MAY. 1917. MONTHLY WEATHER REVIEW. ’ 209
@
SECTION 11.-GENERAL METEOROLOGY.
SUMMEB TYPES OF W F A L L IN UPPEB PECOS VALLEY.
By CLEVE HALLENBECK, Observer.
[Westher Bureau OWce, Roswell, N. Mex., Apr. 30,1917.1
While there are but three recognized primary types of rainfall (cyclonic, convective, and orographic), subdi- visions of these three types are necessary to the classifi- cation and description of the rains of any one locality. Thus, a t Roswell two distinct tvpes of cyclonic rainfall have been observed, one of whch is confined to the colder half of the year, while the other niay occur during an month, and this latter t e can be Iurther subdi- vi -B ed into two groups, one opwhich occurs during the summer and the other during the winter. All tliunder- storms are classed as convective, but they may exhibit lightly Merent characteristics in different portions of the country; the- thunderstorms occurrin at this station
observed by the writer in the Mississippi .Valley and Eastern States. It seems, therefore, that accuracy in classification could be well served by recognizing sub- divisions of the three primary types as long as such subdivisions exhibit individual characteristics. The summer rains of the upper half of the Pecos Valley ma thus be classified as follows: Anvective :
Cyclonic :
The h t
being different in at least one particu 9 ar from those
1. Thunderstorms.
2. Local showers.
3. Transitional rains.
4. Nonconvective rains. up-thunderstorms-includes all rains of this type, E h of purely convective and of cyclonic
origin as the term is generally defined. It might be remarked as an interesting point that the so-called “squall cloud” is seldom observed in thunderstorms at this station. The writer bas not once
OCCUl?f observe this cloud in any thunderstorm during the last two years, although the “s~uall wind” nearly alwa s
squall cloud may be present but hidden by a curtain of rain. The second subdivision of convective rains-local showers-includes light rains of brief duration that occur at this station during the warmer hslf of the year, and which are of convective origin, although they occur a t night as well as during tho da less frequently, however, at night. This type is well gown at stations near the Gulf coast but it is believed, is not of frequent occur- rence in the devated regions of the west. These local showers occur when conditions are favorable to the formation of thundershowers, and i t may he that some
of them are thunderstorms in the first stage of develop- ment. The majority of them, however, are not onibry- onic thunderstorms, as they are frequently observed to die out. As observed at this station, the rain falls from an unusually la e cloud of cumulus formntion, or from a patch of c l o s covering only a portion of tho sky, making a “trail” of precipitation of very limited breadth across the country. On one date in 1916 no less than five such clouds were observed at the same time in
occurs. It is possible that in some of the storms t l e
differcnt parts of the sky, from each of which rain was falling.
The two subdivisions of cyclonic rainfd are believed to be unusua.1 t. 8s not generally hown or recopired in
they are made the main subject of this discussion. They aro primarily cyclonic, but under certain pressure con-
them strikin~y ~i ~~i v i d u s .1 characteristics, which di F er- ditions the to owa. hic. features of the Pecos Vdoy
ontiate them rom the usual t,ype of cyclonic rains. These characteristics are outlined in t,he second paragraph fol-
“‘;Figure 1 the relative amounts of rainfall md the rela- tive hourly frequency of prec.ipita.t,ion from the different t,vpes of rainstornis n.re presented. The left-hand por- tion of the fic*uro gives the total amount of rain for each hour during h e months May to Sept.ember, inclusive, for the 13 ycars 1905 to 1916. The right-hand portion gives the total number of times rain has been recorded in each hour over the same intervals of time. It will be seen
from this figure (and also from fig. 2) that each of the first then t es has its own hours of maximurn rainfall. This phase o YP the subject will be mom fullv discussed in a later paragraph; att.ention is invited to these Sgures at this point to give an idea of t,he relat,ive importance of the cliff erent types of rainstornis. Tho two subdivisions of c clonic rainfall are so nearly alike that! there is some do& as to whether they should, be differentiated, and, after noting their points of dis- similarity, they will be treated in the mam as a single tipe. As observed a t Roswell, they have the followmg c aracteristics in common: 1. The storm always R proaches from the south.
2. The wind direction R uring the storm, and frequmtly for some time before, is northerly. 3. Tho baromet.ric pressure, except for diuinal varia- tion, is practically stationa . 4. Tho rain begins nt n i g t and is heaviest after mid- night.
5. The velocity of the wind is groater toward the ond of the rain than at the beginning. 6. These rains are confined to the warmer half of the year.
7. They occur only when there is a greater than nonnal pressure gradinnt from the middle plains region to Arizona. The two differ in the following characteristics:
1. Group 3 is accompanied by thunder, while group 4
is not.
2. The average time of maximum preci itation inten- sity of group 3 is 10 p. m. to 2 n. m., w h i of group 4 it is 2 to 6 a. ni. It should be said that the above-named points are characteristic of the f!yliical ra.ins of this class, and not of every one of such rains that occur. The rains of roup 3 aro called “transitional” rains . because, while t ley dosely resemble those of grou the are d i e d to thunderstorms, and are recorde as
tion, n .~ well as their period of masimum fre uency o?pre-
other portions YF o the United States, and for this reason
B 4 J
suc i . Their period of maximum intensity of preci ita-
cipitation, is intermediate between that o ? the noncon-
7
210 MONTHLY WEATHER REVIEW. MAT, 1917
Amounts. Frequency.
I I I I I
FIo, l.-To@l hourly anaounts nnd totd hourly frequencies of rilinfdls due to the four.types of precipitation distinguished nt Iiusaell, N. Mes., Iur the 12 years 1905 to 1916 inclusive.
MAY, 1917. MONTHLY WEATHER REVIEW. 211
vective rains and that of thunderstorms. The writer be- lieves that these rains begin as thunderstorms farther down the valley-probably in Texas-and by the time they reach this portion of the Pecos Vdey they have developed into a form closely resembling the nonconvec- tive rainstornis, with a front 50 to 150 miles in estent, and deliverin gcncrl rccipitntion over both the valley
their main coiivective chnrac teristics. It d l be shown
began) ; in 3 it remained southerly throughout, and in the other 2 was variable; 6 of the storms were accom- panied by thunder; 6 delivered “excess+” precipita- tion, each time a t night and 4 of the 6 times after mid- night. In no storm was “excessive” precipitation re-
slopes. By t i e K timn t f) ley reach Roswc11 they havn lost
kt&, however, that the rains of bot.h group 3 and 4 psr- take to a limited esbent of the nature of bot’h convective and or0 raphic rainfall. A stusi of the original rcports of cooperative stations in t.he va ey was made in tin effort. to determine how far to talle north and south of Rosu-oll thesc rains esteiid and whether they prestmt the same characteristics at other points in the valley than Roswell. Five of the above-
named charactenstics were verified for other stations as far south as Carlsbnd and north as far as Fort Sumncr. The third and fdth, of course, could not be verified from the cooperative reports. The . south-to-north movement of the storm was
clearly demonstrated in this investi ation, and in most cases the be innin of rain showef a regular rate of
progress up t %s e v ley. In three instances, the hegin- ning of rain was irregular, and in one stom the rain began at the up er end of the v d e y and
Hower clouds! as observecl at Roswell, moved from the south. A few storms of this class have been observecl a t this station, in which t,he soutlierly direction of t,he clouds changed t,lirough east.erly to northerly t.ownrd the end of the rain, the chitngc being at,t,endc.d by rising barometer. It was more difficult t,o ascertain how fa.r up niid down the valley the northerly wind extended in these sbornis,
as the observers record the prevailing direction for daylight houis while tdie rains occur usu~~llp tit. night. Sufficient data were obtained, however, to warrant, the assumption that this wind niny estcntl a distance of 150
miles, and possibly farthcr. For esaniple, the coo er- ative observer (IT. S. Reclti.niat,ion Service) a t Cads E ad, which is the most southerlp station in the Pccos Valley in New Mesico, reported that during the record rain storm of August 74, 191.6, the wind, which had been from the south and southeast, shifted to north a t the beginning of the storm, and durin the entire time of the
sent’ to interniedintc. stations brought praotic?ally the same information.
111 connection with rains of t,liis type, I find such notes as these eiiterecl on t,he locsl record of this station: “Storm backed against the wind” and “the storm nioved from south to north against the wind. ”
While the weather ninps nisy show n preusurc gradient over eastern New Mesico during these rains, the barome- ter at this station reiiiaiiie practically statiunctry during the storm and usually for 6 tp IS houis before. No heavy rains wcre found of record during which the barometer was rising or falling materially. Tabulated data for 18 of these storms, each of which delivered one inch or more of rainfall a t this station, gave the following results : Thirt,een began a t night and
5 in the afternoon; 15 delivered their heaviest precipita- tion at night (11 of t,licse after midnight) while the remaining 3 showed no well-defined period of masiniuin rainfall; in 13 the wind was northerly throughout the storm (in 6, also northerly for some time before rain
apparent$ rogresse southward. In $1 four of these cases t,he
rain it blew from the north am 7 northeast. Inquirics
Totals.
*rime fu b 6P) Ramfa// N/*/ fCPkLAIt?atajW/
FIG. 2.-Total amounts and total hourly Ire lienpies of precipitation for each of the four types of recipltation, by 4-hour periols, to; the 14 years 196 to 1916, Inclusive, :it Roswell, d Mex.
corded with a southerly wind. Durin 15 of the 18
No rain storms of the type under discussion were found on the local records for the months November to March, inclusive, although they have occurred in April
storms the ressure remained ractical k y stationary at Roswell. Af the storms move B from the south.
212 MONTHLY WEATHER REVIEW. May, 1917
FIG. &-Hvpsometric sketch map (contour interval. 1 OOO feet) of southeastern New Mexlco showing the theoretical air movement m a s the redon when under the inhence of @ HIGH to the northest, a Low tdthe,west, and the Gulf of Mexico. dong thin mows-direction of lower clouds and the general surface wlnd. Short arrows-direction of the underrunning lught current.
MAT, 1917. MONTHLY WEATHER REVIEW. 213
and October, which two months were not included in the data from which figures 1 and 2 were constructed. Such of these storms as are accompanied by thunder exhibit no other thunderstorm characteristics. There
is no abrupt shift in the direction of the wind and no increase in velocity a t the beginning of the rain. Pre-
cipitation usually begins a t n light, rate, and is sometimes more than four hours in reyhing its masirnuin intensity. The barogra. h shows iio m e in pressure such as com- monl atten x s a thunderstorm, nlt.hough two cases were not.eJwhere the pressure was very unsteady throughout the storm. The thunder is seldom heavy, and is nearly continuous. In the storms of this class which have been observed by the writer the elect.ric dischnrgrs apparently took place above the lower clouds, ns 110
In tlie storm of April 16,
“=qzaq 1915, t e t B under WRS in the form of what, might be called “thunder waves” which moved rapidly from south to north a.t interva.ls of 15 or 20 minutes, ench “crescendo” being attended by an increasin rate of
rate.
li htning was seen.
precipitation and each “diminuendo ” by a 8: ecreasing
rains may occur when there is no or LOW located as before described, condition is n moderate HIGH
to the northeast and a LOW over Arizona. Since n LOW
is practically permanent in the huthwest during the summer months, it is only necessary to supply the HIGH
in order to have a pressure distribution favorable to such rains. When the HIGH is resent the Arizona LOW is
increased gradient between tlie LOW and the niicldle Plains region. In this connection attention is invited to figure 5, which is a composite map of five pressure distnbutjons attending five ty ical iionconvective rains
pressure areas are not clea.rly defined there is n
than normal decreasing radient from central
~~~~~~
to Arizona. The norm5 gradient varies soniewhat during the season, being about 0.17 inch in Mny niid 0.30 inch in July. The average during 18 rainstorms was
0.38 inch, or about double t,he normal gradient. Since the barometer a t Roswell remains nearly station- ary on such occasions, it seems that, the HIGH and the
LOW ore about balanced so far as their influence a t this point is concerned. I t has been re entedly observed
elevation of the lower clouds, while if the LOW- is the controlling feature the flow of a.ir will be from the south and southeast a t both elevations, day and night.
Theoretical exphnution.
With these established facts in view-(a) the sout,h-
‘ erly movenient of the ra.instorm, (b) the northerly direction of the wind, (e) the maximum intensity and frequency of precipitation occurring a.t night, and (d ) the attenda.nt pressure distribution over the Southwest- the following esplanation of these rains is offered: Unless the a.ir circulation in the Pecos Valley is co’i-
trolled b.y a redominant high or low pressure aren, the
back to southerly in the forenoon. At Eoswell t c night
reaches north a t about hours nft,er sunrise and reaches south about noon. The shift to northerly is through west and to southerly is through east. The
apparently deepened, but. t K is no doubt is due to the
in the upper Pecos Valley. I! ven in cases where these
that if the HIGH be redominant nort K erly currents will prevail, day and nig R t, both near the surface and at the
R
wind shifts f roni southerly to northerly t i t ni ht ant1
shift in summer normally
back to south begins
a t about 9 p. ni. and The dayt,ime shift
dayt8ime shift from north to south is more clearl marked than the night shift back to north. The diurna s changes are more pronounced in winter than in summer and during the latter season are nearly masked, in the pre- vailing hourly directions, by the predominant southerly winds. Since the pressure distribution is such as to maintain, for the t.ime, nearly stationary barometer at this point, the diurnal shifts of the wind occur without material moclificnt,ion. Or it may he said t,hat the southwestern
LOW controls the lower air movement during the day,
while the HIGH to the nort.l~emt controls it at ni ht. The flow of the air upslope toward the LOW woul f be rein- forced during the day bv t.he normal tendency to flow up- slope during the day, while at night t,he outflow from the
HIGH would be reinforced by the nocturnal tendency of t.he air t,o flow down t,he valley. This implies a deflection of the outflow from the HIGH contrary to the deflection caused by the rot.ation of the cart-h, but it is an observed fact that a HIGH over t,he middle Plains does roduce nort.herly winds in tlie Pecos Vallqy, and materi Bp1 y pro- longs t.he period of downslope mnds at night. If the
HIGH is absent, sout,herly winds will prevail throughout the night, with the possible esce tion of an hour or two
rains of the nonconvective type, is a combination of air draina.ge and of out.flowing air from the HIGH.
Under the part.icular pressure distributeion in question, it seems that at t,he elevation of the lower clouds the Southerly flow of air toward the LOW is maintained
in t.he early morning. This nort, P ierly wind, then, during
age or as a part of the outflow from the HIGH or a com- bination of the two, is an undenvnnin wind flowing
fact that such rains in their typical form occur only when the HIGH is reaent seems to indicate t,hat in ita absence
posint sout,herl , J flow of air, specially in view of the fact that t e ener cloudiness at such h i e s would retard the
the formation of a purely drainage current. h a t this wind is not due to the coohng of the lower air by precipita- tion or by evaporation of precipitation, is shown by the numerous inst,ances in which it starts some time before rain begins. In the two storms whose precipitat.ion areas
are chartzed in figure 4 this was t,rue. In the storm of July 23-34, 1911, the wind was northerly 10 hours before rain be an, and in that of Au t 7-8, 1916, it shifted
although i t did not reach due north unt,il a few hours 1at.w. If this uncierruniiing wind is due to the HIQH, this re- supposes khat the outflow from the HIGH is cooler t R an the inflow toward the LOW, which is contrary to the idea that in summer the HIGH is the warmer. It is believed that. this special case is an esception to the rule. It will be remembered that the condition being dealt with is es- sentially a night condition, and most pronounced in the latter part of the night-. Radiation would be practically unclieckerl in the HIGH, while in the southerly rain-bearin current, which is, as a i d e , cloud laden, radiation woul be greatly retarded. Therefore it seenis reasonable to assume that at night, and specially fifter midnight, the out.flow from the HIGH is cooler than the air moving up the Pecos Valley. It was found that the average 8 a. m. temperatures, for a number of summer pressure distri-
beneath t.he rain-bearing current from t % e south. The
the down-v op1 ey wind would not develop ’against the op-
c,ooling o 9 the lower air by radiation, and thereb prevent
out, of t, a e sout,h t,hree hours be ff“ ore the beenning of rain,
f
314 MONTHLY WEATHER REVIEW. MAY, 1917
butions of the type under discussion were 2 degrees to 5 de 888 below normal over the area occu ied by the IIIQH,
and about normal over southeastern New Mexico and wcstein Texas. It is believed that this undeiming wind is effective in increasing the rate of preci itation, and that it may start precipitation. In severa P instances the beginning of rain was approximately coincident with the arrival of the northerly wind. In the case of the storm referred to in an earlier amgraph ( .211) where the rain began
“backed” southward (the rain clouds moving rom the south), the rain and the norther1 wind began a t about
other stations in the valley, then it seems safe to assume that the northerly wind underran and lifted u the
cipitation. If this is t.rue, these cyclonic rains are allied in that one characteristic to cyclonic thunderstorms, and are to that extent convective.. The down-valle wind would, normally, be most com- pletely establisliei during the later hours of the ni ht,
heaviest recipitation. I t may be that the cooling of
mducing precipitation, since its cumulative effect woulcl gecome greatest during the latter part of the night. In the case of these night rains, where tlie untlerrun- nin wind is catablished in the valley, three causes of
moist air by its ascent up a comparatively stee the further cooling of this air by radiation, and t e cool- ing effect of heino witlrrrun and lifted up by the tlowii-
slope wind. h i $ since this underrunning current must reenter the prevailing circulation, it may in that way be a fourth cause of condensation, since it is, presumably, cooler than the moist air flowing upslope. AI1 these, acting simultaneously, are sufficient to rrccoont for the heavy and sometimes excessirc r i n fill1 that this type of storm frequent.1 delivers in the later houis o f tilie iii-lit.
primarily cyclonic, also partake of thc nahure of both convective and orographic rainfall. An attempt is made in figure 3 (p. 212) to illustratcr this theoretical air movement. On this figure the surface contours are drawn for every 1,000 feet elevation up to
7,000 feet. Higher olei-ations are not shown, but in Lhc northern part of the St&o tho land rises to 12,000 and
14,000 feet. It will he noted that the mountain range west of the Pecos Valley, while low (6,000 to 8,000 feet) is still 3,000 to 4,000 feet higher than the trough of the vallcry throughout its entire course, and that this course, which is southeast-northwest in Tcrxns changes to south- north through Now Mexico. Tho normal direction of
winds over westorii Tesns and eastern New Mexico, moving in toward a LOU- centered over Arizona, should be southeasterly and it is observed that they take this direction over all of that region esce t the Pecos Valley
It is gelieved, therefor(\, that the rango to tho west is effective in deflecting the current to the right as it moves up tlie Pews Valley, until by the time it reaches Roswell it is moving from the south (nlso south at Artctsia, but south- southeast at Carlulad). It also seems probable that the high mountain ranges in the northern part of the State (12,000 to 14,000 feet) deflect the overflowing air from the HIQH toward the
1 IF egree to 2 degrees above normal in t R e Ansona LOW,
P
first at the nortiern P end o P the valley and ap arently
the same time a t Roswell, and i 7 the same was true at
southerly rain-bearing current and thereby startec P pre-
and it is during those hours that such rains deliver t % eir
the rain- E earing current by radiation plays a part in
con 9 ensation are active: the meclianicd cooling of the
1
It will be notec 9 froni the above that these iitiiis. w?de
in New Mexico. and with tho possib P e exception also of ortioiis of the Chiitttlitui niid Grmide Valleys.
left, so that it moves southward through the col con- necting the Pecos and Canadian Valleys. This doflec- tion, as remarked elsewhere, is contrary to the deflec- tion due to the earth’s rotation, but nevertheless it is believed that it occurs. The Canadian Valle is very
while it is flanked on the west and northwest by moun- tains near1 twice as high as the range which borders
doubt deflects the wind moving up the Pocos Vallay, and it seems reasonable to assume that the high ra~iges
to the north would have a similar effect on the outflow from the HIGH.
In figure 3 a system of winds is represented as flowin toward the LOW from the coast regions of tlie Gulf an(
also from the HIGH, this movement being represented by the long arrows. These arrows represent the diroction of the lower clouds as well tu of the surface wind. The smaller arrows represent the underrunning drainage of air down the valley at night, which iiioTes below and in
an op osite direction to the prevailin system of winds.
only durin rains of the type under discussion. Whether
tion continues (luring the night when the weather is clear, is not known. It is possible that similar night drainage may be found, undor the proper prossure con- ditions, in the valleys of tho Canadian md Grande, but such is shown here only for the valley of thc Pectos.
While this is to un extent an idealizocl syatciu of winds, tlie prevailing circulation conforms to thc generally ac- cepted theory of air niovenients under thc given prctssuro distribution, while the iiortherly wind in the Pocos Valley is in accordance with observed conditions at this and other stations in the valley. Further with regard to this underrunning wind as a
factor in producing precipitation, it may bo sitid tlin,t in u nuinber of storms, where tho wind wtis southerly t1.t thtr
sttirt, tlie rate of precipitation at, Roswell bc?gm to in- crousu simultunsously with tho shift of the wind to tlie north. In one lienvy rainstomi, tho wind shiftad from northerly to southerly twicer during the rain, and each shift was marked by a decrease in the intensity of tho rain, which increased again with each shift back to northerly. Among the numerous storms whose r i p i t a t i o n was charted were a few in which the area o heaviest rainfall was in the Canadian Valley, with a secondary area in the valley of the Pecos (e. g., May 13-14, 1911, May 28-29, 1911), and others where general-rain occuqed over west- ern Texas and eastern New Mexico, but which was heavi- est in the Pecos Valley (Sept. 11-12, 1912; Apr. 24-25, 1911; Apr. 13-18, 1915). The conclusions reached by the writer-and they are supported by the available data-are (1) that these rains are primarily cyclonic, bein produced by the mechan- ic& cooling of warm, humif air moving ups10 e, under
a,nd (2) that the heavy nighttime preci itation in the
in wind by the denser valley drainage, with the probable
a a ded effect of the cooling of the saturated air by radia- tion and by mixture with the underrunning wind. It is only under certain pressure conditions that air is brought from the Gulf directly over western Texas and eastern New Mexico. When such conditions exist, the moisture-laden air ascends a gradual slope from sealevel to an elevation of 3,000 to 4,000 feet or more. I n sum-
mer this nearly saturated air is further elevated in the
nearly of the same elevation as the valley of t T ie Pecos,
the Pecos B alley on the west. This lower rango without
f
This, \e it understood, represents t a e air nio~-ements
the prevai B ing southerly flow of air at the higher clcvu-
the esisting pressure conditions, from the Guf f region;
Pecos Valley is due to the underrunning o !f the rain-bear-
May, 1917. MONTHLY WEATHER RBVfEW. 215
formation of convectional and c clonic thunderstorms, and this may account for the zeavy down ours that
States. But this does not explain the fact that when these rains are heavy they are often confined to certain river valleys and, under certain pressure conditions, are always heavier in the Pecos Valle than elsewhere.
rainfall at ths station is much more irregularly distrib- uted over the 24 hours than that of any one of the three grincipal types. This is due to the fact that each t e
mamma, from 4 to 5 m., from 10 to 11 p. m., and from
rainfall from the three principal t pes. The nonconvec-
uency at the time of maximum thunderstorm rainfall. %o typical thunderstorm precipitation has occurred at this station, during the last 12 years, between the hours of 4 to 9 a. m., and no nonconveotive rain of the transi- tional type has fallen between the hours of 1 to 5
Nonconvective rains of roup 4 have occurred a t all ours
ered by storms of this class has been due to rains that be an a t night and continued during part or all of the
Figure 2, which gives the amounts of rainfall and the total hourly frequency b 4-hour periods, shows more
types of rainf&. The total amounts and the total hourly frequency, as charted, are given in Table 1.
sometimes occur over the semiarid portions o P these two
Referring again to figure 1, it wi 1 be seen that the total
as its own eriod of maximum and minimum rai .& . In iigure 1, s %. owing the total hourly rainfall, the three
2 to 4 a. m., mark t P; e hours of maximum intensity of
tional rains eatly preponderate J uring the hours 9 p. m. to noon, wit r thew minimum amounts and minimum fre-
K' mm of the day, but most o B the daytime precipitation deliv-
io f owing day.
clearly than fi ure 1 the (ry ifference between the different
*-4" P" !-
TABLE 1.-Total amounts a d hurlyfieqtu?neiCe of rainfall for the 4 -h r periods, cla8aajW by typea.
Total amounts.
10a.m.Zp.m. 6p.m.lOp.m. 2a.m. 6a.m. Totd.
------- I I I I I I
T Y P .
83.30 18.43
31.0 4.97
Total rainfall. __. . -. . . I 9.42 I 9.47 I 20.50 I 17.40 I 17.14 I 14.371 813.30
. . . -- - - .- - --- ... I
Total hourly frequency.
Types.
Thunderstarms.. . . . . . . . . . . 2 Transltionalrains .._.____.. 1 f I ? 1 I 'g 1 1
35
Nonconreetive rains ...._._ 93 39 32 57 1M 121 Loml showers.. . . . . __ __. . ._ 3 2 3 5 2 S 3 9 1 7
T01.w.. . . . . . . . . . . . . . - 1 146 I 110 I 215 I 2311 I 228 I 175
820
1lM 415 180
1,110
-
For all rains during the period covered, a total of 39.39 inches has fallen from 6 a. m. to 6 p. m., and 48.91 inohes from 6 p. m. to 6 a. ni. Durin the same eriods of time from 6 a. m. to 6 p. m., and in 639 hour-periods from 6 p. m. to 6 a. m., thus establishing the fact that in Summer both the amounts and the frequency of rainfall at this station are greater a t night than during the day. The summer rains of this portion of the United States are essentially da
limited area occupied by the upper half of the Pecoa
rain has been recorded in 471 a our-perio
time rams, and the preponderance of night rains over t i- e
D
FIG. 4.4ODsl ralnloll~ over mutheastern New Mexico and western Texas, resulting from two typical stonns-July 23-24,1911 and August 74,1916.
216 MONTHLY WEATHER REVIEW. M A Y , 191’9
Valley is wholly due to the occurrence of this peculiar ty e of nonconvective rainfall.
it might be ar ed that a comparison of the total amounts of raidfielivered by nonconvective rains with the total hourly frequency indicates that the rains of this type are frequent, but light, while in this discussion emphasis has been placed upon the heavy rainfalls de- livered b this class of storms. It is true that as regards the hour rate of precipitation, most of these rains are light, alt H ough.more than half of the heavy rains of 1 inch or more at. this station were of this type. Pre-
ci itation occumng a t the rate of 0.05 inch per hour is
a i g h t rain, but if it continues at that rate for 24 hours or more the accumulated depth amounts to a heavy rain. Typical storms of this class usually deliver heavy or excessive precipitation for 1 or 2 hours, while the actual duration of recipitation may estend over 12, 24,
ample, “excessive” precipitation was recorded for 2
hours 20 minutes, without a break, while rainfall was
recorded for 13. hours before i t reached the “esc.essive” rate, and continued for near1 three hours after it fell below the “excessive” rate. %his class of storm owes its importance chiefly to the fact that general recipita-
thunderstom rainfall, although it may occasionally be heavy, is much more localized. Figure 4 presents the total precipitation from two typical nonconvective sumnier rainstorms. These two storms were selected from 18 that had been charted, not because they were estrciiie t-ypes as regards the dis- tribution of precipitation, but on ac.count of the heavy rainfall delivered. These two eshihi t a.n ohservcd char- acteristic of a nuniber of the hea.vy rains of this c h s ,
viz, the precipitation decreases more rapidly froni the Rio Pecos westward than it does to the eastward, whereas one would naturally es ect 1iaav.y precipita-
of the tendency of the rain-bearing winds to be drinvn up that slope toward the LOW. 9 few storms, however, carried their heavy rainftiB clear to the crrst (J f the rangc flhnking the valley on the west.
. Moat of the rains that were charted showed B somewhat great.er north-and-south elongation of the prixipitatioii area than the two shown in figure 4. In the stornis where the heaviest precipitation was in the Cansdinn Valley, the center of the southwestern LOW was locatcd near southwestem New Mexico, thus bringing the flow of air from the Gulf directly up the Canadian Valley.
Floods at Roswell due to hmuy rcr?i.)u.
It is an interesting fact that all the serious floods of record in the vicinity of Roswell were caused by storms of the nonconvective type. A number of thunderstorms have delivered heavy rainfd, but in all cases of record
the were too localized to cause serious flood ccuidi tiims.
occurred since this station was established, thc notes
being copied verbatim froni the locd records. The precipitation given is that occurrin at Roswell, during the entire StOrIIi, which sometimes 5 asted more t,hnn 24
hours.
July 25, 1905, 2.75 inches. Flood in Iiondo at reservoir; lo\~lantls of North Spnng River under water.
September 1, 1908, ’1.12 inches. Hondo bank lull; IonlandR under water, pastures flooded.
April 26, 1911, 1.71 inches. Pastures and lowlands tlooded.
Ma 29 1911,1.66 inches. Hondo out of banks, overflowing through city, Towhnds under water.
or 36 hours. I n t f e storm of August 7-8, 1916, for cs-
tion is delivered over the upper Pecos Val P ey while
tion over the western slope of t ! e valley on account
Fo s owing is an,annotated list of the fioods wliich have
July 24, 1911,0.58 inch. High water in Hondo; intake at dam gave wa ; flood waters also coming down Rocky Arroya; Berrendos bank- fulc bridge at Urton’s 4 feet under mtex.
[In this storm the heaviest recipltatlon was north of the station and between Ros- well and the Capitan Younta%s.]
June 12, 1913, 1 inch. Flood conditions in Rio Pecos; water 2.7 feet above flood stage.
October 25, 1914, 2.24 inches. No high water in Hondo; Pecos reported bank-full ; lowlands along North Spring and Berrendoe Rivers under water.
April 17, 1915, 1.91 inches. Lowlands under water; intake at Hondo Reservior gave way, Hondo overran its banks along ib entire course, flooding the city (Roewell) with 2 to 3 feet of water on the streets; basements and many loner floors flooded, including Federal building; several residences partly undermined. Cattle companies report thousands of sheep and large number of cattle drowned; fences nnd bridges ewe It away: boating on the businem streeta the popular pastinie of the day.
Streeta running curb-full; North Spring and Berrendw bottoms under several feet of water; thousands of birda drowned by over two holm excessive downpour; some bridges gone along the Berrendns. washouts reported on railroad; loss of live stock alight.
There dso were a number of serious floods during the yrnrs prior to 1905. cnnccrning which no reliable infnrnia- t.ion is nvailaldc. One, which occurrcd sonic time dur-
dugust S. 1916. 5.57 inches.
Flood in Rio Pecos below highKay bridge.
ing the carly fall of 1901. was of greater that of ,Ipril 17, 1915, during which new channel ovcr part of its (wurse through the city.
FIG. 5.-Average pressure gradient olSummer (MIay-September) over the southwestem United States (- - - -) contrasted with the arerago pressure distribution in the c s e of 5 typical nonconvrctive rains at HmwAl, N. Mex. (-1.
It should be added, in conclusion, that the suninier type of cyc.lonic rainfall is quite distinct from a winter type in the I’ecos Valley, which is produced by the admix- ture of cold air from a winter ~IIOH with warmer and moister air that had previously moved in from the Gulf region. This class of winter storms rarely yields mor’e than light precipitation, in the forin of slow, cold, and clisngreeahle rain in the late Fall and early Spring, aJrd snow in winter. They are also different froni the rains produced by the eastward movement of a LOW. across southern New Mesico. This latter type occurs during the Winter arid Spring. being of most frequent occiirrence in March and A ril.
occurs at this station in thc months of A ril and October.
that began as the suniiiier typc uf nonconvective rain, and toward their end dt~gen~mted into the fiixt-nained winter type.
Eveiy type o r rainfall known to eastern New Mesic0
A number of rains have bern noted in t E cse two months