Abstract

Water tracks are an intrinsic part of the surficial drainage network in the foothills of the Brooks Range, Alaska. They preferentially transport water off hillslopes and represent the interplay between hydrology, vegetation, geomorphology, and permafrost characteristics. This research on mapping the location of water tracks builds on previous work which demonstrated that different types of water tracks exist due to difference primarily driven by geomorphology. We used a combination method where spectral classifications, texture, and topography were fed into random forests to identify the water track classes. The most accurate distributions were obtained for the organic‐rich and wide water track classes. The distinct linear shapes of the water tracks could also be visualized for many of the classes, especially in areas where the water tracks were particularly discrete. The biggest challenges to mapping the water tracks were due to class imbalances and high variability within and overlapping between classes. This research presents a significant step forward in understanding periglacial landscape dynamics.

Key Points

Both water track presence and different classes were spatially identified Complexity of water track classes necessitated use of machine learning Provides a critical baseline for landscape dynamics in the Arctic