Table 2. Percent probability of wind speed and direction for mesh point 2A (Cape Zhelaniya) in August.
Direction
Prob.
Max.
Min.
Wind speed (m/s)
(degrees)
(%)
(m/s)
(m/s)
05
510
1015
1520
2025 2530 3035
3540
4045
4550
>50
000045
15.6
19.3
6.6
5.2
8.4
1.7
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
045090
14.1
20.8
7.2
4.4
6.3
2.6
0.8
0.1
0.0
0.0
0.0
0.0
0.0
0.0
090135
10.8
15.4
5.9
4.5
5.3
0.8
0.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
135180
12.4
17.4
6.1
5.4
5.6
1.1
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
180225
12.3
18.5
6.3
4.7
6.2
1.1
0.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
225270
10.5
16.1
6.0
3.8
5.9
0.6
0.2
0.0
0.0
0.0
0.0
0.0
0.0
0.0
270315
13.0
19.3
6.1
5.2
6.1
1.4
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
315360
11.3
20.2
6.1
5.2
4.9
0.8
0.4
0.1
0.0
0.0
0.0
0.0
0.0
0.0
days, which is probably more realistic. In the cost
General description
calculation, however, we round the number of es-
and assumptions
We wrote our transit model in the FORTRAN77
cort hours to the next greater whole day.
programming language. We assembled and stored
Transit routes. The transit routes that we selected
our data in companion files and lookup tables.
for simulation are designed to cover the range of
The main program is approximately 1000 lines long
paths that might be followed if the NSR were to
and utilizes 21 subroutines. The user can choose
become heavily traveled. We recognize that some
any number between 1 and 500 repeating voyages
of our transit legs are rarely used at present. Our
from which to generate the summary statistics for
objective was to evaluate the full range of costs
transit time and total cost. We used a DOS-based
possible if demand were sufficient to warrant open-
platform running on a 33-MHz 486-SX desktop
ing new routes that now see little or no traffic. We
computer. With a math coprocessor, the model per-
also assume that transit from Murmansk to the
forms 500 repetitions of one set of voyage param-
Bering Strait is non-stop; that is, the model does
eters in approximately 2 minutes if the short out-
not currently allow for intermediate ports of call
put format is selected. The various output formats
to pick up or discharge cargo. The model is de-
are discussed below, under Running the Model. The
signed, however, so that intermediate stops at Si-
following assumptions concerning icebreaker es-
berian ports could easily be accommodated. Fi-
cort, transit routes between Murmansk and the
nally, the calculation of ship's heading needed to
Bering Sea, and the degree to which simulated con-
sail from one data node to the next is performed
ditions reflect real-world conditions underlie the
only at the node embarked from. It is not updated
model.
between nodes to correct for route deviations
Icebreaker escort. In practice, icebreaker escort is
caused by wind, waves, and currents. Transit dis-
mandated by Russian authority in some NSR lo-
tances are calculated along great circle routes, one
cations where navigation is usually difficult. We
consequence of which is the fact that, except for
did not program any voyage segments to always
due northsouth and due eastwest travel, com-
require escort. We triggered the need for escort
pass headings change continuously en route. We
only when the MC-selected conditions reach cer-
assume that the effect of these factors on ship mo-
tion is minor and not important to our overall re-
that escort is instantaneously available when
sults.
needed. In actual practice, delay in a voyage may
Simulated conditions. An artifact of the MC
occur while waiting for an icebreaker to arrive or
method that is inherent in our model is that, to the
to form up convoys of ships. Third, our "escorted"
extent that underlying PDFs permit, conditions set
ship speeds are those of a single ship under escort;
in two consecutive time steps are independent of
i.e., convoys are not considered. Convoys that are
each other. It is thus conceivable that one 8-hr tran-
slower than a single-ship escort might be able to
sit segment with clear skies, no ice, unlimited vis-
transport cargo at a lower cost, but to analyze this
ibility, and light winds might be followed by the
possibility, more complex ship performance tables
next having 2.5-m-thick ice at 100% concentration,
would have to be incorporated. Fourth, our model
with a foggy gale wind causing topside icing. Se-
progresses in 8-hr time segments with ice condi-
quences of events generated in a few segments of
tions being reexamined during each segment. This
a few voyages may not necessarily reflect real-
can result in required escort for isolated 8-hr seg-
world conditions accurately. However, in the more
ments of the transit rather than for consecutive
global sense, when a multivoyage transit is con-
8