Large-scale Clustering of Natural Snowfall: Collective Precipitation Dynamics in Three Dimensions

The interactions between large collections of settling snowflakes and various turbulence intensity levels within the air column make snow precipitation difficult to forecast. Characterizing the multi-scale spatial distribution and transport of snowflakes is crucial for understanding the spatial modulations in the snow deposition process and for interpreting remote sensing signals. In this work, we perform large-scale three-dimensional tracking of snowflakes falling through the atmospheric surface layer in the Swiss Alps. We utilize a novel super-resolution field imaging system that combines 16 high-resolution cameras mounted on arrays and is flexibly deployed in ice-fishing tents at different instrumented field sites with collocated snow and wind characterization. Each camera array is fitted with shifted lenses to stitch an equivalent 100 Megapixel imaging over a 20x20 square Meter field of view at a 2-Millimeter diffraction-limited tracking resolution. Snowflakes are illuminated using white light of 5500 Kelvin at 250′000 Lumens from multiple powerful 1575 Watt stadium floodlight panels mounted on snowboards and retrofitted with lenticular lenses. Shooting data at a 150 Hertz, the system is capable of tracking millions of snowflakes over 10x10x10 cubic Meters simultaneously. We first present collective snow tracking data obtained in a mild wind vector of approximately 3 Kilometers per hour. Analyzing the fall velocity, our data suggests a multimodality for fast and slow falling snow particles, which we discuss in relation to recorded snow particle variability. Subsequently, analyzing the point-cloud data using a Voronoi tessellation, we find a predominance of clusters and voids compared to the clustering diagram for a random Poisson process. Secondly, we present field experiments being caught in a blizzard with windspeeds exceeding 30 Kilometers per hour. We first conduct a qualitative assessment of the observed patterning of snowfall in the atmosphere at high wind speeds, as well as the appearance of saltation and blowing snow layers during the field measurements. We then identify signatures of these field observations in the acquired tracking data and compare events of extreme clustering dynamics against those of the cluster diagram for the mild wind vector.