An iceberg drift and decay model is used to compute the ice-rafted debris (IRD) flux to the glacial North Atlantic under steady iceberg discharge from the Greenland and Laurentide Ice Sheets. I investigate the sensitivity of modeled IRD distribution pattern to formulation of glacial iceberg decay, meltwater drainage of Laurentide Ice Sheet at its southern margin, and the choice of ocean current and wind fields used as model boundary conditions. Modeled IRD patterns are compared with the observed basic glacial mode IRD pattern <Ruddiman, 1977> and the last five Heinrich layer patterns <Grousset et al., 1993>. Model results show that the first-order control on IRD deposition pattern is the prevailing ocean currents, and the choice of current fields is critical in determining the results. Winds can have significant effect when they exert drag in concert with ocean currents. Formulation of glacial iceberg decay directly affects its survival and the expanse of IRD distribution, but the formulation is perhaps one of the largest uncertainties in the model. Drainage of the southern margin of the Laurentide Ice Sheet through the Laurentian Channel by meltwater instead of glaciers predicts no significant IRD deposition in the open North Atlantic in contrast to observations, suggesting that glacier (i.e., iceberg) drainage through the Laurentian Channel was significant during the times of observed IRD deposition. In most experiments, the model predicts a high IRD deposition band that roughly coincides with the latitudinal band of high Heinrich layer deposition but are measurably south of the basic glacial mode high IRD deposition band. This discordance suggests either that the flow fields used as model boundary conditions are inaccurate or that much of the basic glacial mode IRD was not delivered by icebergs but perhaps sea ice.¿ 1997 American Geophysical Union