Quantifying canopy-intercepted snow mass from tree sway observations: a demonstration over six winters in a subalpine coniferous forest

Abstract:

In forested basins, the amount of snow that accumulates on the ground depends on how much snowfall is intercepted in the forest canopy and the subsequent partitioning of canopy-intercepted snow into sublimation, unloading, and melt drip. Quantifying the amount of intercepted precipitation in forest canopies has been an outstanding challenge, and reliable observational techniques are needed to advance understanding and management of forested basins. A method for quantifying interception using wind-induced tree sway observations was proposed nearly a decade ago at AGU, and recent studies have shown the viability of the technique for rainfall in broadleaf canopies. Here we present a novel demonstration of this technique for quantifying the mass of intercepted snowfall in coniferous canopies. We installed accelerometers on coniferous trees in high-elevation subalpine forests of the Colorado Rocky Mountains and monitored tree acceleration at 12 Hz. These sensors collected data nearly-continuously for six years at Niwot Ridge (2014-2020, two trees), and one year each at Grand Mesa (2016-2017, three trees) and Senator Beck Basin (2016-2017, two trees). From these acceleration data, we derive hourly time series of tree sway from a windowed frequency analysis using Lomb-Scargle periodograms. Consistent with mechanical theory, the data show coherent drops in tree sway frequency during snow interception events, but also show increases in frequency when the trees freeze and become rigid. We demonstrate through how decreases in sway frequency can be linked to increases in mass (e.g., snow interception), and propose a temperature correction for freezing conditions. Comparing ground-based measurements to the snow mass derived from tree sway, we find that the sway technique can resolve the magnitude of snowstorms and the maximum snow interception capacity of a canopy (e.g., 20 mm at Niwot). We will discuss uncertainties in this simple methodology and ideas for scaling-up.