Learning Center
Plans & pricing Sign in
Sign Out

Freshwater _ Ice Fluxes West of Greenland_ An Annotated Bibliography


									 Freshwater & Ice Fluxes Through Baffin Bay: An Annotated Bibliography
    A Contribution to the NSF FWI Changes and Attribution Working Group
                                      Kelly Falkner

This bibliography briefly summarizes selected papers addressing the fluxes of freshwater
and ice through the three main passages of the Canadian Archipelago, namely Lancaster,
Jones and Smith Sounds, and Davis Strait into the North Atlantic Ocean. We note there
is another passage called Fury and Hecla Strait that connects the Gulf of Boothia through
Foxe Basin to Hudson Bay. The transport there appears to be relatively small and is not
delivered to Baffin Bay but rather through Hudson Strait to the Labrador Sea. Unlike for
some of the other components of the freshwater system of the Arctic, there is very little
information regarding changes in fluxes through the Canadian Archipelago. Rather the
community is still wrestling with the difficulties of measuring them properly. Limited
records to date suggest high variability at a number of spatial and temporal scales.
Efforts to resolve these observationally are under way. Certain models have attempted to
grapple with the issue of variability. At this juncture, differences between estimates
should be understood to result more from uncertainties in approach rather than true
changes in time.

Classic references and more recent publications that summarize the state of knowledge of
the system are reviewed below. In fact, there is a group effort underway to update such a
summary in the form of a book covering the Arctic Subarctic Ocean Fluxes (ASOF)
program, with an April 2006 deadline for first drafts of chapters. The authors of the book
will be meeting in Torshavn, Faroe Islands on June 27-July 1, 2006 to advance the
project and publication is targeted for fall 2006.

Aagaard K. and Carmack E. C. (1989) The role of sea ice and other fresh water in
the Arctic circulation. J. Geophys. Res. 94(C10), 14,485-14,498.
This is perhaps the most often cited reference summarizing all the freshwater flux
components for the Arctic system. It is also cited for pointing out that convection in the
Greenland, Iceland and Labrador Seas and by extension, deepwater renewal or
thermohaline circulation, could be altered or stopped by fairly small variations in
freshwater exiting the Arctic. The authors reviewed the literature to construct a
freshwater budget and determine storage terms for the Arctic. They provide a table of
fluxes for the Arctic Basin relative to a salinity of 34.80 practical salinity units.
Concerning the three main passages of the Archipelago, they cite Fissel et al., (1988; see
below) for a total volume flux of 1.7 Sv (1 Sverdrup = 106 m3 s–1 = 31,536 km3 yr–1) and
Aagaard and Greisman (1975) for a mean salinity in the passages of 34.2. This salinity is
based on consideration of data from Muench (1971; see below) from whom they derived
a summer mean value of 33.8 for outflow into northern Baffin Bay. They suggest that
winter values may increase by up to 0.8 and so assigned a mid value of 34.2. These
numbers yield a freshwater outflow of 920 km3 yr–1. They do not include an Archipelago
ice flux because they estimate it to be small as follows. They consider that the cross
sectional area of the main passages is 34 km2 and presuming the 1.7 Sv volume flux,
obtain a mean outflow speed of 5 cm s–1. They assumed an average ice thickness of 2 m
and that it would flow for 3 months per year and otherwise be landlocked. These
assumptions result in an ice flux of 155 km3 yr–1 which is smaller than any other terms in
their budget and thus was considered negligible. While this study provides a benchmark,
it must be realized that there can be considerable error in combining average flows with
average properties. As a quantitative demonstration of this, Cuny et al. (2005; see below)
show that using averages for their Davis Strait sections would result in fluxes 46% less
than obtained using the integrated fluxes from the time series.

Fissel D. B., Lemon D. D., Melling H., and Lake R. A. (1988) Non-tidal flows in the
Northwest Passage. Can. Tech. Rep. Hydrogr. Ocean Sci. 98, Institute of Ocean
Sciences, Sidney, B. C., Canada.
The residual (tidal corrected) flow fields from 63 current meter deployments throughout
the Canadian Archipelago beneath the sea ice between 1982 and 1985 are presented.
These were deployed between about 70 to 76˚N in western, middle and eastern sectors in
successive years. Most deployments were for 1-3 months but eight lasted for a full year.
The measurements were predominantly near-surface and in the springtime when the ice is
intact but sunlight has returned. The times series information was combined with
estimates of baroclinic transport computed from density sections undertaken in
conjunction with the mooring program to estimate total volume fluxes in selected
channels. To derive a total volume flux through Lancaster Sound to Baffin Bay they sum
their results for Barrow Strait (0.44+0.25 Sv) and Wellington Channel (0.24+0.05 Sv) to
obtain 0.68 Sv. (I note that these results from the text differ somewhat from what is
presented in Fig. 4-3.) They measured a throughput to Foxe Basin via Fury and Hecla
Strait of 0.04 Sv. They refer to Sadler’s (1976: see below) current meter referenced (2
months in 1971) total volume flux of 0.67 Sv for Nares Strait. They note that previous
baroclinic estimates of flux through Jones Sound are about one-half of those estimated
for Lancaster Sound and so assign Jones Sound a volume flux of 0.34 Sv. These tally to
1.7 Sv. Given limits in temporal and spatial resolution of the measurements and the
necessity of unproven assumptions, this result must be considered to have a large

This reference also includes a concise review of efforts to determine fluxes from 1928 to
then. It is noted that current meters were first deployed with limited success in the
Archipelago in Smith Sound in summer 1963 as reported in Paltry and Day (1968).
Internally recording current meters began to be deployed in 1969 and nearly 400 current
meter records were obtained from 1969-1982. Approximately one-third was judged to
provide useful information on currents but data coverage was generally too limited to
estimate volume transport.

Muench R. D. (1971) The physical oceanography of the northern Baffin Bay region.
Baffin Bay-North Water Proj. Sci. Rep. 1, Arctic Inst. of N. America, Montreal, Que.
This report summarizes findings from hydrographic data primarily from the 1960’s. It
also reports results from very limited (<17 d) current meter deployments in Smith Sound.
The report begins with an excellent review of previous work in the region. With respect
to the subject at hand, geostrophic transports through sections extending across northern
Baffin Bay from Cape York to Bylot Island are computed from summer occupations in
1961-63 and 1966. A variable reference level was used with extrapolation required only
of the 1966 data. Associated heat fluxes but not freshwater fluxes are also computed.
Southward volume fluxes in the Baffin Current were 1.69-3.11 Sv with a mean of 2.28
Sv. Northward volume fluxes in the West Greenland Current were much smaller or 0-
0.37 Sv with a mean of 0.22 Sv. These yield a net southward transport of 1.54-2.74 Sv
with a mean of 2.06 Sv. It is noted that the sum of similarly calculated fluxes through the
three main passageways of the archipelago by the author and others (0.2-1.6 Sv) does not
tally to the net southward flux. This lack of continuity suggests that factors such as tidal
aliasing and barotropic motions are likely at play. In addition, it is noted that the results
are seasonally biased and winter flux information is non-existent.

Melling H. (2000) Exchanges of freshwater through the shallow straits of the North
American Arctic. In The Freshwater Budget of the Arctic Ocean, Vol. 70 (ed. E. L.
Lewis), pp. 479-502. Kluwer.
This reference provides a fairly concise summary of the difficulties of making accurate
estimates of freshwater volume flux through passages of the Canadian Archipelago and
Bering Strait. The reader gains an explicit appreciation of why volume flux estimates
remain uncertain and why freshwater flux estimates are even more tenuous despite
research efforts extending back over a century. Previous estimates are revisited to
demonstrate specific points but no new data are presented.

Prinsenberg S. J. and Hamilton J. (2005) Monitoring the volume, freshwater and
heat fluxes passing through Lancaster Sound in the Canadian Arctic Archipelago.
Atmos.-Ocean 43(1), 1-22.
Times series data from 1998-2001 are presented for western Lancaster Sound at a north-
south transect between Somerset and Devon Islands that obtains 285 m depth and is about
65 km wide. The array consisted of two groupings of 4 moorings each located at
northern and southern sites in the channel. Each closely spaced grouping housed 2
upward looking ADCP’s and 5 CTD’s. A customized compass system made from
commercially available components was devised to determine current direction in this
region, which is so close to magnetic north that direction finding based on conventional
magnetic compasses is unreliable. Measured magnetic headings were corrected for
declination variations by referencing to the geomagnetic observatory in nearby Resolute.
Tidal, seasonal and interannual variability is discussed. Generally stronger net flows to
the east are observed in the ice free summer months and at the southern side of the
channel. A number of assumptions regarding spatial extension of the mooring results to
derive fluxes are explored in the paper. It is cautioned by the authors that that since the
array was sparse, numbers must be considered preliminary. Their best estimates are an
average volume flux of 0.75+0.25 Sv, with an associated freshwater flux 1/15th of this or
0.045+0.015 Sv (1400 km3 yr-1). Sea ice was estimated to contribute less than 5% of the
freshwater flux. The freshwater component may be underestimated since time series
measurements did not include the top 25 m where CTD sections showed freshwater to be
concentrated. (It is noted that a similar array was spaced somewhat differently in the
subsequent year and also included an upper water column profiler in the southern
channel, a mid-channel moorings with upward looking sonar and an additional mid-
channel ADCP mooring. Preliminary efforts to include consideration of this later array
lead to quite similar results as reported in this paper. Details are to appear in an ASOF
book chapter).
Rudels B. (1986) The outflow of polar water through the Arctic Archipelago and the
oceanographic conditions in Baffin Bay. Polar Res. 4, 161-180.
It an acknowledged simplification of the system, the fluxes through the Archipelago are
presumed to be driven by a steric height difference of about 0.3 m relative to an assumed
level surface at 250 db (see Muench, 1971) between the Beaufort Sea and Baffin Bay and
all water passing from the Arctic Ocean to Baffin Bay is derived from the Beaufort Sea.
The author estimates net fluxes from the Beaufort Sea to Baffin Bay in two ways. One
approach models two layer flow in geostrophic balance and the other is based upon a
steady state salt, heat and mass budget for Baffin Bay. Many assumptions are required
for both the model and budget. For example, for the model, it is noted that Lancaster
Sound fulfills the requirement of being wider over all depths than the Rossby radius but
that due to narrowing at depth in Robeson Channel and Jones Sound, fully developed
flow would not be accommodated. To account for this, the flow is scaled by assuming
1.3 or 2.0 effective full capacity sounds. The 1.3 value was chosen for ease of
comparison to match modeling efforts by another author but Rudels thinks 2 might be
better. As another example for the heat budget, a yearly average heat loss through the
surface of Baffin Bay is a key parameter the value of which is cited from a private
communication. The transports for these two approaches have opposite behavior with
respect to the salinity assigned for the Beaufort Sea; the geostrophic model has lesser
fluxes for higher salinity and the budget approach has higher fluxes with higher salinity.
The salinity (32.9) for which the transports are the same by both approaches is taken to
give the best estimate of the transport or 0.7 Sv. The author considers a number of
adjustable parameters in the model and the budget and opines that it is unlikely that the
transport lies outside of the 0.5-1.2 Sv envelope.

Steele M., Thomas D., and Rothrock D. (1996) A simple model study of the Arctic
Ocean freshwater balance, 1979-1985. J. Geophys. Res. 101(C9), 20,833-20,848.
A sea ice model, assimilating velocity from buoys and concentrations fields from passive
microwave satellite data, is used to drive a model of the upper (200 m) Arctic Ocean.
Underlying geostrophic motions are computed from Levitus data that are very sparse for
large areas and in non-summer months. This model takes an approach similar to a
previously published model of the entire basin but it provides some improvement by
subdividing the Arctic into seven major regions. This allows boundary input/outputs to
have better geographic fidelity. Monthly or daily river input is treated as a salinity flux
spread over the region into which it flows. (We note that a total of 1072 km3 yr-1 runoff
is input to this model. Aagaard and Carmack (1998) tallied runoff as 3300 km3 yr–1 for
the entire Arctic but 1373 km3 yr–1 of this is due to large rivers that flow into the Barents
and Kara Seas. The Barents and Kara Seas are considered outside of the domain of
Steele’s model, however, this still leaves 855 km3 yr–1 unaccounted for and so the runoff
term would seem to be underestimated by Steele.) P-E values are taken from a central
Arctic monthly climatology published in the 1970’s and is introduced as snow until the
buoy temperatures reach 0˚C at which point it is melted directly into the ocean. Outflows
from the system are obtained by assuming geostrophic flow from the model regions
through the boundaries into waters of constant density taken to be the density of the
outflowing waters at the assigned (fixed) levels of no-motion. For the Canadian
Archipelago, the Beaufort Sea and Canada basin cells that drain through Lancaster and
Jones Sounds this depth is 150 m and for the North Pole cell which drains through Nares
St. it is 200 m. The ocean model produces salinity profiles and volume transports in the
seven regions. Thus the outflows vary with stratification at the bounding interior region.
The climatology from their model produces a 0.53Sv volume flux and 0.039Sv (1200
km3 yr–1) freshwater flux out through the Archipelago. Variability is examined and it is
noted that small anomalies in outflow through Fram St. seem to lead anomalies through
the Archipelago by a year. Experiments with increased freshwater input (as predicted in
increasing CO2 scenaria) suggest that the Archipelago continues to drain 60% more
freshwater in the form of freshened seawater than does Fram Strait, however the Fram St.
ice export remains a larger term.

Cuny, J., P. B. Rhines and R. Kwok (2005) Davis Strait volume, freshwater and heat
fluxes, Deep-Sea Res. 52:519-542.
Data from a six mooring array, deployed across Davis Strait from Sep 87-Sep 90, by the
Bedford Institute of Oceanography were analyzed along with hydrographic data collected
historically and in conjunction with the mooring deployments. Each of the moorings had
conventional current meters at three levels (150, 300 and 450 m) that also recorded
temperature and salinity. Velocity records from a single mooring tended to correlate with
each other but correlations across different mooring sites were weak. Thus the array was
too sparsely spaced to reliably capture cross-strait variability. Also there were no
moorings located over the Greenland shelf. Analyses of this data by others have
appeared previously in print, however, this work differs appreciably from previous ones
in the way that the upper 150 m of the water column are treated, which leads to
significantly higher flux estimates. From Dec to May when the Strait is ice-covered, it
was assumed the upper 150 m of the water column was homogenous in TS and so the
velocity at the top mooring was applied to the overlying water. For months without ice-
cover, hydrographic data were mined to develop monthly TS profiles for each of the
mooring locations. These were combined with the TS from the 150 m mooring to
develop daily TS profiles from which baroclinic transports were computed. The
meridional velocity from the top moored instrument was added to this. It is
acknowledged in the paper that the results depend strongly on this set of assumptions and
so entail uncertainties that cannot be quantified. The net southward volume flux derived
is 2.6+1.0 Sv and the net liquid freshwater flux is 0.092+0.034 Sv (2900+1100 km3 yr-1).
Verification awaits deployment of a more tightly spaced data that provides better vertical
coverage and particularly in the upper 150 m where the majority of their freshwater
transport appears to be concentrated. Satellite data were used to estimate a net ice flux of
528 km3 yr-1. Interestingly, this times series suggests that freshwater and volume fluxes
peak in Oct-Nov and that thoughputs from the upper Arctic Ocean are more important
than the seasonal ice cycle in driving seasonality in the transport of freshwater. The
paper ends with a brief reconsideration of the upper Arctic Ocean volume and freshwater
budgets in light of these Davis St. findings.

Münchow A., Melling H., and Falkner K. K. (2005) Observational estimates of
volume and freshwater fluxes leaving the Arctic Ocean through Nares Strait. J.
Phys. Oceanogr. submitted.
Shipboard ADCP data from the USCGC Healy August 2003 mission to Nares Strait are
presented. Data providing effectively continuous cross strait coverage and accompanied
by CTD casts on August 4 in Kennedy channel are analyzed. Sustained southerly winds
during the observation period appear to have weakened surface flow. Tidal contributions
to the flow are shown to exceed non-tidal ones by a factor of 2. Two methods of
correcting for tidal signals are presented and generally compare favorably although the
least squares fitting procedure was preferred over a barotropic model since it permits
quantification of variations in the vertical. This is of course key for freshwater flux
estimates. The net volume flux, pertinent to this observation period, was 0.8+0.3 Sv
southward accompanied by a liquid freshwater flux of 0.029+0.019 Sv (915+600 km3 yr–

Sadler, H. E. (1976) Water, heat and salt transports through Nares Strait, Ellesmere
Island, J. Fish. Res. Board. Can. 33(10):2286-2295
This was the first attempt to incorporate distributed current meter measurements into
estimates of volume flux through Nares St. Three moorings were suspended through the
ice in southern Robeson Channel with records extending up to two months in May-June,
1971. The records were interpolated in space to derive a volume flux of 0.67 Sv. This
can only be considered a very preliminary estimate, pertinent to that one period, since the
moorings were spaced so widely across the channel. Sadler attempts to compute salt and
heat as opposed to freshwater fluxes through the channel but these again are only very
crude estimates and the errors he presents are constrained largely by speculative

Stayed tuned for papers from Ron Kwok and Tom Agnew and others regarding
areal ice fluxes via satellite observations through the Canadian Archipelago. Their
findings will also be included in the ASOF book.

To top