Statistical Classification of Self-Organized Snow Surfaces

Abstract

Wind-swept snow self-organizes into bedforms. These bedforms affect local and global energy fluxes but have not been incorporated into Earth system models because the conditions governing their development are not well understood. To address this difficulty, we created statistical classifiers, drawn from 736 hr of time-lapse footage in the Colorado Front Range, that predict bedform presence as a function of wind speed and time since snowfall. These classifiers provide the first quantitative predictions of bedform and sastrugi presence in varying weather conditions. We find that the likelihood that a snow surface is covered by bedforms increases with time since snowfall and with wind speed and that the likelihood that a surface is covered by sastrugi increases with time and with the highest wind speeds. Our observations will be useful to Earth system modelers and represent a new step toward understanding self-organized processes that ornament 8% of the surface of the planet.