Dmitrenko et al.,' Oceanographic Causes for Transarctic lee Transport
85
The sum of freezing degree-days S( -T) was calculated using mean daily air temperature from
October to June 1994-1995 and 1995-1996 at the polar stations Terpyai-Tumus, Dunai, Tiksi,
and Kotel'nyy (Figure I). The convective heat exchange was estimated using the ratio given in
(I ).
The amounts of heat passing through the surface due to convective processes are rather small.
Calculations show that density stratification in the river water outflow zone prevents the
development of convective heat exchange (st. KD9502 and KD9503 in Figure SA). However,
at the periphery of the outflow zone, where density stratification is weaker, the convective heat
exchange becomes much more efficient. In the outflow zone, about 10% of the heat
accumulated in the intermediate layer can be transferred to the surface due to convective heat
exchange. At the periphery, this value increases up to 40-60% (st. KD9504 in Figure SA). As a
rule, the convective heat exchange is more efficient at the beginning of the winter season up to
the time the fast ice edge becomes quasi-steady. On average, during the first three months of ice
formation, the amount of transferred heat is four times as large as during the next four months
up to the time the ice growth ends (Figure SA, position I, II). When the fast ice edge becomes
stable, the ice growth is reduced by 27 and 7 cm at the periphery of the river water outflow
zone and beneath it, respectively (process 2, Figure 7). At the time the ice growth ends, these
values increase by 15 and 3 cm, respectively. On average, these values are even less at the
periphery of the outflow zone and do not exceed 10 cm at the time the ice edge becomes stable.
Heat exchange governed by double-diffusion
The criterion for the possible occurrence of double-diffusion convection is a certain value of the
density ratio, Rr = bDS'aDT-i, for the density boundary, where temperature makes an unsteady
contribution to the density gradient. Here b = S·1O- 4 is the coefficient of salinity compression,
with dimensions inverse to salinity, a = 7.10- 5 [DC-I] is the temperature expansion coefficient,
and DT and DS are the salinity and temperature differences at the boundary. In our case, these
are the temperature and salinity differences across the upper boundary of the warm intermediate
layer (the upper part of the seasonal pycnocline). It was found that Rr = 15 is the upper limit for
the development of double-diffusion convection and accompanying step-like thermohaline
structures (Fedorov, 1976; Huppert, 1971; Neshyba et aI., 1971). At Rr > 15, exchange
through the boundary surface becomes purely molecular. Estimates of the actual density ratio at
the upper boundary of the seasonal halocline at the time preceding ice formation, Rr (only for
transect I); at the beginning of ice formation, Rrf; and at the time the fast ice edge becomes
quasi-steady, Rrj, are presented in Table 2. To calculate the density ratio, Rrj, the thermohaline
characteristics of the upper sub-ice layer calculated using formulae (4) - (S) were used.
Our estimates show that double-diffusion convection before ice formation is impossible (st.
KD9502-KD9509, Table 2). During the initial period of ice formation, the double-diffusion
convection is possible only at the periphery of the river outflow (st. KD9505, KD954S in Table
2). When the fast ice edge becomes quasi-steady, the area expands (st. KD9502, KD9503,
KD9505, KD9507, KD9509, KD954S, KD9549). However, in the central part of the river
outflow zone, the double-diffusion is either extreme or absent during the entire winter season
due to the high stability of the seasonal pycnocline.
The heat exchange values governed by double-diffusion have been calculated using a ratio
obtained experimentally by Turner (1965, 1973). Based on the Turner's experiments, Huppert
(1971) suggested a formula for calculating the convective heat flux which expresses the
dependence of this flux on the Rr value:
Ft = b . Ft* . Rr- 2,
(10)
85
The sum of freezing degree-days S( -T) was calculated using mean daily air temperature from
October to June 1994-1995 and 1995-1996 at the polar stations Terpyai-Tumus, Dunai, Tiksi,
and Kotel'nyy (Figure I). The convective heat exchange was estimated using the ratio given in
(I ).
The amounts of heat passing through the surface due to convective processes are rather small.
Calculations show that density stratification in the river water outflow zone prevents the
development of convective heat exchange (st. KD9502 and KD9503 in Figure SA). However,
at the periphery of the outflow zone, where density stratification is weaker, the convective heat
exchange becomes much more efficient. In the outflow zone, about 10% of the heat
accumulated in the intermediate layer can be transferred to the surface due to convective heat
exchange. At the periphery, this value increases up to 40-60% (st. KD9504 in Figure SA). As a
rule, the convective heat exchange is more efficient at the beginning of the winter season up to
the time the fast ice edge becomes quasi-steady. On average, during the first three months of ice
formation, the amount of transferred heat is four times as large as during the next four months
up to the time the ice growth ends (Figure SA, position I, II). When the fast ice edge becomes
stable, the ice growth is reduced by 27 and 7 cm at the periphery of the river water outflow
zone and beneath it, respectively (process 2, Figure 7). At the time the ice growth ends, these
values increase by 15 and 3 cm, respectively. On average, these values are even less at the
periphery of the outflow zone and do not exceed 10 cm at the time the ice edge becomes stable.
Heat exchange governed by double-diffusion
The criterion for the possible occurrence of double-diffusion convection is a certain value of the
density ratio, Rr = bDS'aDT-i, for the density boundary, where temperature makes an unsteady
contribution to the density gradient. Here b = S·1O- 4 is the coefficient of salinity compression,
with dimensions inverse to salinity, a = 7.10- 5 [DC-I] is the temperature expansion coefficient,
and DT and DS are the salinity and temperature differences at the boundary. In our case, these
are the temperature and salinity differences across the upper boundary of the warm intermediate
layer (the upper part of the seasonal pycnocline). It was found that Rr = 15 is the upper limit for
the development of double-diffusion convection and accompanying step-like thermohaline
structures (Fedorov, 1976; Huppert, 1971; Neshyba et aI., 1971). At Rr > 15, exchange
through the boundary surface becomes purely molecular. Estimates of the actual density ratio at
the upper boundary of the seasonal halocline at the time preceding ice formation, Rr (only for
transect I); at the beginning of ice formation, Rrf; and at the time the fast ice edge becomes
quasi-steady, Rrj, are presented in Table 2. To calculate the density ratio, Rrj, the thermohaline
characteristics of the upper sub-ice layer calculated using formulae (4) - (S) were used.
Our estimates show that double-diffusion convection before ice formation is impossible (st.
KD9502-KD9509, Table 2). During the initial period of ice formation, the double-diffusion
convection is possible only at the periphery of the river outflow (st. KD9505, KD954S in Table
2). When the fast ice edge becomes quasi-steady, the area expands (st. KD9502, KD9503,
KD9505, KD9507, KD9509, KD954S, KD9549). However, in the central part of the river
outflow zone, the double-diffusion is either extreme or absent during the entire winter season
due to the high stability of the seasonal pycnocline.
The heat exchange values governed by double-diffusion have been calculated using a ratio
obtained experimentally by Turner (1965, 1973). Based on the Turner's experiments, Huppert
(1971) suggested a formula for calculating the convective heat flux which expresses the
dependence of this flux on the Rr value:
Ft = b . Ft* . Rr- 2,
(10)