Numerical simulations of Columbus’ Atlantic crossings

Abstract

The transatlantic route of Columbus was simulated incorporating historical winds, currents and hypothetical magnetic variation in order to estimate where the first landfall occured. Earlier simulations using an 1899 map by Van Bemmelen and assuming zero magnetic variation in the Bahamas to produce a landfall near San Salvador (Watlings Island). New theories postulating a Geometric league of 2.67 nautical miles and a westerly magnetic variation of approximately one point (11.25°) for the western terminus result in a landfall near the Turks and Caicos Islands. A westerly variation of this magnitude in the Bahamas has been inferred from early charts - the islands are shown several degrees too far north, which would have occurred if early navigators had been set imperceptibly southward by westerly variation - by studies of directions within the islands, and by studies of early navigation books. The simulation of subsequent voyages by Columbus lend further weight to a westerly variation of about one point in the region of Bahamas. Our work shows that a Grand Turk landfall cannot be ruled out based on the transatlantic portion of the voyage as has been suggested in the past. A more accurate simulation of the voyage and first landfall still requires a more accurate representation of the field of magnetic variation.