«THE CLASSIFICATION OF SNOW METAMORPHISM By R. A. SOMMER FELD (U.S. Forest Service, R ocky M ountain Fores t and Range Experim ent Station, * Fort Co ...»
J fl lI l"1Ja! (.f G/aciolvg)', \ '01. g, :\"0. 55, [ 970
THE CLASSIFICATION OF SNOW METAMORPHISM
By R. A. SOMMER FELD
(U.S. Forest Service, R ocky M ountain Fores t and Range Experim ent Station, * Fort Co llins,
Co lorado 805~2I, U.S.A. )
and E. LA C HAP E L LE
(Depa rtm ent of Atmosph eri c Scien ces, Uni versity of Washington, Seattle, Washington 98 105,
.-\ BSTRACT. A n ew class ificat io n ot' snow o n th e ground is based on th e maj or physica l processes involved in th e m eta morphism of a snow cove r. T he m aj or divisions a re based on (1) the mecha nica l damage to snow crysta ls during precipi tation, (11 ) th e tra nsport of water va p or at constant temperatCl re b ecause of surfaceenergy differen ces, (Ill ) the tra nsport of water va por a long a th erm a l g ra di, nt, a nd (IV ) firnifi ca ti on because of me lting and refreezin g, a nd pressure consolid a ti on.
R ESUME. La claSJijica lioll dll ndtamorphisme de la neige. U ne nou vell e cl assifica ti on d e la neige a u sol es t basee SUI' les principa u x procesms phys iq ues du m etamorphi sm e d ' une couverture d e neige. L es divisions prin cipa les sont basees sur (I ) les d ommages m ecani=lues ca uses pal' le, cri, taux d e neige pe ndant la precipitation, (Il ) le tra nsport d e la va peur d 'eau it tempe ra ture constan te du a ux differences d'en ergi e superficieJl e, (Ill ) le transport d e vapeur cI'eau le long d ' un g racli ent th ermiqu e et (IV ) la formati on cle neve due it la fon te a u regel d e meme qu'a u tassement.
ZUSA MM E NFASSUNG. Die Klassijikatioll VOII ScllIleellletalllorplwse. Ei ne neue Kl assifikation von li egend em Schn ee geht von clen wichtigsten ph ys ika lisch en VOl'ginge n bei d el' M eta m orphose e iner Schn eecl ecke a us.
Di e H a uptglied erung beruht aut' ( I) d el' m ech a nisch en Besch ad igu ng d el' Schn eekristaJle beim N ied erschl ag, Ill ) clem T ranspo rt von W assercl a mpf bei konstanter Temperat ur infolge VJn U nterschieden in cl er OberAach enenergie, (Ill ) d em Transport von W asserclampf iangs ein es T emperaturgra diente n, un ci (I V ) d el' \ -erfirnung infolge von Sc h m c l z ~ n unci Wi ecl ergcfri cren sow ie d el' Druckve rfestigung.
[. DEFINITIONS :'I. certain amo unt of confusion exi sts in the classification of m etamorphosed snow. For exam pl e, th e term crystalline, whi ch is a precise, scientific term, is misused by many workers in the fi eld. A solid is crystalline ifits atom s are arranged in an ord erl y array. U nd er this d efinition, a ll ice found in nature is crystallin e w heth er it consists of snow-flakes, m etamorpho, ed snow. la ke ice, or ice cubes in a g lass. Thus, it is nonsense to speak of snow grains losi ng or ga ini ng crystalline ch aractc l'. A crystal may be rounded, partfy faceted, or face ted, accordin g to th e d egree to which the crys ta l faces are d eveloped.
Structure in the crystallographi c sense refers to the m olecu lar a rrangement in th e crys tal. In common usage structure also refers to the layeri ng or stratigraph y of th e snow cover. Because th e scale of these two usages is ve ry different, no confusion should res ul t. W e su ggest texL:lre to d esc ribe the relationships a mong th e snow crystals.
The snow cover is an aggregate of ice crystals, and its text ure is polycl)1stalline. Th e individ ual crystals in a pol yc rystalline mat eria l a re called grains a nd are separated b y grain boundaries. In snow, th e term grain is more loosely d efin ed as one of th e obvious sub-units whi ch m a ke up the snowpack. Thus, in speaking of snow, th e term grain does not m ean the same as the very specific term in metallograph y. Special tec hniques, which are too cumbersom e for use in th e fie ld, a re n ecessary to make the g ra in boundaries visi ble. The precise locations of the grain boundaries would give an unnecessal-il y d etailed d esc ripti on of th e snowpack for normal purposes. If the snow is no t bonded, the separatio n b etween th e g rains is real a nd the iden lificati on of separate grains is easy and precise. Once th e snow becomes bond ed into its usua l, complex, three-dim ensional network, th e placem ent of the boundaries between grains becomes increasing ly arbitrary and inexact with increasing d ensity. Furtherm orc, the exact
m ea surem ent o f the sizes o f the complicated shapcs of snow g rains is much too tedio us for a n y prac tical fi eld routine. T h ese distinctions do not h ave to be a bsolute, however, to be useful.
T he d escription of a mow laye r as small, medium, or large grained has obvious m eanin g a nd is important in characteri zing the condition of the snowpack.
The terms crystal and grain may be used interch a ngeabl y under certa in circumsta n ces.
T here are times when each gra in is a single crystal, as in new snow a nd in som e of th e stages o[ metamorphism. On th e other hand, a grain may be p olycrys talline, or a singl e crys tal may consist of more than on e grain (Fig. I). The latter possibility becom es comm on onl y [or ad van ced stages offirnifi cation (see below). Even w hen the crystals a nd grains a re ve ry nea rl y id en tical, the arbitrary b ound a ries o f th e grains may not correspond exa ctl y to the crysta ll ogra phic boundaries b etween crys ta ls. Unless the bound a ri es of the crys tals a re well d efi ned, as \\·hen th e crys tal faces a re well d eveloped, the m ore genera l tCl"m /:;rain sho uld be used.
Exi sting schem es for classifying sn ow on the g round are based on the shapes of th e sno w g ra ins. The "Internati on al cl assification for sn ow" (C anada. N ational R esearch Co un cil, J 954) uses grain size a nd th e presen ce or absen ce of facets a nd stepped surfaces to disti ng ui sh am on g snow typ es. Inad equa te referen ce in thi s classifi ca tion to the ph ysical p l'Ocesses resp onsible for these extern a l fea tures often leads to con fu sion a m ong snow observers. T he expl a nations of the processes a re sketchy and in som e cases in erro r. Th e use of p hrases like " los t its crys talline cha r ac ter " is unfortuna te.
Eugster ( 1952 ) divided snow metamorphism into the ca tegories: " d es tru cti ve", '"constructi ve", and " retrograd e". T he distinctions we re based on wh ether the process form ed or d estroyed surface features such as d endrites, face ts, a nd steps, but the emph asis \\'as on th e forms and not the processes.
Bo th the International a nd th e E ugster classifications a re essentiall y m o rphologica l. T he m aj or processes of sn ow m etamorphism are sufficiently well understood today that a simil a r cl assification sch em e for d eposited snow may b e fra m ed on generic grounds. W e fee l th at such a sch em e is hi ghly d esira bl e a t this stage in snow r esearch. I t will enable field wo rkers to understa nd b etter th e ch an ges occurring in th e sn ow, a nd to communi cate their findin gs mOI'e easily to other wo rkers both in snow r esearch a nd in related fi elds.
3. T H E PH YSICA L BASES OF TH E CLASS IFICAT IONS
L'mnetam01phosed snow In our classification system (sce p. 9-17 ), we idealize certa in conditions which strong ly a ffect the character of th e sn ow cover. An y ac tual snow cover is influen ced by som e mixt ure o f these idealized conditions. F or instance, the first category " unm etamorphosed SI1 0 \ \'" has th e di visions " n o wind a ction " a nd " wind bl own ". N ature does not sh a rpl y di vid e these TH E C LA SS IFI CA TI ON OF S:;OW ;VI ETAMORP II1 5 M 5 ex tremes. Snow is deposited und er a lmost any possibl e wind speed, 50 a choi ce, invo lving considerabl e judgem ent, must b e made about the importan ce o f wind ac ti on on an y given snow
la ye r. The distinctions in this category are governed by a stri ctl y mechani cal criterion:
whether or not the origina l snow crys tals were badly broken 01' d eposited in a relatively undisturbed fashion. W e recomm end th e use of Magono a nd Lee's ( 1966) classifications for th e sub-categories und er " unm etamorphosed snow " wit h " no wind action ".
The classification "surface hoar" is used to cover direc t d eposition o f water vapor fr oIll th e atmosphere on to the sn ow surface. Surface h oar and d epth hoar sh are similariti es offo rm, but surface hoar tends to d evelop as Rat plates instead of th e three-dimensional d evelopmen t characteri sti c of d epth hoar. Once incorporated within th e sno w cover, surface hoar te nds to resembl e d epth hoar and foll ows a similar metamorphi c patt ern.
Hqlli-temperatllTe metamolphism W e have divid ed th e important metamorphic proce ·ses in bdow-fi'eezing, seasona l snow cove r between "equi-tempera ture m etamorphism " an d ·' temperature-gradient meta morphism". These terms a pproximately correspond to the commonly used terms. contructive and d es tructive m etamorphisms. Because the current usage is not cl earl y d efin ed, we propose new terms to avoid confusion. T he id eali zed cond itions, of equal temperat ure throughout the sub-freezing snow cove r, or of a constant temperature gra di ent, do no t exist in nature; but by understan ding what wou ld h appen und el· idea li zed conditi ons, we can better unders tand w hat happens under the mixture of na tural co nditions. Act ual classificat ion should be based on th e history of th e snow cover if it is known.
W e distinguish between equi-temperature and temperature-g ra dient meta morphism because different physi cal processes are important under the two id ea lized condition s. Under equi-temperat ure conditions, th e wa ter m olecul es m ove, la l'gely by vapor diffusion (l-Iobbs a nd Mason, 1964; Hobbs and Radke, 1967), to new positions to d ecrease the sUlface free energy.
A sys tem wi ll change to a state of lovve r free energy if it can, and th e lowest possible energy tate of any given sys tem is its equilibrium state.
W e know that the surface energy is proportional 10 surface area in a given mass. Snow ta il s fi·o m the sky with man y d endrites, needl es, etc., whi ch have very large rati os of surface a rea to mass, and thus hi gh surface free en ergy per unit mass. If snow crys tals are broken by strong wind action, the res ulting small fragm ents, sha rds, a nd splinters ha ve even hi gher total sUlface fi'ee energy for th e same mass of materi a l. During equi-t emperature m etam orphism there is a continuous d ecrea se of surface free energy. Th erefore, 11 0 basis exi sts for absolute divisio ns, a nd our divi sions mark arbitrary points in th e d ec rease of surface free energy.
An important point is th a t both tim e a nd temperature arc major fac tors in d eterll1ining the stage of m etamorphism. If th e sno w is very co ld, it wi ll change ve ry slow ly; but if it is close to th e freezing point, it can change ve ry rapidl y. Age a lone will no t d e termin e the stage o f metamorphism.
The first manifes tati on of equi-temperature metam orphi sm (stage [J. A. I ) is th e d ec rease o f c urvature of th e sharp poin ts. This ca uses a distin ct rounding and a loss of the fin e ma rkings on snow crystal s (Fig. 2). This change is well illustra ted by th e carefu l wo rk of Bad er an d
6 J OURNA L OF GLAC I OLOGYothers ( 1939) and Yosida ( 1955 ). In badly shattered, blown snow, this process may proceed very rapidly, even as the snow is being d eposited. Thus blown snow may appear to skip thi first step. A little later, but still within the sam e step, the small necks disappear, causing th e snow crys tal to lose most of its complicated shape and to break up into small er grains with less total surface area (see Fig. 2). (The Eugster classification fail s to recognize this process. ) W e have d efin ed a grain as th e obvious sub-unit in the pack; during the first stage this m eans a snow crystal or a sno\\' crystal fragm en t. When t h e small n ecks are d estroyed, th e grai n size d ecreases and the number of grains increases.
During the next ste p in the metamorphic process (11. A. 2), the to tal surface a rea is furth el' d ecreased a s the grains becom e more equi-dim ensional. Splinters and n eedl es becom e shorter and wider ; plates becom e thi cker whil e their maj or diam e ters d ecreasc. These first stages a re equivalent to Eugster 's ( 1952 ) " d estructive m etamorphism ".
For grains of the sam e sha pe, th e smaller the size th e la rger th e ratio of surface area to mass.
T herefore, surface energy d ecreases further as th e smaller grains disappear and the larger grains grow, leading to an increase in grain size an d a d ecrea se in grain number. T his proces is not as striking as the ro unding of th e grains, and while it occurs a t earlier stages (arrows in F ig. 2), it is very difficult to see until the grains have fairly unifol'm shapes. W e have used its apparent onse t to mark a n ew step (11. B. I).
Th e various forms associa ted with d epth hoar, such as th e surface steps and the latti ce and cup crystals, also result in a sta te of relati vely high surface en ergy. Because of the large size of th e crystals, the total surface en ergy is probabl y sm a ll er than that of a corresponding mass of fin e-grai n ed snow. The large crystals can furth er reduce th eir surface energy by losing their surfa ce fea tures a nd b ecoming m ore equi-dim ension a l. If th e temperatul'e gradient necessary for the formation of d epth hoar (see n ext section ) is removed, th e sn ow will undergo equitemperature m etamorphism, eventually losing many of its facets and o ther surface features.
As the grains b ecom e more equi-dimensional and larger, the p ossible surface energy reduction available for furth er change d ecreases, and the process continually slows d own (ass uming constant temperature). Therefore, there is a practical li mit to the size the grains can atta in ( ~ 1 mm ) under sub-freezing a nd stri ctly equi-temperature m etam orphism in th e seasonal snow cover. Face tin g visible at thi s stage may be evidence of slight tempera ture gradi ents.
T emperatllre-gradient metamorphism T h e physical basis for the next major category, whi ch is equi va len t to E ugster's " con stru cti ve metamorphism ", is th e process of vapor transport a long a strong thel'mal gradi ent.