Annals rifGlaciology 19 1994 © International Glaciological Society Bent icicles and spikes N. MAENO, instilule of Low T emperalure Science, Hokkaido University, Sa/J/Jo ro 060, ] aj}(1n L. MAKKONEN , Laborat01) of Structural Engineering, T echnical Research Center of Finland, 02150 Espoo, Finland T. T AK AHASHl Hokkaido University of Education , Sapporo 002, ] a/Jan ABSTRACT. Formati on mec ha ni sms of bent icicl es a nd thin spikes were studi ed by observatio ns o f na tura l icicl es a nd by a rtifi cia ll y growin g icicl es in a co ld labo ra tory, a nd th e res ults a re disc ussed with reference to th e ge neral growth process o f icicl es . It is shown th a t icicl es can be bent by th e displ ace ment o f th eir roo ts, cha nge in water suppl y ra tes, and wind . Spikes were observed to g row by th e freez in g ofa wa ter column captured at th e centre of icicles a fter th eir length a nd di a m eter growth stopped. INTRODUCTION Th e physical process of icicl e growth has been in ves tigated extensively by G eer (1981 ), Maeno a nd T a ka hashi (1984a, b ), M a kkon en (1988 ) a nd M ae no a nd oth ers (in press). However , th ere rem ain several importa nt phenom ena, whi ch were no ted in th eir studi es but have not bee n surveyed in d etail. Maeno a nd Taka has hi ( 1984b ) tri ed to clarify mechanisms for th e bending of icicl es a nd for th e formation of thin spikes often found on th e surface of a n icicl e. This pa per d escribes furth er studi es o n bent icicl es a nd spik es . BENT ICICLES Ben ti cicl es a re often o bserved to grow from roofs a nd trees (Fig. I). Maeno a nd Taka has hi ( 1984 b) showed th a t a t least three fac tors can result in th e form a tion of ben t icicles . Th e first is an y inclina ti on of roo ts of icicles, e.g. b y th e creeping d eform a tion of ice/snow on a roof edge or b y bending of a tree bra nch from whi ch icicl es a re growing. Su ch ben t icicl es are often obse rved to grow fro m trees exposed to spray a nd waves of ri ve r, la ke 0 1" sea wa ter. Th e second is th e cha nge o f wa ter suppl y conditi ons during icicle growth . A d ecrease in water supply ra te by a n y meteorological cause will lead to a vari ation of waterflow pattern ; a single fl ow pa th on th e icicle wall will res ul t in a ben t icicle, a nd two fl ow pa th s, forked icicl es . In o ur icicle growth experim ents th e a bove two types of ben t icicles could easily be p rod uced . Th e third fac tor to bend icicles is th e effec t of wind. To study various effec ts of wind on icicle form a ti on, a n icicl e growth d evi ce was constru cted in a co ld la boratory, th e 138 tempera ture of whi ch was va ri ed between 0 a nd - 30°C . Th e d evice was simil a r to th a t used by M ae no a nd T a kahas hi (1984a ) a nd M ae no a nd o th ers (in press ), in whi ch icicl es could be grown from a horizonta l meta l ba r wrapped by wet gauze, to whi ch pure wa ter just a bove th e melting tempera ture of ice was suppli ed a t a co nsta nt ra te betwee n 0 a nd 50 mg s I . Wind speed co uld be vari ed between 0 a nd 6.0 m S I. During th e growth process th e icicl e's leng th a nd di am eter were measured , a nd rela tive humidity, mainta in ed a t a bou t 80 % , was a lso monito red. It was found th a t wind affec ts icicl e growth in two ways: it in creases th e h ea t excha nge and cha nges the wa ter fl ow a t the wa ll a nd tip. In our experiments th e la tter effect was mo re impo rtant tha n th e fo rmer ; it i observed th a t th e horizon ta l cross-sec ti on of a n icicl e grown in wind y co nditions becom es "anti-streamlined " Fig. 1. N atural bent icicles growing from a roo]. All the icicles are bent loward the righhand side by wind. The horizontal width of the photograph is 2.5 m. Mamo and others: Benl icicles and spikes 80 on ly slig htl y with water suppl y rate a nd is kept consta nt a t 2.44 mm in ave rage irresp ective of temperature and th e R ey nold s number (R e = 2ParU f ry , ry: d ynami c viscosity of a ir) . R e is roughl y in th e ra nge of 350 to 3500 in th e expe rim ental conditions, when U = I 10ms 1. The drag coeffi cient is a fun ction of R eynolds number a nd can be empiri ca ll y ex pressed as follows (Morsi and Alexander, I 0/ Cl) / ID / ~ 60 / / Cl ID / / "'0 I / / / ID Cl 40 )3/ C ctI "'0 c 20 / .- / "" 19 72 ) : CD = 98.33Re- " Q) - 2778Rc- 2 + 0.3644 (2) for 100 < R e < 1000, and / CD = 148.62Re- 1 g!eO"O en t - 4.75 x 104Re- 2 + 0.357 (3) I rTfRl';;U; 0 0 246 Wind speed 8 10 -I ms Fig. 2. Nfeasured bend angle versus wind speed. 0: present results ( temperature around - 9°C and water-su/Jply ra le around 10 mg s \). 0 : dala from M akkonen and Fujii ( 1993). The dashed and solid curves refer respectively to the empirical (rP = 0.684rP+ 0.140+ 4.0 degrees, 0 in 111S- 1 and calculated (Equation (1 ) ) bend angles. with th e sharper end windward, the lee end becoming blunter a nd thi cker du e to th e more ac tive growth of ice. The faster growth is attributed to the wind-driven film Oow of water, which is coo led effecti vely by th e wind , toward s leewa rd at th e wall, leading to an ti-streamlined or Oa t icicl es . l\1i croscop e observation of thin crosssec tion of these icicles und er pol a rized light showed th a t crystal g rains a re generall y mu ch sma ll er and contain more tin y a ir bubbl es at the lee, suggesting faster freezing of water. "Vh en wind was sta rted a fter some growth of an icicle , mostly 20 cm, th e icicle began to bend sha rpl y. The angle of bend, measured from vertical, was found to increase with th e wind sp eed in th e ra nge studi ed ; in th e meas urement th e air temperature a nd water supply rate were ma in tained at abou t - 9°e a nd 10 mg s- \ res pecti vely. Figure 2 gives th e m easured bend a ngle plotted aga inst wind speed together with th e result by Makkon en a nd Fujii ( 1993), who grew icicles on a horizontal wire in an icing wind tunnel. In the fi gure th e d ash ed lin e is a curve fitted to th e m eas ured d ata. Th e general increase in bend a ng le with wind speed is r easo nab le beca use an aerod yna mi c force acts on th e pendant drop a t th e tip where th e icicl e is g rowing lengthwi e. rf we ass um e th at a spherical pend a nt drop (radius r ) is in a mech a ni cal ba la nce between gravity and drag force due to a steady wind (hori zo nta l speed U ), a force balance equation leads a pproxim a tely to (1) wh ere c/J is the bend a ngle, Pa a nd Pw a re th e d ensities of air a nd water respectively , CD is th e drag coeffi cient, a nd 9 is the acce leration o f g rav ity. According to th e o bserva ti on of Maeno a nd oth ers (in press), r increases for 1000 < R e < 5000. Th e solid lin e in Figure 2 gives th e bend a ng le of a n icicl e at va rio us wind speed s calcul a ted by Eq ua tion ( I) togeth er with Eq uations (2) a nd (3) . Equ a tion ( I ) can explain on ly th e genera l tend ency of in crease in th e measured bend angle with wind speed. The m agn itud es of th e meas ured a ng les a re much la rger than those calc ul a ted , impl ying that th e bend of icicles is ca used by wind ac ti on but not onl y through the simple mechanical aerodynamical force actin g on a pend a nt drop given by Equation ( I) . In wind y co nditions, part of wind -dri"en flow of effec tively coo led water freezes a t th e leewa l-d wa ll formin g a n anti -streamlin ed icicle, as stated above, and th e res t ([nally enters into a pend a nt drop preferentially a t the lee, which leads to faster g r o wth there a nd co nseq uentl y a llows th e icicl e to bend. At this stage of study, howeve r, the d eta iled flowin g and freezing process can no t be fo rmul a ted reasonab ly, because more quantita ti" e info rm a ti o n is need ed on sevel-a l important pmcesses involved such as rh e a ir Oow a round a n icicle which controls th e water flow a nd heat exch a nge, th e viscous flow of a thin , rou ghl y 100 j1m , water film in which radial (hori zonta l) temperature a nd velocity di stributions are se t up, a nd so on. FORMA TION MECHANISM OF SPIKES The photographs in Fig ure 3 show na tural icicles with spikes. Spikes a loe thin ice needl e. typicall y I mm in di amete r and 5 to 20 mm in leng th. Some, however , are mu ch thicker and shorter. Under th e pol a ri zed light, microscope observations were mad e of spikes prod L1 ced on na tura l and a rtifi ciall y form ed icicl es . Th e sp ikes were fo und to be composed of two or a few sing le crys tals. Th e growt h direct io n of spikes is ra nd o m a nd shows no sys temat ic rela ti on to that of main icicl e g rowth. Som e icicl es have several spikes a nd some none, d ependin g on how their g rowth stops , as ex pl ained below . I t was found from observations of natura l icicles a nd tho. e formed in a co ld labo rato r y that spikes a re no t a peculiar or unu sua l ph eome non. In o ur icicle form a tion experim ent most icicles were observed to produce sp ikes a fter th e cessa tion of th eir grow th of leng th a nd dia meter and th e su bseq uen t slow freezing of a wa ter column capt u red at the ce n tre of eac h icicle. Th ese spik e production eve nts so metim es took p lace a few hours a ft er th e ove ra ll growth of a n icicl e depending on its growth history , size a nd su rro unding temperature . 139 Gold: Elastic modulus oJfresh-water ice confin ed with a th in ice shell after the cessation of growth. It was noticed that spikes a re more frequently form ed at the tip than near th e root, a nd in th e late afternoon when most icicles stop growth by th e relative decrease in the water supply rate a nd increase in the freezing rate. The formation mecha nism of spikes on icicles is esse ntially the same as th e growth of tubu lar or columnar ice, which develops from an ice layer covering supercooled water (Dorsey, 1921 ; Bally, 1935; Blanchard , 195 1; Hayward , 1966; Kra usz a nd others, 1967; Tusim a a nd Suzuki, 1972; Wascher, 1991 ). ACKNOWLEDGEMENTS The work was partly supported by G rants-in-Aid for R esearch in Natural Disaster, Co-operative R esearch a nd Scientific R esearch of the Ministr y of Education , Science a nd Culture. The authors th a nk a nonymous referees for their helpfu l comm ents. Fig. 3. Natuml icicles with spikes. The vertical length the jJhotogmph is 180 mm. of REFERENCES As shown by M ae n o and Takahashi ( 1984a ) , Makkonen (1988) and Maeno and oth ers (in press), an icicl e grows by a simu ltaneous freezing of water at the tip a nd wall , the growth being maintained by a balance between the suppl y rate and freezing rate of water. When the balance is d estroyed a nd dripping of water drops at the tip ceases, the pendant drop begins to freeze from outside forming a thin ice shell . As freezing proceeds the press ure inside increases steadil y. Since th e pendant drop is connected to a water cylind er, 5 mm in diameter and a rew cm in length , in the central part of a n icicle, its subsequ ent freezin g produ ces a high internal press ure which squ eezes out unfrozen water from small holes a nd cracks. Water exiting from suc h holes and cracks ra pidl y freezes to form spikes. Th us various forms of spikes can be produ ced in various directions depending on the mode of water sq ueezing a nd freezing. When a n icicle stops growi ng, the water fi lm on th e wall also begins to freeze from th e outer su rface which is in contact with th e cold air. In this case, only very fine needle spikes may be form ed since the volume of confined wa ter is small so that the pres ure produced is also small. Spikes are so thin that they easily d isappear by melting or sublimation . This may be th e reason why spikes have not been frequently noticed , although th ey shou ld form in every icicle when unfrozen water is 140 Bally, O. 1935 . Ober ein e cigenanige Eiskrystallbildung. H elD. Chim . Acta, 18, 475-476. Bla nchard, D. 1951. A verifi cation or th e Ba ll y- Dorsey theory or spicu le formation on sleet pellets. J. Meteorol., 8, 268- 269. Dorsey, H. G. 192 1. Pec uliar ice formations. Phys. Rev., 17, 162- 164. Geer, I. v·.,. 198 1. The not-so-ord ina ry icicle. Weatherwise, 34 (6), 257259. H ayward , A. T.] . 1966. Growth of ice tubes. Nature, 211 , 172 173. Krausz, A. S., B. H a rron and G. G. Li tva n. 1967. Tubular ice crystals. Nature, 215 (5098 ), 27 I - 273. Maeno, N . and T. Takahashi . 1984a. Studies on icicles. l. General aspects of the structure and growth of an icicle. Low T emp. Sci. , Sa. A 43 , 125- 138. [In J apanese wi th English summ ary.] Maeno , N. and T. T akahas hi . 1984b. Stud ies on icicles. 11. Wave-forms, spik es a nd bent icicles. Low T emp. Sci. , Ser. A 43 , 139- 147. [ In J a panese with English summ a ry.l Maeno, N ., L. :v1akkonen, K. Nishimura, K. K os ugi a nd T. T aka has hi . In press. Growth ra tes of icicles. J. Glacial. Mak kollen, L. 1988. A model of icicle growth. J. Glacial., 34 ( 11 6), 6 70. Ma kkonen , L. a nd Y. Fujii. 1993. Spacing of icicl es. Cold Reg. Sci. T u /wo!. , 21 (3), 3 17- 322. Morsi, S. A. and A.J. Alexander. 1972. An in vestigation of part icl e trajectories in two-phase Oow system. J. Fluid M ech., 55 (2), 193- 208. Tusima , K . and S. Suzuki. 1972. Column a r ice protruded to the air. L ow Temp. Sci., Ser. A 30, 23- 33. [I n J apa nese with English summa ry.] Wascher, T. 1991. Generation of sla nted gas-filled icicles. J. Cryst. Growth, 110, 942-949. The accuracy of references in the text and in this list is the responsibility of the aulhors, to whom queries should be addressed.
© Copyright 2026 Paperzz