current on iceberg drift. Hand plots suffered from a number of problems, including the fact that 

 wind and currents were the only parameters considered. In addition, gathering and plotting 

 collected information was time-consuming (IIP, 2004). In 1971, IIP began using a computerized 

 model that employed the vector-addition routine described in Morgan (1971). The model area 

 selected for this routine covered the area from 40° to 52° North and from 39° to 57° West. This 

 model helped eliminate cumulative errors associated with hand plotting and improved the ability 

 to model all icebergs accurately (IIP, 2004). 



In 1980, the Coast Guard Research and Development Center released a report that 

 discussed the physics involved with iceberg deterioration. Theoretical Estinuite of the Various 

 Mechanisms Involved in Iceberg Deterioration in the Open Ocean (White, Spaulding, & 

 Gominho, 1980) supplied equations and figures for the erosion of icebergs due to buoyant 

 convection, wind-forced convection, insolation, and wave erosion. This report gave rise to the 

 ICEPLOT computer program, which was used to predict iceberg drift and to predict the melt of an 

 iceberg over time. 



In 1992, IIP implemented the precursor to BAPSNT, a system that the Canadian Ice 

 Service (CIS) had been using since 1986. This system permitted estimation of iceberg-melt rates 

 based on real-time sea-surface temperature (SST), significant wave height, and significant wave 

 period. The BAPSNT deterioration algorithm uses the parameters of stability, water velocities, 

 drag, surface melt due to insolation, air convection surface melt, buoyant vertical convection, 

 forced convection, calving, and wave erosion to compute estimated iceberg melt. The 

 combination of wave erosion and convection based on SST is responsible for over 90% of the 

 deterioration of icebergs (El-Tahan, El-Tahan, & Venkateshl, 1987). Any assumptions that are 

 made by the model are conservative, meaning that the model will calculate an iceberg to melt 

 more slowly than it would in the ocean (Anderson, 1983). 



The most recent advance in the modeling of iceberg deterioration from the Ice Patrol 

 perspective is the Canadian Ice Service's implementation of BAPSNT 1.7 in January of 2004. 

 This model is based on the same prediction routines as BAPSNT 1 .4, but results have not been 

 tested to ensure that re-coding did not alter the algorithms or induce sensitivity differences. 

 Sensitivity differences in the two models would result in the programs generating slightly 

 different prediction results, which have yet to be compared or understood. It should be noted that 

 although newer versions were implemented in Canada, IIP continued to use version 1.4 

 operationally. 



In this study, we used iceberg size, wave erosion, and SST to review iceberg-deterioration 

 results computed by the What-If Model (WIM) of BAPSNT 1.4 and compared them to the WIM 

 of BAPSNT 1.7. The WIM provides the ability to deteriorate selected icebergs, while applying 

 pre-selected environmental data to gain estimates of percent melt versus time. Estimates of 

 percent melt versus time were used to compare the algorithms/melt routines used by BAPSNT 1 .4 

 and BAPSNT 1.7. Additionally, the sensitivity of melt estimates to each of the environmental 

 parameters were briefly analyzed. One of the main goals of this work was to ensure that version 

 1.7 is an operationally acceptable replacement for version 1.4. The drift comparisons will be 

 completed in another project. 



Methods 



In order to complete a successful evaluation of the models, an appropriate study area had 

 to be chosen. A location off the northern Grand Banks was chosen based on the large number of 

 icebergs that flow through the region, giving the area the nickname "iceberg alley." The specific 



58 



