contents.gifindex.gifprev1.gifnext1.gif

Summary of Working Paper No. 150-1999

I.1.6 A Time and Cost Prediction Model for Northern Sea Route Shipping

By N.D. Mulherin, D.T. Eppler and D.S. Sodhi, U.S. Army Cold Regions Research and Engineering Laboratory, Hanover, New Hampshire, USA.

In an earlier study, we created a Monte-Carlo based numerical model that calculated the time and cost required for shipping by way of the Northern Sea Route for three cargo ship types during the four seasons of the year. The model used historical probabilities of occurrence for sea-ice cover, meteorological, and oceanographic conditions to establish the physical environment at each transit way point in the model. It then set the speed of the ship for the upcoming trip segment or the need for icebreaker escort according to a series of decisions related to the physical conditions. Shipping costs were related to the type of ship, the time of year, how long the transit takes, and how much of the time an icebreaker escort is needed.

For this study, we conducted an extensive sensitivity analysis of the transit model to reveal the most significant parameters determining the shipping time and cost. We provided this information to other researchers to help them decide which variables to include and their degree of resolution in their creation of an improved transit model. We then modified the CRREL model, further segmenting its spatial grid by a factor of ten and adding the necessary environmental data files that were provided to us from another INSROP project. Obtaining and incorporating more recent cost data or ship performance criteria were outside the scope of work for this project. The cost data used herein were those which was available as of 1993. The cost calculations therefore are useful only in a relative sense. Our new model results indicate that better information in these two areas is needed.

In this report, we compare the structure, the data, and the logic used in the old and the improved CRREL models. We describe our sensitivity analyses and identify those variables having the greatest influence on time and cost. We present and discuss the new model results, and compare those with results from the original CRREL model.

Whereas our earlier model results agreed reasonably well with historical experience, these were somewhat less successful in that way. The new results were similar to the earlier model for April transits but their August transits differed significantly. In fact, the new August time and cost results were not much different from those calculated for April. This was contrary to historical experience, which indicates that shipping in August requires half the time for shipping in April.