number of teachers and others to whom the material is of immediate prac- tical use. The H,umirEi.ty folio is still to be lithographed and printed, The complete Pre.cipitat.ion and Humidity section will therefore robably not be ready for dis-
Temprutwe section is, unfortunately, still further from Advance ublication has been made of the
all of which have previously been reviewed by the present writer.< General description cf the pPecipitation wtion.-The preci itation section comprises 55 figures, of which 16
are &grams and 69 are maps. Of the maps, 16 show the monthly and seasonal, and one shows the mean m u d precipitation. With these, the present discus- sion does not deal. The text covers 53 pages. Taken altogether, this new series of charts and diagrams con- stitutes one of the most notable contributions to clima- tology of recent years. Not only is the series remark- ably complete, more so than any previous set of similar charts, but, what is even more e.ssential, records for a uniform basic period of 20 years (1895-1914) were em- ployed in the construction of all the princi a1 ma s and
2O-year eriod. Additional shorter records from about
within the 1895-1914 period-vary in length from 10
to 19 ears, and were reduced to the basic %)-year
a more complete but a far more accurate compilation of recipitation data for the United States than has hitherto gee, possible. Especially is this true of the more recently settled portions of the country and of the
not yet bound or distributed. Sets of the P oose sheet3
tribgtion in its h a l P orm, for several month. The
new mean annual, mon &I y, and seasonal rtiinfall maps,s
graphs. About 1,600 stations have recor C f P s for t le full
2,000 ot E er stations were also used. These all come
period t y well-known methods. The result is not only
~
1 Sw revlew b R De C. Ward “Frost in the Unlted States,” Geogr. Reu., vol 7, M f 1919 %it&. Ward: “Some Characterlstics of United States Temperatures,’’ Mbo. WEATHER REV., November 2,1921.49: 595-606: Charts SX.
8 R De C. Ward “Mean Annual Rainfall of the Unlted States with Notes on the New ‘Chart of Average Annual Prwlpltatlon from the ‘Atlas of American Agr&d- turd (Advance Sheet) ’’ ibfd, Jul , 1917, pp. 45: v: chart. I B. Kincar: “The ’Se&onal 8lstrlbutlon 01 Preei itation and Its FTeqnency and Intkdty In the United States ” ibid September l9lB 47: 624-63P Charts SVL.
4 R. De C.Ward “New Month1 add Seasonal hainfall Yaps of h e United States ’’ Gro RLV vol 9 September, I&, pp. 173-181. with the four seasonal charts redradn
k. Dd
,rUfLm.Wbt iimpllnea.
103464-51
higher altitudes of the West. In the case of these last-named districts the mountain sriowfall records were found very useful as a guide in drawing the isohyetal linea. In addition, the amounts of precipitation at the higher elevations (where there is a great lack of actual rainfall records) were inferred from the known increase of precipitation with elevation; from the character of the vegetation, and from the stream flow, wherever reliable data on these conditions were available. Departures -from the average rainfall.-The principal and most immediate object in view in the preparation of this new precipitation folio was to benefit agriculture. From the a ricultural standpoint, the mean annual and mean mont y rainfds are by no means alto ether satisfactory. In the case of the former, the perio is so long that deficiencies in certain parts of the year which are
of critical importance to crops may be who11 concealed by excess precipitation at other times 01 the year. Further, the annual amounts are often of secondary importance to the seasonal distribution. Nor are the monthly charts always of reat value, for they show
in the growth of crops. For these and other reasons, increasing emphasis has lately been laid upon seasonal precipitation charts, which, combining as the do the
regarding the amounts of rainfall available for crops during their critical periods of growth. This new monograph very rightly em hasizes
charts, one each for the four seasons warm season, April-Se tember, inclusive. An especially
emphasis which i t lays upon the de artures which may be expected from the average r a i n f d More and more, as climatological investigation progresses, is it realized that far too much attention has in the past been tncans. It matters little to the farmer to know t at the mean or average rainfall over his section is sufficient for the rowth of a large crop if in some seasons his fields are arc a ed during an (‘unusually” dry time and in others k s crops suffer from (‘excessive” rains. He needs to know what departures from the average he may expect in the run of the years. He is then in a position to decide for himself what crops he may plant with the greatest probability of success. It is ood, therefore, to
graph which shows, for a group of selected stations, the seasonal or monthly precipitation for each of the 20
f 8.
only the avera e conhtions d uring the arbitrary unit of a calendar mont f 1, which unit often has but little importance
conditions of three ‘months, give better in 9 ormation
seasonal precipitation by inc P uding five
noteworthy feature o H the new folio is the marked
!Fd to
see at the bottom of each seasonal an % monthly chart a
117
118 MONTHLY WEATHER REVIEW. MARCH, 1922
ears of the basis period on which the chart is based. Ghese diagrams show the variations which are likely to occur from year to year. They indicate the “relative dependability’’ of the means. Similarly, a full-page set of diagrams (fig. 5 ) shows the annual precipitation a t 56
selected stations, arranged by geographic districts, for each of the ears 1895 to 1914, inclusive. The facts here
from year to year in different sections and the general geogra hic distribution of these variations. It is seen
over the Great Plains, and in the Gulf States. Many other new and economicall important details the variability .of rainfa% are also shown. From a c art of the relative frequency durin
1895-1914 of an annual precipitation less t an 85 per
cent of the aver e (fi 7), it appears that in the vicinity
‘cipitation fell in half the ears, while more than 85
years in parts of the Lake region, of the tlantic Coast States, and of Tennessee. Another chart, of the relative fre uency of warm-season preci itation (April-Septem-
facts of great agricultural im ortance %i. 11). The
average in 8-11 years (of the 20-year period) in southern California and parts of the adjoming States, but, fortu- nately for the great agricultural’interests of the region east of the Rocky Mountains, a warm-season precipitation of less than 75 per cent of the average occurred in only two to four seasons during the 20 years. For each of the eight months July to October the relative frequency of recipitation leas than half the average is
ZE!? (E gs. 58-65). The practical value of these charts may be realized by an examination of the facts which appear on any one of them. Take, e. g., that for July, a cntical month for many of our great staple cro .
Cotton Belt there are a number of localities where only one July with a deficiency of more than one-half of the average rainfall occurred in 30 years. At some points in no year was the Jul rainfall less than half the average
few deficiencies of 50 per cent. n the other hand, the ercentages of frequency of a July reci itation less than galf the average are 1ar e on the 8 a c d slope, espcially in California. As a wiole, the variations above and below the normal are more frequent west of the Rocky Mountains, especial1 in summer. Rain all t y p e s .d e question ‘( WAen does the rain fall!” leing, for the apculturist, often of more impor- tance than the question ((HOW much rain falls?” it is natural and lo ical that considerable attention should be
tribution through the year. The discussion of this mat- ter occupies somewhat over a page of the text (pp. 16
and 37). In place of the 11 ty es of seasonal distribu- tion recognized by Prof. A. J. genry, 6 types are here adopted as sufficient “with res ect to their agricultural significance and areas covereg.J’ * These are named Pacific, Sub-pacific, Arizona, Plains, Eastern, ‘and Flor- ida and are illustrated by a series of percentage graphs
in their roper locations on a general map of the fared- Tnitcd for selected stations, the separate graphs being
permanent or might in some respects be altered if R
resented s I ow the variations which may be expected
that t 1 e larger variations occur on the Pacific slope,
sthe period
of Yuma, Ariz., P k ess t an 85 per cent of the annual pre-
per cent of the annual mean 3 ell in over 90 er cent of the
be3 less than 75 per cent of t % e aver e brings out
warm-season rainfall was less t !i an three-fourths of the
5;
It appears that in the central and eastern portions of t F e
The districts east o 9 the Mississip i River, as a rule had
8
paid to rainfal K types and to illustrations of rainfall dis-
States ( R ’g. 13). The question whether these types are
8 Inconnection with t h i s referencemay be made tp R. De C. Ward: “Rainfall Types of the Unlted State?,” C&. Rev., vol. 4, Aukust, 1917, pp. 131-114.
different series of years were adopted was investi- gated. It appeared that the Pacific type is less constant in relative monthly distribution than are the Eastern and the Plains types, but the distinguishing characteristics are maintained, even if different eriods of years are
sizing the characteristics of the less distinctive types as
indicated by the relatively short period of 20 years. The Eastern type (including the originally forested eastern United States except the Florida peninsula) has compara- tively uniform preci itation through the year. In gen- eral, autumn has t g e least seasonal rainfall. This is particularly true of the Cotton Belt, and it there consti- tutes a very favorable condition for cotton picking. The Plains type (including the prairie and Plains re ‘ons and
has a marked late spring and summer maximum, with
a dry winter. In the Anzona t pe (western Texas, New
rainfall. The Sub-Pacih ty e (covering most of the country between the Rocky 5 ountains and the Sierra Nevada and Cascade Ranges, and north of the Arizona type) has most of its precipitation during the winter and The Pacdic type (between the Sierra
%k%~~~!%x and the Pacific) has wet winters and
9 summers. A heavy late summer or early autumn ramfall characterizes the Florida peninsula. Four small charts (figs. 21, 31, 41, 51) show the per- centages of the annual preci itation which come in each season. In winter the area % aving the highest percent-
ages of the annual is on the Pacific coast (40-60 per cent) ; in s ring the northern Rock Mountain and East-
35 per cent), but high ercentages (over 30 per cent) are
well onto the Plains. In summer the area of hlghest percentages is still farther east over the Plains proper
(40-50 per cent). There is thus seen to be a seasonal migration of the area of (relatively) heavy precipitation from the Pacific coast eastward to the Great Plains be- tween winter and summer correspondin to the seasons
trols. The Plains have roughly the following seasonal distribution: Winter, less than 10 per cent; spmg, 25-30 per cent; summer, 40-50 er cent; autumn, 15-20 per
ring between April 1 and September 30 is highest (over
70 per cent) over most of the eat agricultural re oqof
and the Prairie States, a fact which brings out in s t a r t l i i prominence the agricultural importance of the “Plains type of rainfall, with its late spring and early summer maximum (fig. 3). A small diagram (fig. 4) shows the period of the year within which 50 per cent of the annud
used. Caution may, however, we1 r be used in.empha-
extending westward to the crest of the Rocky 3 ountains)
Mexico, and Arizona) July and K ugust bring the heaviest
ern Foothil P regions have the hig i est percentages (30-
found westward icto f dah0 and Nevada and eastward
of rnhximum and of minimum manne an d continental con-
cent. The percentage of t E e annual precipitation occur-
the eastern United States, em r racing the eastern % lams
itation occurs. A double-page chart (fig. 15) gives showing for selected stataons the precipitation for the 12 months for each of the 20 years (1895- 1914). The amount of precipitation is shown by a dot and the aver e monthly amounts by heavy lines. The graphs show 7 t e seasonal variation of precipitation, and
also the Variations from the monthly preci itation, which may be expected in different sections. $he frequenc of subnormal monthly rainfalls can be easily determined: On the Pacific coast the variations are large. East of the Rocky Mountains the variations are, as a rule, largest where the avera e amounts of precipitation are greatest.
tively small. East from the 8 reat Lakes the vanations are compara-
MARCH, 1922. MONTHLY WEATHER REVIEW. 119
120 MONTHLY WEATHER REVIEW. MARCH, 1922
Frequency and intensity of precipitation.-Of critical im ortance in agricultural operations is the fre uency
good ded less rainfall m the course of a year than another, and yet the former, because of a better distribution, and of a less torrential character of the rains, ma offer more favorable conditions for agriculture. Specidatten- tion is therefore given to this subject. For each of the seasonal and monthly, and also for the warm-season precipitation charts, smaller charts are given which show the average number of rainy days (0.01 inch or more) (figs. 12, 24, 27,34-37,44-47,54-57). Perhaps the most h t out is this: In eastern sections the g e m annual r s d increases from north to south, but the frequency of occurrence (number of rainy days) often increases from south to north, especially in spring. In other words, rain falls at shorter intervals where the total annual amount is less, thus compensating to some extent for the smaller amount. However, the rule does not always work. For example, the number of rainy days is small in the drier western section while on the northern Pacific coast, where the annud amounts are large, the number of rainy days is also lmge. Two full- awe diagrams show, for eight selected stations, the
#say amounts of reci itation during the four years
1911-1914 (figs. 681, 68k)
For the whole country, the average annual number of days with precipitation of 0.014.25 inch, 0.26-1 .OO inch, more than 2 inches (figs. 71-73) brings out some inter- The average annual number of days with r$?f0.014.25 inch) rains is less than 20 in southeastern C ’fornia and reaches 120 along the North Pacific coast an 3 in the upper Lake region. The average annual number of da s with moderate (0.26-1 inch) precipita6ion
parts of the lnterior plateau region to 50 days alon the north Pacific coast and in the central Ap lachian bfoun-
occurs on the aver e on between 40 and 60 days over
Mountains (fig. 12). During the 20-year peiod, the heaviest rainfalls in 24 consecutive hours varied from over 10 inches along the Gulf Coast in Texas and Louis- iana to 4 inches over the Plains and the Northeastern States (fig. 70 in Atkrs, fig. 1 here). On the north Pacific coast the 24-hour amounts have not exceeded 5 inches, although the mean annual rainfall is heaviest there. The average number of da s with over 1 inch of precipita-
West and northern Lakes area to six days along the Gulf coast (fig. 74). The maximum hourly rainfall (1395-
1914) has varied between less than 1 inch over most of the country west of the Rocky Mountains to about 4
inches over the central Gulf coast States (fig. 75 in Atlas, fig. 2 here). Rains of the “cloud-burst” type, bringing very much larger amounts, occasionallv occur in the southwestern interior. Over the Midale and South Atlantic States and the lower Missouri valley there
occurs on the aver e about once in a season.(March- September) a 3 of 80 or more consecutive days
the same rainfall amounts and season, but a period of 30
or more consecutive days without 0.25 inch in 24 hours, it is seen that the fewest such drought periods are found in the Middle and South Atlantic States, tche Upper Ohio valley, and the northern parts of New York and New England, where they occur on an averagc about once in three sensons (fig. 69). The longest period of consecu-
an t; intensity of precipitation. One section may 1 ave a
. stnking fact here bro
facts.
ranges from. r e88 than 10 in the far Southwest and in
tain region. In the warm season (April 2 eptember) rain
the principal agric 3 tural districts east of the Rocky
tion in an hour ranges P rom less than one day in the
without 0.25 inc K“‘“ of rainfall in 24 hours (fig. 66). Taking
tive days without 0.25 inch rainfall in 24 hours during each warm season for 20 years has also been worked out for selected stations east of the Rocky Mountains (fi . 67).
gay and night rains.-The percentages of rainfall
occurring between 8 p. m. and 8 a. m., sevent fifth meridian time, during the warm season have been c i- arted
(fig. 9). The facts here indicated are of great economic importance. Over the great agricultural States east of the Rock Mountains large sections receive over half and
season rains a t night. There is, therefore, much less rapid evaporation of the fallen rains from the surface, the moisture penetrates more deeply into the soil, a crust is less likely to form on the cultivated surface, and har- vesting and threshing are carried on to the greatest advantage. I t is to be noted, further, that the dominant night rains occur in an important wheat area, and par- ticularly in the harvest season. The district of maxi- mum day rainfall is in the southeast. West of the Rocky Mountains the summer rainfall is about equally divided between day and night, except in Arizona. SnourfaZL-The maximum mnual depth of snowfall shown on the new aver e annual snowfall map (fig. 76)
central Sierra Nevada. The next greatest de th shown is 459 inches, over the northern Cascades, in dshington. Then follow 430 inches over the Cascades of Oregon, 337
inches on the mountains of the Colorado-New Mexico border, 389 inches in north-central Colorado, and 346
and 247 inches in Idaho and in northeastern Oregon, respectively. The heaviest snowfalls in the United States are those of the mountains of the Pacific coast (400-500 inches in some places). The Rocky Mountains have less snow, as is to be ex ected from their more con- tinental location. Neverthe P e s , 200300 inches and more are recorded for some of their western slopes. The greatest depths of snow in the eastern United States occur from northern Michi an to New England. The Adirondack Mountains of a ew York State have 150
inches, and part of the Lake shore of northern Michigan has 120 inches. To the south there is a rapid decrease,
the amounts in the Gulf province being ne ligible (under
lower elevations of southwestern Arizona and southern California, more or less snow falls in all parts of the United States. On the immediate west coast south of
latitude 42’ it is practicall unknown. The snowfall over the Plains ranges from a Jj out 1 inch in central Texas to about 30 inches in northern Kansas, and 20-30 inches farther north.’ The average number of days (not necessarily consecu- tive) during which the ground is covered with snow east of the Rocky Mountains decreases from four months (120
days) along the northern border to one day in the central portion of the Gulf States (fig. 78 in Atlas, fig. 3 here). West of the Rocky Mountains the data concern the lower altitudes only. Most of the central Plateau rovince has
Pacific coast, less than 10 days; southwestern Anzona, 0 days. A new chart (fi . 79 in dtlas, fig. 4 here), p e s
(0.01 inch or more, melted). The range is from one day
over the Gulf province and the central and southern
considera 3: le areas receive about two-thirds of their warm
anywhere in the Unite 7 States is 527 inches, o v e the
1 inch). With the exception of southern A orida and the
from one to two months of snow cover; t !I e northern
the average annual num % er of days on which snow falls
e In connection with snowfall, reference Fa be made to the following: Charles F. Brooks “The SnowfalloftheUnlted States &art.Journ.Roy. Met Sor Pol 39 1813
p 8 1 4 : R. De C. Ward: “The Snowfsll’of the United States,” Sk. ddA.. 401. 9: &9, pp. 39745.
MARCH, 1922. MONTHLY WEATHER REVIEW. 121
- .. ..
I
FIG. &-Average annual number of days with snow cover.
~
____ ~-
~ ~~
FIG. I.-Avemge annual number of dnys with snowfall (0.01 Inch or more, melted).
122 MARCH, 1922 MONTHLY WEATHER REVIEW.
MARCH ,1922. MONTHLY WEATHER REVIEW. 123
124 MONTHLY WEATHER REVIEW. MARCH, 1922
California lowlands to 80 days in extreme northern Michigan. In the West no attem t is made to take account of the greater elevations. considerable popular interest attaches to the avera e date of first autumn
is before Se tember 16 locally in the Rocky Mountain
ber 16 in the central portions of the South tlantic and Gulf States. Farther south, snow falls very irregularly, or may often not occur at all during a whole winter, ns
alon the Gulf coast.
l'funde~storms and haX-Another quite new chart shows the average annual number of days with hail dur-
snowfall (fig. 77 in Atlas, fig. 5 fl ere). The earliest date
region; Octo g er 1 in extreme northern Michi an; Decem-
s
fog (dense fog obscures objects a t a distance of 1,000 feet from the observer). The Pacific and the northern At- lantic coasts have the greatest fog frequency (over 40
days a year). Over the interior dmtricts there are en- erally fewer than 10 days a year with dense fog. %e western Gulf has over 15, as has the upper Lake re- gion. Two charts showing the average number of clear and of cloudy days complete the folio (figs. 83 and 84 in dtlas, figs. 8 and 9 here). Fewer than 100 clear days a
year are usually recorded in the Lake region and on the northern Pacific coast, while in the southwestern interior there are more than 300 clear clays. Cloudy days aver- age 160 a year in northernmost Michigan and 180 a year
ing the crop- rowing season (fig. 80). The Plains States and the Roc5 Mountain area have the most frequent hailstorms. d s t of the Mississi pi, hail occurs on the avera e only about once R year luring the cro season.
days annually) on the central and estern Gulf coast, n
considerable section of the southeastern United States having over GO thunderstorm days a year (fig. 81 in Aths, fi . 6 here). On the Pacific coast, thunderstorms
annually. Fog and cloudineBs.-The first fog map of its kind for
the United States (fig. 82 in Btlas, fig. 7 here) gives the distribut.ion of the average number of days with dense
Thun 3 erstorms have tsheir greatest frequency &O to 90
are usua 5 ly recorded on not more than two to four days
on the northcrnmost Pacific coast.. Fewer than 30 da,ys a year nre! as R rule. cloudy in southwestern Arizona and sou theas tern Cnliforn ia. To do full justice to the new precipitation folio is quite impossible in a discussion such as this, which must neces- sarily be little more than a catalogue of the ne- charts, with n few words of description in each case. The cnrto- graphic work is escellent : the colors on the mn.ps are well chosen; the text, while brief, is quite adequate. Thc whole folio deserves, and will surely receive, very careful and serious study on the part of all who have any interest in United States climates, and will do much to establish American climatology on a higher plane of scientific accuracy.