October 1970ON THE POSSIBILITY OF WEATHER MODIFICATION BY AIRCRAFT CONTRAILS WALLACE B. MURCRAYGeophysical Institute, University of Alaska, Colleye, AlaskaABSTRACTThe possible effect of contrails in modifying the weather is reconsidered in the light of information obtained fromground-level contrails in Alaska. It appears likely that inadvertent cloud seeding by jet aircraft may be of the sameorder of magnitude as that attained in commercial cloud seeding operations. Further investigation is needed; but inthe meantime, the possibility of contrail contamination should be kept in mind when evaluating the results of seedingoperations.1. INTRODUCTIONAircraft contrails first attracted public attention duringWorld War 11; but as air traffic has built up to its presentlevel, they have come to bo accept,ed as part of the environ-ment. Even during World War 11, it as difficult to watchthe cloud cover laid down by a large bomber formationwithout wondering what it might be doing to the weather;at present,, there is widespread belief among the generalpublic and some feeling among scientists (Fletcher 1969,Reinking 1968, Livingston 1969, and Schaefer 1969) thatcontrails are increasing cloudiness, if nothing more, insome regions. The writer himself has seen instances inwhich a single conhi1 seemed t,o grow until it became a.novercast covering the whole sky. If t,he contrail were indeedresponsible, which is by no means certain, this wouldconstitute definite proof that contrails are capble of asignificant effect on local n-eat8her, and even possibly onglobal climat,e, if such occurrences arc widespread andfrequent. It seems wort.hmhile, in view of all this, t.0consider quantitatively whether or not there is reason tobelieve contrails are capable of exerting a significant influ-ence on weather.9. OBSERVATIONS ON THE COMPOSITION OF THE CONTRAILAny quantitative discussion should start with thecontrail itself. A medium-sized jet airliner-the Boeing 727to be specific-burns over 3100 kg of fuel per hour whilecruising, producing over 1.2 times as much water (morethan a kilogram per second). It is this water that forms thecontrail, and a discussion of the basic processes by which itdoes so was given by Appleman (1953). His discussion alsogave a fairly accurate description of the conditions underwhich a contrail mill be formed, but did ndt discuss thenature of the contrail particles further than to specify t,hatthey are ice crystals. Some attempts have been made tosample the contrail particles in flight, but there is alwaysskepticism as to how well the particles collected under suchcircumstances represent those in the contrail. During thewinters of interior Alaska, however, conditions arefrequently such that a contrail is left while the aircraft ist,axiing and t'aking off or landing so that it is easilyaccessible for study. (On t'he ground, the contrail is calledice fog, and it can become a serious problem in flightoperations.) This discussion is based on observations madeunder the direction of Ohtake (1967; see also Huffman1968) during the course of investigations of ice fog fromthis and other sources. As a result of these observations,enough is known about t8he may the contrail is formed tomake it, safe to state that' t'he contrail formed at cruisingaltitude is very unlikely t'o differ greatly from that formedon the ground.Ephemeral contrails are of no particular interest here,although a great deal can be learned a,bout conditions inthe atmosphere by observing their formation and disap-pearance; the discussion will therefore be limited t,o thoseformed while the aircraft is cruising in atmosphere at leastapproximately saturabed with respect to ice and at atemperature of approximat,ely -4OOC or lower. When thejet exhaust enters such an atmosphere, it is very rapidlycooled by expansion and turbulent mixing, and the watervapor condenses in tho form of very small droplets. Al-though the cooling is so rapid that levels of supersaturationat which homogeneous nucleation could take place shouldbe rea.ched, electron microscope photographs of thecrystals shorn no evidence that it ever happens. Thedroplet,s all appear to form on nuclei, which are abundantlyprovided by the exhaust products, even if not otherwisepresent,. The liquid droplet phase is very short-lived-ofthe order of a second-since the ambient atmosphere isbelow the spontmeous freezing threshold. The dropletsfreeze, and the result is a cloud composed mainly of thesmall, roughly spherical crystals called "droxtals" by theStanford group mho first reported them (Thuman andRobinson 1954). The time before the cloud is completelyfrozen is much too short to allow mass exchange t,o takeplace between liquid droplets or between liquid drops andice crystals, but there may be some crystal growth whilethe environment is between mater and ice saturation. Theend result seems to be dependent mainly upon ambienttemperature and little upon details of the may the exhaust146MONTHLY WEATHER REVlEWVol. 98, No. 10enters the atmosphere, the ice fog produced by an auto-mobile exhaust being practically identical to that from ajet engine. One therefore feels quite confident in using theresults of the ground observations to discuss the contrailformed at cruise altitude.Nearly 80 percent of the crystals collected near theairport runway had a diameter near 4 p. All collectionmethods are open t'o the suspicion of weighting in favorof larger particles; furthermore, very small particleswould not have been detected at, all. Ma.ny of the crystalscollected showed evidence of having ruptured a,nd emitteda spurt of water as they froze from the outside in (Bally-Dorsey effect, Bladchard 1951); this may have producedcrystals too small to be detected, though t'here is no evi-dence of this effect'. At, any rate, assuming sphericalshape and 2 p radius gives an uverttge reciprocal mass ofapproximately 3 X 1O'O droxt'als gm-', which is believedto be a very conservat,ive lower limit.3. CONTRAIL CRYSTALS AS FREEZING NUCLEIA cruising Boeing 727 then produces 3X loi3 crystalssec-l if all of the water rapor produced by the fuel com-bustion goes into crystal formation, as probably happensunder the stated conditions, or lOI3 if only a third of itdoes. (The writer feels that the act,ual crystal productionrate is probably at least IO'* ant1 possibly 10'5 see-'.This seems to be confirmed by the area and optical depthof the contrail.) Except for hail suppression operations,large-scale weat,her modification tl.ttempts currently inprogress employ ground-based generators to introducefreezing nuclei into the atmosphere. Under these condi-tions, silver iodide, the nucleating material most generallyused, is considered to deliver lOI3 freezing nuclei gm-'of silver iodide on the average. Under optimum conditions,it may deliver as many as 1015 nuclei gm"; under adverseconditions, it, may deliver none at all, since silver-iodidenuclei deteriorate rapidly in effectiveness after enteringthe atmosphere. It can be seen, then, that a medium-sized jet forming a persist'ent contrail is introducingnuclei in numbers equivalent to ground-based generatorsburning more than a gram per second of silver iodide.Since silver iodide is introduced with the object of formingice crystals from supercooled water droplets, it seemsobvious that the contrail ice crystals can produce anyeffect of which the silver-iodide nuclei are capable, andin addition can do things of which silver iodide is notlikely to be capable, such as extracting water from airsupersaturated over ice but under water saturation.(Silver iodide apparently can be effective in the absenceof water droplets by acting first as a condensation andlater as a freezing nucleus, but there seems to be somedoubt about how efficient it is in this.) It is then evidentthat, whenever and wherever persistent contrails arebeing produced, the airlines are conducting a seedingoperation in which each passing aircraft deposits nucleiin numbers at least comparable to a ground-based seedingattempt. The coverage is global in extent, and whilerandom in the sense that airline flights are not scheduledwith weather modification in mind, it is not really randomwith respect to meteorological conditions. Airline opera-tions are affected by the weather, and such things aspressure pattern navigation for instance could well intro-duce a significant bias. The formation of a persistentcontrail is itself a, function of ambient conditions thatwould seem to lead to the crystals being deposited pref-erent,ially in situations favorable to modifications of theweather being produced.4. MODIFICATION BY PROCESSES OTHER THAN SEEDINGThe contrail crystals are so small that they fall veryslowly in still air, even at jet cruising altitudes, and in thepresence of even slight updrafts will not fall at, all. Undersome conditions, it, seems likely that they can accumulatesufficiently to cause some high cloudiness along a busyairway even if they remain essentially unchanged in sizeand numbers after thq- are first introduced. This, how-ever, seems to be marginal, and it is unlikely that simpleaccumulation, even if accompanied by modest grou-t'h, iscausing much effect, at, present. There is, however, goodreason to believe that some mechanism exists that resultsin crystal multiplication.Careful observation of contrails shows that, in general,their behavior is consistent, wit.h t,he higher rates of crystalproduction given in the discussion above; but in manyinstances, the contrails attained dimensions that are verydifficult to explain by simple growth of the existing crystalsunless the numbers given here arc very gross under-estimates. The crystals in the contrails that. exhibitedthis behavior were obviously growing very rapidly, sincewell-defined fall streaks developed from them in less thanhalf an hour; so they were evidently in air highly super-saturated with respect to ice. (Natural air can becomevery much more highly supersaturated with respect toice than it ever does with respect to water because of thevery low content of nuclei capable of removing watervapor from the air once it is below water saturation.) It isalso obvious that, if the contrail c,rystals really did producethe complete sky coverage in the instances mentioned inthe introduction, some rapid and efficient process ofcrystal production must have been triggered. (It is quiteevident that systematic and quantitative observation ofcontrails can lead to a considerable increase in our knowl-edge of the meteorology of the 200- to 300-mb region andof cloud physics in general, in addition to the value ofsuch observations in connection with the subject of thispaper.)5. GENERAL DISCUSSION AND INFERENCESIt has been shown above that there is good reason tobelieve that aircraft contrails may be capable of con-siderable weather modification by producing cirrus cloudsat altitudes close to those at which the aircraft iscruising. It is obvious that the contrail crystals may in-October 1970 Wallace B. Murcray 747fluence cumulus development by getting into clouds thatextend to and beyond normal jet cruising levels. A littlereflection mill show that there is also good reason to believethat contrail crystal seeding may be producing all theeffects of which deliberate seeding with freezing nuclei iscapable. Crystals falling from cirrus clouds through whatappears to have been at least moderately dry air have beenobserved to survive for astonishing distances (Braham andSpyers-Duran 1967). The contrail crystals must grow consid-erably before they fall, but they are frequently seen to developrespectable fall velocities within a few minutes after theyare formed. Whether or not they are capable of survivinga long fall through dry air, it is likely that they are oftendeposited in air that, is saturated with respect to ice allthe way from t,he altit,ude of formation to the cloud bases.Under such conditions, they mill of course at least surviveto the melting level, and it is likely t,hat they will growand possibly multiply. Under such conditions, they canenter clouds anywhere above the freezing level andpossibly for some distance below. (It is difficult to seehow they could be injected into a cloud in a mannerthat mould exactly duplicate the conditions of a seedingflight through it, or a silver-iodide shell exploded inside,but, not, so difficult to see how t,he contrail cryst'als mightproduce essentially the same results.) It' is not intendedhere to undertake an exhaustjive discussion of the possibleeffects of cloud seeding by contrail crystals, but one, ortwo possibilities seem n-orth mentioning.The orographic cloud situation in which a wet airmassis being driven upward over a mountain range is con-sidered to be one of the more favorable for increasingprecipitation by ground-based silver-iodide seeding. Thissituation, however, leads t,o an airmass that is rising andcooling over a region that extends for a considerabledistance upwind and t,o altitudes well beyond t'hose atwhich jet airliners normally cruise. It is likely, therefore,that, the air dl be at or near saturation, at least over ice,from the cloud base up, and that a persistent contrailwill be formed that can and will seed the clouds. Whenone considers the number of airline flight,s over an areasuch as that west of Denver, Colo., one immediat,elywonders what the airlines may be doing t'o the statisticsused in evaluating the results of commercial seeding.(It should be noted that the writer is not contending thateither or both operations actually do increase precipitn-tion, he is merely pointing out that both should beconsidered.)The plains region of the United States is traversed byheavy air traffic; in the summer, it is also heavily populat~edby towering cumulus so that a jet aircraft is often flyingamong them. The flight path rarely goes through theclouds because of the turbulence and the likelihood ofencountering hail, but the whole region is likely to besaturated with respect to ice. The air is also likely t,o con-tain considerable numbers of natural ice crystals, so it isnot immediately obvious that the introduction of the con-403-235 0-704trail crystals can have any effect. However, while ridingan aircraft through this region, one sees numerous con-trails left by other flights, so it is evident that the naturalice crystal concentrations are much exceeded along theflight path and that this condition persists for some time.Schaefer (1969) mentions this in discussing inadvertentweather modifications. He implies the seeding capabilitiesbut does not discuss them specifically. The probabilitythat a developing thunderhead will entrain such a highice crystal concentration is thus increased; and a littlecalculation shows that, if the Russians and the French areright, it could lead t.o suppression of destructive hail.There seems little doubt that contrail crystal seedingsometimes disperses clouds in the same manner as seedingwith Dry Ice does. Over the years, many instances ofcloud dispersal by aircraft have been reported in Weather,and many very good photographs have been published.In most instances, not enough data are available to form afirm opinion of how the dispersal as accomplished, butt,he resemblance of the phot,ographs and the descriptions ofthe occurrences to t,hose of t,he results of seeding super-cooled wat'er clouds with Dry Ice is so striking that itseems safe to conclude t,hat most, if not all, of the reportsare examples of the seeding of clouds by contrail icecrystals. The possible consequences of this are consider-able; in fact, it seems probable that one of the projects formodifying the global climat,e discussed by Fletcher (1965),namely modification of the cloud cover over t'he NorthPolar Basin by cloud seeding, is already underway, al-t,hough the sca.le is still more modest t,han he envisioned.6. CONCLUSIONSThe conclusions reached as a result, of the abora dis-cussion are that, nearly all result's that can be produced byseeding with ice crystals are in fact, being produced as aresult of routine airline operations; that, the offoctivenessof ice cryst'als as nuclei over such a wide range of metero-logical conditions and the scale on which they are beingdeposit,ed make it, likely that they are affecting precip-it,ation t,o a much greater extent than are present delib-6rat.e seeding operations (although the deliberate op-erations may stsill show more net results because of theselectlion of situations suitable for seeding).It, cannot be concluded t,hat any net effects are neces-sarily being producnd on even local climat,es, although itseems certain that tho seeding is taking place, because thepossible result,^ of any kind of seeding are not 1%-ell under-staood at present. and because the very massiveness andfrequency of t,he seeding makes it likely that such a pro-fusion of effects in contradictory directions is being pro-duced that the net result may be small. The actualmagnitude and direction of the effect on the weather canonly be established by analysis of the records. The purposeof this paper is to show t,hat there is sound basis for theMention ofcommercial products does not constitute an endorsement.