The need for a transit model
foreign passage. CRREL was also assigned to over-
The various reports from each of the primary
see the coding for the model, perform the simula-
team members were delivered to the Alaska Dis-
tions, and report the results. CRREL subcontracted
trict in the fall of 1994. It was the task of the Alaska
the actual work of writing the computer code to
District, in the second year, to assimilate the in-
Bronson Hills Associates (BHA) of Fairlee, Ver-
formation from these reports into overall recon-
mont. In addition to being skilled computer pro-
naissance recommendations. Due to the complex-
grammers, the BHA principals were two veteran
ity of the data accumulated in that first year, the
Arctic researchers for the U.S. Navy and were ex-
Study Manager further identified a need for a tran-
perienced in sea-ice processes and navigation.
sit model that would help to compare the NSR
All participants worked closely at all stages of
with the conventional southerly routes. That is,
the model's development, from defining its pur-
the NSR's shipping costs would be compared with
pose through flow-chart conception to final ex-
those of the Panama and Suez Canal routes.
ecution. Direction and decision-making was ac-
The USAED Study Manager then selected UAF
complished through constant communication that
and CRREL to combine their respective environ-
included telephone, facsimile, and electronic mail.
mental and operational databases to produce a
During the course of the modeling work, the par-
computer model that would estimate NSR transit
ticipants gathered at two workshops. The first was
times and costs. The Study Manager guided the
held at the beginning of the project, and the sec-
modeling investigation and provided the conven-
ond occurred approximately midway through its
tional cargo ship cost data that were the basis for
development.
estimating similar costs of the ice-strengthened
vessels that are needed for NSR shipping. USAED
On-site workshops
provided the rationale to adjust these warm-water
Two 3-day workshops were held at CRREL in
ship costs upward to reflect added construction
Hanover, New Hampshire. The first took place in
and operational costs for ice-going vessels.
December 1994, during which the USAED's needs
UAF assembled the model's climatological in-
were fully explored in light of the collective capa-
put, which portrays the meteorological and oceano-
bilities and the data available. We defined the vari-
graphic conditions of the route. The data were
ous legs of the northern route to be modeled and
available to predict the probability of occurrence
established a set of ship routing decision nodes,
of winds, wave heights, ocean currents, ice condi-
where two or more alternative route choices exist.
tions, and visibility factors (such as fog, snow-
The NSR is not a unique shipping channel but
storms, atmospheric icing, and darkness). Each of
rather is generally regarded as any and all pos-
these were simulated as functions of time and lo-
sible routes from the Atlantic to the Pacific Ocean
cation.
through the passages, open seas, and island groups
The fact that much of the environmental data
north of the Eurasian land mass. We agreed on the
were available in the form of probability distribu-
types of ships to model and the class of Russian
tions was the major reason we chose to use a Monte
icebreaker to serve as the cargo ship's escort when
Carlo modeling technique. This method derives
needed. We agreed to consider cargo transits in
its name from the city on the French Riviera where
three different seasons to simulate best-case, worst-
games of chance and gambling are popular. In the
case, and intermediate transit scenarios. We estab-
fashion of gambling, Monte Carlo simulation "rolls
lished a timetable, discussed the sources and suit-
a die" (randomly selects) for the existence or mag-
ability of data, constructed a preliminary flow
nitude of variables from their respective probabil-
chart, and agreed on the task assignments noted
ity density functions at each trip node. The voy-
in the previous subsection.
age is repeated a large number of times (100500)
The second workshop was held in February
to allow the time and cost calculations to reflect all
1995, during which a preliminary version of the
probable conditions to be encountered on a typi-
model was tested and modified. The model proto-
cal voyage. In other words, the model produces
type provided feedback on data suitability, the pro-
Gaussian distributions of voyage time and cost
gram algorithms, and indicated where refinement
parameters.
was needed. At this time, we incorporated in-ice
CRREL, having recently completed studies of
ship performance criteria that we formulated from
the history and current administration of the NSR
1) consultation with Lawson W. Brigham, USCG,
(Mulherin et al. in prep.), provided input on ship
an experienced captain of a U.S. Coast Guard Po-
operational capabilities and NSR cost factors for
lar-class icebreaker and noted expert on Russian
3
Previous Page
