IRREGULARITY OF THE GREAT LAKES WATERSHED BOUNDARY: QUESTIONABLE INFLUENCE OF TOPOGRAPHIC RELIEF

The boundary of the Great Lakes / St. Lawrence River watershed exhibits a somewhat higher degree of map-view trace irregularity than observed in prior results for the Continental Divide in the conterminous U.S. The complexity measure used is C_T, percentage difference in length estimates for a trace obtained by divider walks at 1 and 10 km step sizes. While C_T values for 65 segments of the Continental Divide trace range from 3% to 45%, the values obtained for 50 similar-length segments of the Great Lakes watershed boundary range from 16% to 64%. This runs counter to an expectation that more active dissection in a region dominated by high mountains will produce more fine-scale irregularity in its watershed boundaries.

Documentation of local relief associated with each segment of the Great Lakes watershed boundary allows more direct testing of its relationship to boundary complexity. Relief within a 5 km radius of each watershed boundary polygon node was assessed from 20 m contour coverage, and these values were averaged for each boundary segment. Strong relief contrasts among sets of segments in three major physiographic regions do not consistently predict shifts in watershed boundary complexity. While the relatively small group of Central Lowland segments (n = 7) have a somewhat lower mean C_T value of 29%, Canadian Shield and Appalachian regions share mean C_T values of approximately 40%. Plotting of segment C_T values against corresponding relief values suggests that relief accounts for an insignificant amount of the variation in watershed boundary complexity in the full data set (R² < 0.01), in a subset of segments excluding relatively rare high-relief cases (R² < 0.1), and in physiographic province subsets (R² < 0.25). Derived trends in these regressions are inconsistent, further supporting a general conclusion that watershed boundary complexity has little relationship to regional relief value.