J ournal oJGlaciology, Vol. 41, No. 137, 1995
Variations of ablation, albedo and energy balance at the
tnargin of the Greenland ice sheet, Kronprins Christian
Land, eastern north Greenland
THOMAS KO NZ ELMAKN
Department oJ Geograph)l, Swiss Federal i nstitule of T echnology, CH-8057 Zurich, Switzedand
ROGER
J.
BRAITHWAITE
Grrmlancis Geologiske Unciers@gelse, DK-1350 Kf) benhavn K, D enmark
ABSTRACT. A meteorological a nd glaciological experiment was carried o ut in
July 1993 a t the ma rgin of th e. Greenl and ice sh ee t in KrOl"))Jrins. Christia n L a nd ,
eas tern north Gree nl a nd. \Vlthm a small area (a bo ut 100 m- ) d ally measurements
were made on ten a bl a tion stakes fi xed in " light" a nd " dark " ice and were compa red
to eac h oth er. Simulta neo usly, the components of the energy bala n ce, including n et
radia tion, se nsibl e-h ea t flu x, la tent-hea t flu x a nd conductive-hea t flu x in the ice were
d etermin ed . Gl o b al radiation , lo ngwave in co ming radia tion a nd albedo wer e
measured, a nd longwave outgoing radi a tion was calculated by ass uming th a t the
glacier surface was melting. Sensible- a nd la tent-h ea t f1u xes were calcula ted from a ir
tempera ture, humidity and wind sp eed. Condu c tive-hea t flu x in th e ice was es tim a ted
b y tempera ture-profil e meas urem ents in th e upperm os t ice la yer. N e t radia tion is the
m aj or source of a bl a tion energy, a nd turbulcnt flu xes are smaller en ergy so urces by
a bout three tim es, whil e hea t flu x into the ice is a substanti al hea t sink, redu cing
energy ava il a bl e fo r ice melt. Albedo vari es fro m 0 .42 to 0.56 within th e ex perim enta l
site a nd ca uses rela ti vel y large differences in a bl a ti o n a t stakes close to each o ther.
Sm all-scale albed o variations sh o uld th erefore b e carefull y sampled for large-scale
energy-bala n ce calcula tions.
INTRODUCTION
An a bla tion- clima te stud y was m ad e a t the ma rgin of th e
Greenl a nd ice shee t in Kronprins C hristi an La nd , eas tern
north Greenl a nd (Fig. I). The stud y was part of a 2 year
progra mme on world sea-level ch a nges supported by the
European Community. The specific obj ective of th e stud y
was to collec t d a ta on clima te, a bl a ti on, radia tion a nd ice
tempera ture to estim ate th e so urces and sinks of th e
energy bal ance during the period of meas urements (8- 27
J ul y 1993) a nd to com pare condi tions wi th those found in
W es t Greenland (Braithwaite a nd Olesen, 1989 , 1990).
The fi eldwork was carried o ut a t two study sites. One
was located o n the tundra in front of the ice sh ee t (base
camp ). Th e o ther was establish ed o n the ice sheet (glacier
sta ti on). Th e glacier sta ti on was loca ted a bout 300 m
inla nd of th e ice m argin a t a n eleva tion about 50 m a bove
th e ground-level a t the edge . Th e si te was well-ex posed
with out ma rked slope, i. e. it represents a large a rea close
to the ice-shee t margin. Th e surrounding a rea was
covered by hummocks about 0. 5 m in height a nd 5 m in
wa velength , i. e. th e surface to pogra ph y is rougher th an
th e site in W es t Greenland (Camp IV; Fig. I ), wh ere
Ambach (196 3) m ade energy-b a la nce studies. Th e glacier
sta ti on was a wa lk of only 10 min from th e base camp a nd
174
was vi sited seve ral tim es d a il y during th e experim ent. Th e
position of th e glacie r station was d e termined by th e
g lo b a l p os iti onin g sys tem (GPS ) to b e la titud e
79°54'43" N and longitud e 24°04'25" W a t an altitud e of
380 m a.s .l. Abla tion was determin ed as the ave rage of
readings a t ten sta kes drilled within a n a rea of a bo ut
100 m 2 . T wo th ermi stor strings were also insta ll ed to
m eas ure tempera ture g radients in the top 3 m of the
glacie r surface. At the sa me tim e as th e a bl a ti on measurements, meteorological d a ta (air temperature, rela tive
humidity and wind sp eed ) were obtained once-hourl y
a nd r ecord ed by data loggers at base camp and at the
glacier sta tion . Radi a ti o n conditions were studied by
co ntinuous logging of global radiation a nd all-wa ve
(sho rt- a nd longwave ) incoming radi a tion at th e base
camp , while albedo was meas ured a t the glacier sta tio n .
"Yea ther conditions were ra ther sta ble throughout th e
whole field period with nearl y co ntinuous sunshine, liule
cl o ud a mount, nearly consta nt tempera tures (daily means
of 3- 6°C) and strong wind s from th e ice sheet (d aily
mean s of 3-8m s I) . Cloud obse rva tions (amount a nd
type ) were mad e six tim es a day using World M eteorologica l Organiza tion cl assification sch emes. The interior
of Kronprins Christi a n L a nd is obvio usly drier and has
more sunshine tha n the coast in general, e.g. Sta tion ord
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
Konzelmann and Braithwaite : Ablation, albedo and energy balance in Kronprins Christian Land
climatic data using profile-aerodynamic formul ae (Ambach, 1986 ) , thus saving the weight of heavy instrum ents.
By analogy with preliminary geological surveys, such an
a pproach m ay be termed a reco nnaissance energybalance stud y and it is hoped th a t the present results
legitimate the approach.
ABLATION
S0NORE
OAMANARSSUP SERMIA
cf/
• NOROBOGlETSCHER
Fig . 1. Location of glaciological studies
ice sheet referred to in the text.
011
the Greenland
is no to ri o us for fog and late-lying snow. The fine weather
con ditions durin g th e p eriod of observa ti on were
proba bl y not co mpletely typical as there was a stro ng
a nticyclo ne over Greenland for most of the field period
(Gcb and Naujokat, 1993 ) .
Ambac h (1963) showed for the upper a blat ion area
(Camp IV ) that the ener gy gain at the s urface in summ er
is ca used mainl y by the radiative flu xes as the sensibl eand latent-heat flux es almost cancel each o th er. At ETH
Camp (Ohmura a nd others, 1991 , 1992 ; Fig. 1), which is
located cl ose to th e m ean equilibrium-lin e altitud e
(ELA ) , n et radi a tion provides almost a ll of the abla tio n
energy (Ohmura a nd oth ers, in press ) . R adiation was a lso
expec ted to be the main so urce of a bla tion energy a t the
present stud y site. In th e lower ablation area of a glacier
or a n ice sheet, la rge varia tions of th e surface reflectivity
(albed o ) can be expec ted. This is du e mainly to different
ice conditions, e.g. age, surface undulation , conta mination, slope and orientation. Th ese variations in albedo
a ffec t the energy balance and th e a bla tion ra te, and th e
relation between sm a ll -scale va ri a ti ons of a bla tion and
albedo is treated as thc m a in poi nt of th e prese nt paper.
The sma ll er, less importa nt energy-b a la nce compo nents, such as turbulent Duxes a nd condu ctive-heat flux in
the ice, were evalu ated b y methods req uiring only lig h tweight eq uipment. Morris (1989) sugges ted tha t acc uracy
in es tim a tin g energy-b a lance co mpon ents should be
co mmensurate with th eir relative importa nce in the
overall energy ba la n ce. W e ex tend this co nce pt b y
replaci ng "acc uracy" with " logistic cos t". For example,
the turbul ent components were es tim a ted from simple
Previous experience has shown that ab latio n measurements in vo lve considerab le error and tha t abla tion itself
varies greatly even on a scale of metres (Ol esen and
Braithwaite, 1989 ), suggesting that many m eas uremen ts
are needed to obtain a representative value for a site.
The ten stakes were read daily close to 1900 h local
time (about 171 5h solar time ) . As stakes are usua ll y
surround ed b y an "ab la ti on hollow" of 0.1 ~ 0.3m
di ameter, stake readings wer e made by meas uring fi'om the
top of each sta ke to a str aight ed ge laid on the upstream
sid e of th e stake which defines an ave rage level of the
su rroundin g surface. Differe nces in successive daily
readings are conve rted into abla tion valu es ass umin g an
ice density of900kgm 3. We find it conven ient to treat
ablation as a positive rath er than negative quantity as
recomm end ed by Anon. ( 1969 ) to avoid the clumsiness of
having to say tha t (negati \"e ) a blation d ecreases with
in creasing availa bility of energy.
Th e stakes were placed so that fi ve were loca ted in
"dark" areas a nd fi ve in " light" areas . Th ese were
sec ti ons of the ice surface whi ch had " d ark" a nd " light"
co lor, respec ti\'ely, based o n visual inspectio n at the
beginning of the fi eld wo rk . The difference between the
two kind s of surface a ppea rs to be whether d ebris is
spread evenl y over the surface (" d ark " ) or co ncentrated
into a few deep cryco nite h oles surround ed by ver y white
ice (" light" ) .
Abl ation vari ations from d ay-to-d ay and between
stakes were a nalysed by the simple linear model of
Lliboutry ( 1974):
YjI
=
Qj
+ (31 + Ejt ,
(1)
where Yjl is the ablation at stake j on da y t, Qj is th e mean
ablation at stake j, (31 is the ab lation deviatio n which is
the same for a ll stakes on d ay t, a nd Ejt is the error in the
model which is ass umed to be random. Measurement
errors con tri bu te to th e error term as long as th ey are
rand om bu t if they are syste m atic, i.e. sh ared by a ll
stakes, th ey contribute to th e abla tion d eviation . Accord ing to Bra ithwaite and O lese n ( 1989), readings at the
indi vidua l st a kes should b e a pproxim a t e ly equall y
correlated w i th the a bla tion d ev iation (3jl . At ab la ti on
stake B, howeve r, the correla tion coeffi cient between th e
ablation d evia tion and d ata is onl y 0. 36 while co rrelati ons
Unfortunately,
for oth er stakes are in the range 0 . 83 ~ 0.95
stake B was loca ted on a steep slope which made it
difficul t to use the straigh t -ed ge method. Beca use of
inconsisten cy with the other stakes, data from that stake
are dropped fi'o m any furth er analysis. Th e ablation
va ri a tions a t the remaining nin e stakes a r e shown in
T able 1. Th e error in th e lin ea r mod el for th e nin e-sta ke,
20 d matrix is surprisingly low with a sta nd a rd deviation
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
175
Journal of Glaciology
Table 1. D aily ablation at nine stakes ( A- J, excluding B ) on the margin of the Greenland ice sheet, Kronprins Christian
2
I
Land, J uly 1993. Units are kg 111 cl
Date
A
C
D
E
F
G
H
1
]
M ean
s.d.
July 1993
8
9
10
I1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
36
58
58
40
45
3I
58
63
58
45
45
58
58
54
45
50
36
31
23
36
40
63
58
31
54
31
40
54
58
31
45
50
58
54
40
54
36
27
31
36
31
54
40
40
27
27
50
50
54
40
45
54
54
50
45
54
31
45
18
31
31
50
45
36
27
31
45
58
40
31
45
50
50
50
36
54
27
27
31
40
31
50
45
27
27
27
31
68
40
31
31
50
54
50
40
50
31
18
18
40
36
50
54
36
31
31
50
63
45
40
45
45
50
45
36
45
31
31
27
31
23
45
45
27
27
27
40
50
36
40
36
50
45
45
40
45
31
31
27
31
31
45
45
27
23
23
40
54
40
27
36
45
50
36
36
50
23
27
18
27
31
58
45
31
36
27
45
63
45
40
50
50
54
54
45
50
36
3I
31
31
32
53
48
33
33
28
44
58
46
36
42
50
53
49
40
50
31
30
25
34
M ea n
s. d.
46
12
45
12
42
40
10
38
13
41
10
37
9
35
43
II
II
41
10
II
of only ± 5 kg m 2 d I which is certainl y small co mpared
with the range ± 13 to ± 19 kg m 2 d- I qu oted by
Braithwaite (1985, p.21 - 22 ).
Thc hypothesis th at the four " d a rk " stakes (A, C, D
a nd G ) have the sa me mea n abla tion as the fi ve " lig ht "
stakes (E , F, H , I and J ) is tested with the Studcnt t test
(Kreyszig, 1968 , p. 219 ) . The mean a nd stand ard deviations for the two data se ts a re 44 ± 9 kg m- 2 d- I a nd
39 ± 10kgm- 2 d- l , respec tively. Th e m ean valu es a re
different (a t 5% sig nifi cance leve l) a nd it ca n b e
conclud ed that " da rk " stakes have sig nificantl y greater
a blatio n than " light" stakes. Th e time series of a bla tion
energy for " lig'ht" a nd " dark" sta kes a re presented in
Figure 2.
Th c glacier surface co nsisted almost always of a 2030 mm thi ck "wea thering crust" (MU ller and K eeler ,
1969) with a relatively low density. The effect of this crust
on the eva lu a tion of ab lation deserves discussion . The
abla ti on in a period is the loss of ma teri al in the surface
layer of thickness h relative to a stake fi xed in the ice. Th e
total surface-layer thickn ess h consists of a layer of ice of
thickn ess hi a nd a layer of crust of thickness he. Th e
abla ti o n is given by:
5
6
7
5
10
3
8
7
8
6
6
4
4
6
4
4
4
7
6
5
a blation is evalua ted using only the first term with an ice
density of 900 kg m 3 which refers to slightly bubbly
glacier ice . Some attempts were mad e in 1986 at
Q a ma narssllp sermi a Lo m easure the d en sity of crust
with a m iniature co rer (unpublished data from R.J.
250
1--0---
light
dark
200
E
:;:
>-
e>
150
Q)
cQ)
c
o 100
~
:0
«
50
(2)
Julian day
wherc 11h is the cha nge in surface-layer thickness, 11hc is
the ch a nge in crust thickness, and Pc a nd Pi a re th e
densities of crust a nd ice, respectively . As the depth of
crust is not observed a nd its density is unknown , the
Fig . 2. A blation energy for" light" and " claTk" stakes, 827 Ju ly 1993 (Julian days 189- 208), at Kronprills
CIzTistian Land. Unit is W 172 2.
Downloaded176
from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
I
KOllzelmanll and Braithwaite: Ablation, albedo and energy balance in Kronprins Christian Land
Bra ithw aite) . Although th ese dat a ca nn ot be very
acc ura te, a crust d ensity of 500- 700 kg m 3 was indi cated. Over the whole 20 d period of the fi eld experiment
with an average i1h of 0.9m, th e second term , involving
cha nge in crust thi ckn ess, is entirely negligible. For th e
d a il y a bl ation readings, however, th e second term cou ld
be a signifi cant so urce of error althou g h the weathering
cru st neve r co mple te ly d isa pp ear ed in th e present
ex pcrim ent as obse rved by MUller a nd K eeler ( 1969 ) in
their data se ts. It should be noted that a n error of this
kind is not detected by the Llibo utry m odel becausc the
error is more-or-Icss co mm on to a ll sta kes.
RADIATIVE FLUXES
Du c to the importa n ce of radiativc Ouxes in th e energy
ba lance, priority was given to accuratc radiation measurem ents. At the base camp , glo b a l radi a tion a nd allwave in coming ra diation were m eas ured direc tl y with a
Swissteco SS-25 p y ra nometer a nd a Swissteco ST -25
py rradiometer, res pec ti vely. The lo n gwave incoming
radi a ti on was then calcula ted as th e difference between
th e a ll-wav e in co ming radi a tion an d th e g loba l
radi a ti on, a nd compen sa ted for th e emission loss of the
instrument (0'T;4) , where 0' is the Stcfan- Bo ltz mann
con sta nt ( = 5.67 x 10 8 Wm 2 K .,) a nd 1i th e tempe rature of the p yrradiometer in kelvin. Th e cosine
error of th e u pfacing p yranomete rs was co rrected for
ze nith a ngle > 70 ° using a pol y n o mi a l function,
whereby onl y the part of direc t sola r radia tion was
taken into acco unt. Based on meas ure m ents made und e r
simil a r meteoro logical cond-itions a t ETH Camp in
June 1990, it was ass um ed th a t 70% o f g loba l radi a tion
is ca used by direct sola r radiation (K o nzel ma nn , 1994) .
Du e to a poss ibl e und erestim a tion of th e a ll-w ave
in co ming radiation du e to thermal co nv ec tion as
proposed by Ohmura and Gilge n ( 1993 ), the longwave
in co ming radiation was correc ted accordingl y. Shortwave radi a tivc Ouxes at th e g lacier sta tion were
m eas ured with a Swissteco S'vV-2 two- co mpon e nt
p yra nom e te r (a lb e dom e ter ). Th e in str ument \ov a
mounted on a tripod a t a height of abo ut I m a nd was
reloca ted every seco nd d ay to sampl e as many ty pes of
surface as possible. In genera l, g lo b a l radi a tion a t th e
g lac ie r sta ti o n and a t base camp ag reed closely.
Howe\'Cr, to avoid poss ible sys tem a tic errors due to
a n y tilt o[ th e r a diometer a t th e glacier sta tion , g loba l
radia tion meas ured at base ca mp w as used to calculate
a lbed o toge ther with shortw ave refl ec tcd radiation at
th e glacier stati o n. A ss uming melting co nditi ons a ll da y
a t th e glacier station , longwa\'e o utgoing radiation was
se t a t 3 16 \V m 2 according to th e Ste[an Boltz mann
law.
Th e response tim e of the radi om eters is abo ut 5 sa nd
th e sig na ls of the instruments were co ntinu ously recorded
every 15 s. Every 30 min th e average was computed and
stored. The opera tion of the instrumen ts a nd da ta loggers
was checked sever al tim es a da y. Th e un ce rta inty in the
longwave in co ming radiati on was se t at ±IO W m 2
acco rd i ng to th e correc tion meth od m en tioned above
a nd a n in str um e nt intercomp a ri so n perform ed b y
D eLuisi a nd oth ers ( 1993 ).
TURBULENT FLUXES
Th e sensible-heat flux (SHF ) is estimated by th e method
o[ Amb ac h (1986 ) from m eas ured wind speed and
temperature:
(3)
where K s is the exchange coeffi cient for turbulent-h eat
flux , P is the atmosph eric press ure, U2 is the wind speed at
2 m above the glacier surface and T2 is th e a ir
temperat ure a t 2 m a bove the glacier surface. Th e glacier
surface is ass umed to be a t the melting point. For a
Prandtl-ty pe neutral boundary layer with loga rithmic
profi le for wind speed , temperature a nd vapour press ure,
the exch a nge coeffi cient is given by:
where cp is th e specific heat of air with co nsta nt pressure
(1005Jkg 'K ' ), kisvo nK a rma n'sco nsta nt (0.41 ), po is
the densit y of air in the sta ndard conditi on ( 1.29 kg m \
bo is the stand a rd atmosph eric pressure ( 1.01 3 x 105 Pa ),
Z is th e in strum ent heig h t (2 m ) and ZOw a nd ZOT are th e
roughn ess lengths [or loga ri thmi c profi les o[ wind and
temperature, respectively.
Th e la tent-h eat flux (LHF ) is similarly es tim a ted from
meas ured d a ta [or wind speed a nd vapo ur press ure:
(5)
wh ere KL is th e exc hange coe fIi c.ien t for late nt-h ea t flux
a nd i1e2 is th e difference be tween \'apo ur pressure at 2 m
a bove th e g lac ier surface a nd that at th e melting glacier
surface. For th e same assumpti ons as befo re, th e exchange
coeffi cie n t is given by:
where L is th e latent hea t o[ evapo ra tion o r sublimation
as app rop ri a te (( 2.514 o r 2.849 ) x 106J kg ' ) and ZOe is
th e ro ug hn ess length [o r th e logarithmi c profil e o[ water
vapour.
Followi ng Ambach ( 1986 ), th e roug hn ess length [or
\\'ind ove r an ice surface is ass um ed to be 2.0 x 10 3 m
while the ro ughn ess len g th s fo r temperature a nd water
va pour a r e both assumed to be 6.0 x 10 b m . As
formu la ted by Ambach ( 1986 ), th e exchange coefficients K s and KL are valid for a neutral boundary layer
while th e air just over a glacier surface m ay be ra th er
sta ble.
Th e tu rbu lent-heat 0 uxes were calcu lated ever y
hour according to Equ a tion s (4) and (6 ) using th e air
temper a ture, relative hum idity a nd wind sp eed measured a t 2 m. Some r a di a tion hea tin g of th e tempera ture/humid ity screen was noted whereby temperatures
sud-d en ly rose when wind speeds were low, and th e
d a ta we re co rrected by h a nd. Hourl y values of th e
calcu lated turbu lent Duxes and ha l[-hourl y values of
th e radiative flux es were then summ ed up to daily
values e nding a t 1900 h loca l time, so th ey co uld be
co m pared direc tly wi th a b lation meas u re m en ts mad e
a t th a t tim e.
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
177
Journal of Glaciology
ICE TEMPERATURES AND CONDUCTIVE-HEAT
FLUX IN THE ICE
4
It was exp ected tha t the co ndu ctive-h eat flu x in the ice
wo uld b e a significant en ergy sink in contras t to the icesheet m a rgin in Wes t Gree nl and where it was ass umed by
Braithwaite a nd Olesen ( 1990) to be n egligibly small.
Englacial tempera tures were therefore m eas ured ever y
day to calculate the conductive-h eat flu x in the ice. The
measurements were mad e a t two strings with thermi stors
which were drilled a t 0. 5 m intervals dow n to a maximum
of 3 m d ep th.
M eas ured tempera tures were even lower than expected with the ooe isotherm only 0.2- 0 .3 m below the
surface (even with the possible help of radia ti on warming
of th e thermistor cables) . This probably accounts for th e
ra ther difficult drilling conditions en countered when
placing th e stakes, i.e. a wet drill b a rrel penetra ting
very co ld ice and frequently j amming . During the period
of m easurements, the ice tempera tures a t th e various
depths rose by 0.2- 0.3 K d- I.
The ra te of change of ice tempera ture with time is
aT I at given by:
3
aT l at = - (11pic) dH (z)/ dz,
(7)
where Pi is density of ice (900 kg m-3) , C is specific heat of
ice (2009J kg- I K - I), H (z) is the englacial-h eat flu x and z
is the d epth below the surface . Ass uming that aT I at
equ als a consta nt C du ring the period of m easurements,
then H (z) is a linear fun ction of depth:
H(z)
=
Ho - CPiCZ,
(8)
wh ere Ho is the heat flu x through the glacier surface a t
Ho can be calcula ted
as the intercept in a regression equ a tion of h eat flu x versus
depth where th e hea t flu x H(z) is calcula ted from
englacial-temperature d a ta according to the equ a tion
z = 0 m. Equa tion (8) suggests tha t
H (z) = - KaTl az,
(9)
where K is the th ermal conductivity of ice, 2.1 W m- I K - I
according to Paterson ( 198 1, p. 186 ). The heat flu xes for
each d ay a re strongly co rrela ted with d epth with
gen era lly similar equ a ti ons and , assuming that an y
differen ces are du e to sta tistical flu ctu a tions, the mean
hea t flu x Ho for the whole period was calcula ted by
combining all the availa ble da ta in to a single regression
equ a tion (Fig. 3). The intercept is 17.6Wm 2 and th e
slope of the regression lin e correspond s to a constant
temperature change of 0.21 K d- I during th e field period .
Extrapola tion of the regression line (Fig. 3) to greater
depth sugges ts that the h ea t flu x becomes zero at a bout
4 m d epth bu t it is more likely tha t th e a pproach to zero
flu x is asymptotic at greater depths tha n 4 m .
ENERGY BALANCE
Th e en ergy used for a blation ABL is given by:
o
o +-.~
o
20
15
5
10
2
Conductive heat flu x in the ice W m -
Fig. 3. Ca lculated conductive heat flux in the ice versus
depth below the glacier surface, 8- 27 July 1993 .
ABL + CHF = SWR + LWR + SHF + LHF + ERR,
(10)
where ABL is calcula ted from measured a bla tion using
the la tent h eat of fu sion (3.34 x 105Jkg I), CHF is
conductive-heat flu x in the ice, SWR is sh o rtwave net
radia tion , LWR is longwave net radia tio n , SHF is
sensible-heat flu x, LHF is la tent-heat flu x, a nd ERR is
th e total error du e to measurement erro rs, simplifying
assumptions a nd disregard ed t erms in the ene rgy-balance
equation . D efin ed as above, ERR can be regarded as an
unknown , extra source of en ergy which is es tim a ted from
the residu al of th e oth er terms in Equ a tion ( 10 ). The
magnitud e of ERR is a usefu l ch eck on th e accuracy of the
other terms. F or calcul a tion of SWR an albed o of 0. 48
was used . The condutive-h ea t flu x in the ice was taken as
17.6 W m- 2 . The time series of all compon en ts of th e
energy b ala n ce is presented in Figure 4.
The calcula ted energy b ala n ce for the 20 d is shown in
T able 2 a nd summ a rised in T a b le 3. As disc ussed below,
the main error is probably du e to the turbulent flu xes so
tha t they p ro b ably constitute more tha n th e 26% of
abla tion en ergy shown in the ta ble.
ERROR IN THE CALCULATED ENERGY
BALANCE
On average, the error term ERR represents n ea rl y I1 % of
th e total en ergy balance. I n view of the uncerta in ty of the
various ass umptions, it is good tha t the mean er ro r is not
even larger. For example, Braithwaite and O lesen (1990)
found errors of -40 to + 29 % of monthly a bla tion at
Nordbogle tsch er and Q a m a n a rssup sermia . In terms of
variance, th e error corresponds to about on e- third of the
ablation varia nce. There is therefore a substa nti a l source
of both sys tem a tic and ra ndo m error in th e energybalance calculation .
The error ERR has almost no correla tion with the
radi ation terms SWR and L WR. On the other h a nd , the
error has mod era te correla tion with the turbulent flu xes,
i.e. r = + 0 .63 for SHF a nd r = - 0.53 fo r LHF , whi ch
Downloaded178
from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
Konz elmann and Braithwaite : Ablation, albedo and energy balance in
III
Wm- 2
i'R
11 SH' 0
OF IIIlI! LHF
~
ABL •
Ern
I
suggests that they maybe the main sources of systemati c
error in the calculated energy balance. For example, if
both SHF and LHF are multiplied by a factor of 1.5
before calculating the energy balance, the mean error is
essentially reduced to zero (th e error standard deviation is
littl e changed ). In order to assess the plausibility ofa 50%
error in the turbulent fluxes it is necessary to examine in
more detail how they are calcu lated, i.e. with regard to
sta bility and to the assumed surface roughness.
The exchang e coeffici ent [or sensible-h ea t nu x
(Ambach , 1986 ) given by Equation (4) does not take
account of the stability of the boundary laye r over the
m elting glacier. Price and Dunne ( 1976) propose stabili ty
corrections in terms of the bulk Richardson numb er, and
application of this r ed uces the sensible-heat flux in the
present stud y by 5- 18 %, thus increasing the error in th e
calc ulated energy balance rath er th a n reducing it.
The Ambach ( 1986 ) formulation uses different rough n ess lengths for wind , temperature and water vapour
(over ice) while Price and Dunne (1976) a nd Moore
( 1983 ) assume the same roughness lengths (over snow).
Pu tting ZOw = Z OT = Z Oe = 10- 3 m in Equations (4 ) and
(6) would exactly account for the mean error in the
calculated energy balance. Such a roughness length falls
well within the wide range of values quoted in the
literature [or ice, e_g. by Grainger and Lister ( 1966),
Streten and Wendler (1968), W endler and Weller ( 1974),
Poggi (1977 ) and Hay and Fitzharris (1988 ), although
Morris (1989 ) suggests that some o[ the larger roughness
300
c
'(ij
200
Cl
>,
Cl
Q;
c
100
w
0
(/)
(/)
.Q -100
>,
~
Q)
c
w -200
-300
o
ill
Ol
Ol
(\j
Ol
Ol
l()
o(\j
o
(\j
~
co
o
(\j
Fig . 4. Energ)1balance 8-27 July 1993 (Julian days 189208) at Kronprins Christian Land with net radiation
(NR ) , sensible-heat flux ( SHF) , latent-heat Jlux
( LHF), conductive-heat flux in the ice ( CHF), ablation
energy ( ABL) and error ( ERR ) . AB L conesjlonds to
stake measurements. ERR appears either on the top or on
th.e bottom depending on whether there is a surplus or dificit
oJenergy. Unit is Wm 2 .
Table 2. Daily energyl balance 011 the margin oJ the Greenland ice sheet, Kronprins Christian Land, Ju0' 1993. Units are W m- 2
+
LWR
+
Date
Jul y 199 3
SWR
SHF
8
9
lO
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
186.0
182.9
168 .9
178.8
149.2
145.6
175 .7
174.6
174.1
159.1
168.9
170.5
162.7
159.6
159.6
156.5
158. 1
159.6
158. 1
155.5
- 58.0
- 54.0
- 5 1.0
- 55.0
--43.0
- 35.0
--48 .0
--46.0
--46.0
- 38.0
--4 1.0
- 49.0
- 52.0
--44.0
- 50.0
- 51.0
- 56.0
- 60 .0
- 63.0
- 58.0
59.9
79.7
78.4
43.8
45.6
3 1.2
65.1
88.2
90.5
59.5
49_8
82.7
87.5
75.6
67.5
82.2
57.3
33 .8
33. 1
38.8
Mean
s.d.
165.2
11.0
--49.9
7.4
62.5
20 .1
+
LHF
+
+
ERR
ABL
- 25.4
- 26.6
- 14.6
- 14.0
- 18. 5
- 9.1
- 18.4
- 23.9
- 33.3
- 18.6
- 12.4
- 31.9
- 33 .2
- 21.8
- 25.0
- 34.8
- 28 .6
- 15 .9
- 19 .2
- 17 .5
- 20.1
39.4
23.3
- 8.9
12. 3
- 5.2
15. 1
50.0
11.5
- 4.4
15.1
40.0
56.4
36.9
21.9
59.4
8 .3
15.6
5.1
29.3
124.9
203.8
187.4
127.1
128.0
109.9
171.9
225 .3
179.2
140.0
162.8
194.7
203.8
188.7
156.4
194.7
121.5
115.5
96 .5
130.5
17.6
17 .6
17.6
17.6
17.6
17.6
17.6
17 .6
17.6
17.6
17.6
17 .6
17.6
17 .6
17.6
17 .6
17.6
17.6
17.6
17.6
- 22 .1
7.6
20.1
22.0
158.2
37.8
17.6
0.0
CHF
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
179
Journal oJ Cla ciology
values may be ca used by slope errors . It appears,
therefore, th a t the calcula tion of the energy b a lance is
ex tremely sensitive to th e ass umptions a bout s urface
roughn ess in agr eement with Munro (1989 ) . This is an
importa nt point for future attempts to co upl e ice-shee t
models to ge neral circula tion mod els (GCMs) .
EFFECT OF ALBEDO ON ABLATION
190
~
A
180
E
•
. c
;= 170
>-
CJ)
•
D.
ID
cQ) 160
c
0
fij 150
G
•
•
:0
«
H •
E
•
F
140
M eas urem ents of a lbedo are norm a ll y mad e a t one point
a nd therefore d epend strong ly o n th e local s urface
conditions. For a n area of 50 m x 100 m in th e upper
a blation zone (Camp IV ) , Ambach ( 1963 ) gives a
variation of the a lbedo from 0.34 to 0.58 for a particular
d ay . Preliminary res ults from a n a ltitudinal profile of
a lbedo on bare ice close to the prese nt study site show
a lbedo variations of between 0. 30 and 0.62 (p ersonal
co mmuni ca tion from H. O erter) .
On a total of 16 d, a lbed o was meas ured a t different
places insid e the area of ablation m easurements. Parts with
" d ark" surface (8 d ) and " light" surface (8d ) were selected
subjec tively. For the " dark " surface a mean albedo of 0.43
was fo und and the " light" surface showed a mea n value of
0.53, while the m ean albed o of all sites was 0 .48. The
a lbedo difference between " light" a nd " dark" surfaces is
here 0.10 . For a d ail y mean glob al radia tion of320 W m 2,
this gives a d iffer ence in calc ulated melt of 8.3 kg m - 2 d I
which is about 20% of the mean a bla tion over the whole
d ata set (T able I ). Th e difIc rence in mean a blation
between the four " dark" and th e fi ve " li ght" stakes is
5 kg m 2 d I while the greatest inter-sta ke difference is
I1 kg m 2 d I between stakes A and I (Table I ) .
The above disc ussion refers to separate a blation a nd
a lbedo variatio ns within a n area, a nd it would be
interesting to extend th e disc ussion to th e relation
between a bl ation a nd a lb edo for indi\'idual stakes.
H owever, it is impossible to measure albed o exactly a t a
stake because th e stake a nd the albedometer interfere
with eac h other , a nd a mo re subjective procedure [or
estimating albedo a t each stake is used . In thi s method,
the su rface co nditions under the albedom eter were
co mpa red on eac h day to the situations at n earby
stakes, and the measured a lb edo va lu e was assign ed to
one or more stakes that appea red to have a simil a r
surface. The a lbedo for each stake was then d etermined as
the average of the assigned albed o values. The estima ted
a lbedos o[ the four " dark " stakes (A, C, D and G ) agree
close ly but there is quite a large r a nge [or th e fiv e " light"
stakes . Average albedos for the gro ups are higher than
previously stated , but this simply refl ects the different
sampling. Th e relation between mean daily a blation a nd
th e estim ated a lbed o for the individual stakes is illustr ated
by Figure 5. The correlation sh ows a \'alue of r = - 0.82,
whi ch aga in suggests that a lbedo is a main factor 111
ablation varia ti o ns within a small area.
The above indi cates that a point meas urem ent of
a lbed o is no t sufficient for acc urate energ y- balance
calc ula tions in a blation areas. Therefore, a n albedo
value for a larger area should be determined , either by
m easurements from high towers (L a ngleben, 1968) or by
enough point m eas urements at gro und level to sa mpl e
small-scale a lb edo variations as attempted here.
180
•
130
0.4
0.45
0.5
0.55
0.6
Albedo
Fig. 5. Ablation energy in IV 171 2 as a function of albedo
based on mean daily values on the margin oJ the Greenland
ice sheet in July 1993, Kronprins Christian Land.
COMPARISON WITH WEST GREENLAND
I n ,1\1 es t Greenland , energy-bala nce studies were carried
o ut b y Braithwaite an d Olesen ( 1990 ) at Nordbogletscher
(NBG ) and Q a ma n a rssup sermi a (QAM ), a nd by Van de
Wal a nd O erleman s ( 1994 ) near S0ndre Stmmfjord in
en vironments similar to those found a t th e present stud y
site (KCL). At all of these stations the net radiation was
meas ured at one poin t and no albedo variation was taken
into acco unt . Ambach (1963 ) performed a more sophistica ted energy-ba la nce stu d y at Camp IV , but the
sta ti o n was located in the upper a blation a rea, a nd the
a lb edo variations and their efIect on abla ti on were
unfortunately not a na lysed in detail. The energy balance
from different experiments is summ a ri zed in Table 3. Net
radiation is th e main so urce of ablation energy at a ll
locatio ns. Turbulent flu xes are a n important energy
so urce in the ablation zo ne co mpa red to the situ a tio n
close to th e ELA (Am bach, 1963; Ohmura and others, in
press) , corresponding to an ab lation rate of about 10-
Table 3. Energy balance in July 1993 in Kronprins
Christian Land ( KCL ) compared with July values at
Qamanarssllp sermia ( QA M ) Jor 1980- 86 and ./I"ordbogletscher (NBC) for 1979- 83 }i"om Braithwaite and
Olesen ( 1990) . Signs correspond to the difinition given in
Equation (10) . Small letter ( a) indicates a value assumed
to be 0 Wm 2
Elevation
D ays
Net radiation
Turbulent flu xes
H ea t cond uction
Error
Melting energy
m
Wm
Wm
Wm
vV m
Wm
2
2
2
2
2
KCL
Q;1M
N BC
380
20
11 5
40
18
20
157
790
185
123
79
a
-4
198
890
155
89
44
a
2
135
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
Konzelmann and Eraithwaite: Ablalion, albedo and eneTgy balance in Kronprins Chrislian Land
')
1
20kgm - d . In contrast to West Green land where heat
conduction into the ice is assumed to be sma ll , it is an
important heat sink in the present stud y. The possible
ab lation is red uced by about 4.5 kg m 2 d I.
Ablation rates a nd degree-da y factors for KCL , QAM
and NBG are summarized in Tab le 4. Despite the low
elevation of the statio n compared with the two West
Greenland sites, the ab lation rate is quite low. This is
partly du e to lower air temperature , but that is offset by a
high er degree-day factor. At KCL a value of 9.8 mm
d 1 cC 1 was found compared with 7.2- 8.1 mm d 1 QC 1 at
the West Green land sites (Braithwa ite, 1992 ), which may
be th e res ult of higher wind speeds . Boggild and others
( 1994) a lso got a higher degree-day factor (9.6 mm d 1
QC I) at Storstr0mmen in north eas t Green la nd (Fig. I ).
Table 4. Ablation rate and degree-da)1 faclor ill July 1993
in Kronpl'ins Christian Lalld ( K CL ) comjJared with Ju£y
values at QJlfnanarssu/J sel'mia ( QlIM ) for 1980- 86 and
~ V ordbgletsc
h er
(}vBG ) for 1979 83 from Braitlzwaite
(1992)
m
E levation
Da ys
mmd- I
Ab lation ra te
QC
Air temperature
Degree-d ay
mmd 1 QC
factor
ms
Wind speed
NEG
KCL
(LAM
380
20
40
4.2
9.8
790
185
53
6.3
8.1
890
155
35
4.6
7.2
6.2
4.8
3.2
CONCLUSIONS
Net radiation is th e major so urce of a bla tion energy at the
margin of th e Green land ice shee t in Kronprins Ch ristian
L a nd as in other parts of Green la nd. Albedo varies
greatly within a sma ll area and, because of the high
in co me of g loba l radiation , causes relati\'ely large
differences in a blation at stakes close to eac h other.
Sma ll-sca le a lbedo variations mu st therefore be ca refu ll y
sa m pled to 0 btai n represen ta ti ve a l bcdo val ues for largescale energy- balance calculation.
Th e ca lculated energy balance in the pre ent study
a lso has a substantial error w hich may be caused by
und erestimation of the turbu le nt flux es using the
aerodynamic formu lae of Ambach ( 1986 ). The reaso n
is probably that the glacier surface is rougher than
assumed. Th e co nducti\'e-h ea t flux in the ice is a
substantial hea t sink and reduces th e energy available
for ablation. The average ab lation rate in Kronprins
Christian Land in J uly is low co mpared wi th values in
West Greenland, re flecting lower temperature a lth o ugh
this is panly ofIset by a higher d egree-day factor due to
high wind speed. The degree-day factor is significantly
higher in north Greenland than th e ones found in West
Greenland.
ACKNOWLEDGEMENTS
This paper is published by permission of The Geologica l
Survey of Greenland (GGU ). K. SchroIT (Department of
Geogra phy, Swiss Federal Institute of Technology) and
O. O lcsen (GGU ) prepared instruments. Logistic support
was supplied by GGU' s camp at Centrum So, Kronprin s
Christian Land, led by GGU Stat. geolog N. Henriksen.
The research is supported by th e European Community
und er co ntra ct numb er EV5V-CT91-0051 which is
coordina ted by the C limate R esearc h Unit, U niversi ty
of East Anglia. Th e first draft of this paper was prepared
in November 1993 when R.J. Braithwaite was a guest of
Professor A. Ohmura, D epa rtm ent of Geography, Swiss
Federal I nstitute of T ec hnology (grant No. 0-04-509-93 ) .
Professor Ohmura a lso made useful comments on the
m an uscri pt.
REFERENCES
t\ mbach, \\'. 1963. UnLersuchungen ZU1l1 Energieumsatz in der
Ablationszone des gronhindisc hen Inland eises (Camp J\' EGIG,
69 40'05":-.1, 49 ~ 37'58"
\\' . . I/ etlt!. Grolll" 174 (4 ).
Ambach, 1\'. 1986. :'\Iomographs ror the determination or meltwater
" 'om snoll'- and ice surfaces. Berichte des . \'atlll'1(,issntschaj'tlich.I/edi::.inischen I'ereills in IIIIIJbl'l/ck, 73. 7 15.
Anon. 1969. I\[ass-balance terms. J. Glacial., 8 (52 ). 3 7.
Boggild , C. E.. N. R ee h and H. Oener. 1994-. :\lodelling ablation a nd
mass-balance sensiti"ity to climate change or Storstromll1en. northeas t Greenland. Global al/d Planetill} Change. 9, 79 90.
Braith waite, R .j. 1985 . Glaciological investigations at QamanarsSL'I)
se rmia. West Greenland, 1983 198-1. Gl'olllal/ds Geologiske Clldersogelse.
Gletschr-/~vdogi
ske
,I/ eddelelser 85 /3.
Brait h \\'aite, R . j. 1992. D egree-day ractor. energy balance. ancl the
increased melting or the Greenland ice sheet under a lI'armer cl im ate.
Gronlands Geologiske Cndersogelse. Rapt!Ol't 155. 79 83.
Brai thll'aite. R.j. and O. B. 01 (,5e n. 1989. Calcu lation of'glacier ablation
rrom a ir temperature. \\'cst Greenland. III Oerlemans. j. , ed. Glacier
fluctuations and climatic change. Dord recht. Klu\\'e r Academic Publishers, 2 19 233.
Braith wa ite. R.j. and O. B. Olescn. 1990. J\ simpl e energy- balan ce
m odel to ca lculat e ice ablation a t the margin o f' th e Greenland ice
sh ect. J. Glaciol., 36 ( 123 ), 222 228.
De Lui si, j. , K . Dehn e, R. \ 'ogt, 'J'. K onzclma nn and i\. Ohmura. 1993.
First results or the baselin~
suditec rad ia tion nctwork ( BSR~
)
broadband inrrared radiometer intcrcolllpariwn at FIRE 11. i ll
K ecvallik , S. and O. K ;il'ller. eds. Cllrrent j))'oblelllJ ill atmosj)hl'l'ic
radiatiol/. IlItemational Radiatioll ,~)'mjOil.
Tallillll 1992. Hamp to n. \ ' A.
A . Deepak Publishin g, 559 56+.
Geb, 1\ 1. and B. :'\aujokat. 1993. :'\ordhcmi,phii ri,cher Klimabc ri c ht
ZUIll juni /juli/Augll st 1993. Beilage ~lIr
Berliller lI'ellerkarte. Be rlin.
Fre ie Univcrsi tat Berl in. I nstitut f'Lir \leteo ro logie.
Grainger, 111. E. and H . Li stel'. 1966. Wind speed , stability and eddy
viscosi ty over melting ice surra ces. J. Glacial., 6 (4-3 ). 101 127.
H ay, j. E. and B. B. Fitz harri s. 1988. i\ compa ri son of thc encrgybalance and bulk-aerodynamic a pproaches ror estimating g la c ie r
mclt.J.Glaciol.,34 ( 117 ), 1+5 153.
K onzelmann . T. 1994. R adia tio n condit ions o n th e Greenland ice sh eet.
.<)i'l'c/icr Geograj)/iische Sclzrijiell 56.
Kreysz ig, E. 1968. Statistiscfle ,llethodell IIlId i/ire AlIll'l'IIdllllgen. GOllingen.
V andcnhoeck and Ruprec ht.
Lan g leben , 1\ 1. P. 1968. Albedo measurements or an Arctic ice cover
rrom hig h to\\'crs. J. Glacio/., 7 50 ). 289 297.
L1ibout r) , L. 1974. i\l ulti variate statistical a n a lysis or glacier a nnu al
balances. J. Glacial., 13 (69 ). 371 392.
I\[ oo rc , R . D. 1983. On the use or bulk ae rodynamic formulae ove r
melting snow . ."ord. H),drol., 14 (4) , 193 206.
I\ lo rri s, E.1\ 1. 1989. Turbul e nt transfer ove r sn o\\' and ice. J. H ydro/.,
105, 205 223.
\1Lill er, F. and C . I\1. K eele r. 1969. Errors in short-te rm a blatiun
measurements on m el tin g ice surraces . J. Glacial" 8 (52), 9 1 105.
M unro, D. S. 1989. Surrace roughness and bulk h ea t transrer on a g lac ier:
compa rison with edd)' co rrela ti on. J. Glacial., 35 ( 121 ).343 3+8 .
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X
181
J ournal of Glaciology
Ohmu ra, A. a nd H . Gilgen. 1993. R e-eval uation of th e g lo ba l energy
bala nce. Geophys. M onogr .. Am. Geophys. Union 75 , 93- 11 0 .
Ohmura, A. and 8 others. 1991. En ergy a nd mass bala nce d uring the melt
season at th e eq uilibrium li ne a ltitude. Paa ki tsoq, G ree nland ice
sheet. ETH G reen land Expediti o n . Progress Report JVo . 1. Zurich,
Swiss Feder a l I ns titute of T echno logy.
Ohmura , A. and 10 others. 1992. En ergy and mass ba la n ce during th e
melt season a t th e equilibrium lin e a ltitud e. Paakitsoq , Greenl and ice
shee t. ETH G reenl and Ex pedi tion . Progress Report No. 2. Zuri ch,
Swiss Fed era l Institute of T echno logy.
O hm ura, A., T . K on zelm ann, j. Fon 'er , A . Abe-Ou chi a nd H. T oritani.
In press. E nergy balance for th e G reenl and ice sheet b y o bserva tion
a nd model comp uta tion. Intmwtional Association oJ !flldrological Sciellces
Publication (Sy mpos ia at Yokoh a m a 1993 - Sl1ow Caller al1d its
In teractions with Climate and Ecosystems; Processes oJ IvI ass and Hllergy
Exchange betweell the Atmosphere and P olar Swiace ).
Ol esen , O . B. a nd R .j. Braithwa ite. 1989. Field sta ti on s for glacier
climate resea rch , vVes t Greenla nd . [n O erl ema ns, .1 ., ed. Glacier
fluctuations alld climatic change. Dord rech t, Klu wer Academi c Publishers, 207- 218.
P a terson , W. S. B. 198 1. The ph),sics oJ glaciers. Second edition. O xfo rd , etc..
P ergamon Press.
P oggi , A. 1977 . H ea t b a la nce in th e a bla ti on a rea of the Ampere Gl acier
(K erguelen Isla nd s) . ]. Appl. Meteorol., 16 (1) . 48- 55.
Pri ce, A. G. and T. Dunn e. 1976. En ergy ba la nce computatio ns of
snowmelt in a suba rcti c area. Waler Resour. Res., 12(4). 686- 694.
Streten, N. A. a nd G . W endler. 1968 . Th e midsummer hea t ba la nce of
a n Alaskan ma ritim e glacier. J. Glaciol., 7 (5 1), 43 1- 440 .
\"'al, R. S. W. va n d e a nd.J . O erl em a n s. 1994·. An en ergy b a la n ce
m od el for th e G r eenl a nd ice shee t. Global alld Planetmy Change, 9,
11 5- 131.
W e nd1er, G. a nd G . W el ler. 1974. A hea t-ba la nce stud y o n M cCall
Gl acier, Brooks R a n ge, Alaska: a contributi on to th e Intern a tiona l
H vdrologica l D ecad e. ]. Glacial. , 13 (67 ), 13- 26.
MS received 27 Janua ry 1994 and in revisedJorm 28 J une 1994
182
Downloaded from https://www.cambridge.org/core. IP address: 168.151.137.40, on 23 Sep 2017 at 20:48:02, subject to the Cambridge Core terms of use, available at
https://www.cambridge.org/core/terms. https://doi.org/10.1017/S002214300001786X