Category:

Vlog

Duration: 1:13:43

2,558 views

14

Worldwide, a billion people—including those in western North America—depend on winter snowfall and subsequent spring melt for their water. In the mountains themselves, the distribution and duration of snow drive ecological processes. So how do we measure the topographic and temporal variability of snow (its water equivalent) and subsequent melt, at scales of whole mountain ranges? Direct measurements with satellites currently flying or available in the next couple of decades are not feasible. Instead, we track the seasonal progression of snow cover and its albedo with remote sensing, and we model melt rates by combining that information with assimilated climate data. When the snow disappears, we can run the model backwards to estimate how much snow existed at peak accumulation, everywhere. This information is great for scientific analysis, but not much use in forecasting. Are there patterns we can observe earlier in the season that correlate with eventual runoff? This end-to-end scenario illustrates eScience problems that combine hydrologic and computer science: physics of the processes, interpretation of surface properties from satellite measurements, management of many disparate data records that are themselves big, computations that cover dimensions of time and 3D spatial coordinates at a scale where Earth curvature matters, pattern recognition and correlation, and possibilities for people with special expertise to contribute to parts of the whole. Our goal is to assess seasonal snow resources, relative to historical trends and extremes, in mountains with meager infrastructure, sparse gauging, challenges of accessibility, and emerging or enduring insecurity related to water resources. Background reading, in case someone wants to do homework first. It's just 3 pages: Dozier, J (2011), Mountain hydrology, snow color, and the fourth paradigm, Eos, 92, 373-375, doi: 10.1029/2011EO430001, http://www2.bren.ucsb.edu/~dozier/Pubs/2011EO430001_rga.pdf