kilometers and found in roughly the same loca-
Probabilities of ice thickness for April, June,
tions each summer, are known as ice massifs. Apart
August, and October are read from ICTHCK**.DAT
from these regional accumulations, the interannual
files (Table C.6) using subroutine GETDAT. These
extent of the ice cover is markedly variable. In
PDFs contain five categories: ice free, <120 cm,
summer, much of the coastal route may be entirely
120180 cm, 180240 cm, and >240 cm. For sim-
ice free, although the straits still are more likely to
plicity, these ranges are converted to discrete thick-
have ice. Other summers have resulted in very
nesses by NSRSIM01: 0 cm, 60 cm, 150 cm, 210 cm,
little melting such that ships needed nearly con-
and 240 cm. Ice thickness PDFs are updated at
tinuous escort. Ice concentration, thickness, and
each data point from the MAIN program, and a
ice pressure are the major direct factors influenc-
new thickness is selected at 8-hr intervals using
ing ship speed. These three characteristics of the
the MC algorithm via subroutine ICETHICK.
ice cover were included in the database used to
simulate transit navigation.
Ice pressure
Ice pressure, or ice compression, is one of the
Ice concentration
most important factors that can slow ship speed
We used ice concentration data for August and
or even stop an icebreaker (Buzuev 1977, Voevodin
April taken from Romanov (1993). For June and
1981a, b). We simulated ice compression and its
October, we digitized this information from the
probability along the NSR on the basis of atmo-
Sea Ice Climatic Atlas of USNOCD (1986a, b) and
spheric pressure for the period from 1946 to 1988.
from Arctic and Antarctic Sea Ice, 19781987
We assumed the divergence of the drift velocity of
(Gloersen et al. 1992). In the Chukchi Sea region,
ice to be proportional to divergence of the wind
the ice concentration data were corrected using
after Doronin and Kheisin (1977). That is
information from Alaska Marine Ice Atlas (LaBelle
et al. 1983) and USNOCD (1986a, b). We input
Pi = Ap[div (Vi)]
(2)
the probabilities of ice concentration for April,
June, August, and October to the model from
where Pi is ice pressure, div is operator of diver-
CONC**.DAT files (Table B.2) using subroutine
gence, Vi is ice velocity, and Ap is a coefficient of
GETDAT. Concentration PDFs contain five cat-
ice compression where
egories: ice free, 1030%, 4060%, 7080%, and
90100%. For simplicity, these concentration
Ap = 0 if div (Vi) < 0
ranges are converted to discrete concentrations
Ap = 107 if div (Vi) > 0 .
by NSRSIM01: 0%, 20%, 50%, 75%, and 100%. Ice
concentration is updated at each data node from
On the other hand, the ice divergence is inversely
the MAIN program, and a new concentration is
related to the divergence of atmospheric pressure:
selected at 8-hr intervals using the MC algorithm
via subroutine ICECON.
d d2P d2P
div (Vi ) = -K
+
dt dx2 dy 2
(3)
Ice thickness
We obtained ice thicknesses for April and Au-
where d/dt is a time derivative, d2/dx2 and d2/dy2
gust from Romanov (1993). We computed our ice
thickness data for October using the equation of
are second-order space derivatives, and P is atmo-
Zubov (1944):
spheric pressure. In general, the coefficient K de-
pends on the compactness of the ice and diver-
gence. In a region of low atmospheric pressure, an
I2 + 501 8R = 0
(1)
increase in compactness (i.e., convergence of ice)
where I = ice thickness (cm)
takes place, and in a region of high atmospheric
R = cumulative freezing degree-days (C).
pressure, thinning occurs.
For simulation purposes, we categorized ice
The air temperature information required for cal-
pressure into four levels of severity: no ice pres-
culating cumulative freezing degree-days was
sure (when div (Vi) < 0) and low, medium, and
taken from Proshutinsky et al. (1994).
high ice pressure. We thus assigned probabilities
We interpolated or extrapolated ice thickness
of occurrence for each category at each data node
for June using April and August observations from
on the basis of atmospheric pressure data from
coastal and island stations.
NCAR (1990).
24
Previous Page
