Can remote sensing transform our use of snowmelt resources?

Current forecasts of snowmelt runoff depend mostly on multiple regression between measurements of snow water equivalent and subsequent streamflow. Like all predictions based on statistics, they work reasonably well most of the time, but the law of large numbers gives rise to occasional significant errors. Streamflow and snow accumulation both have long-tailed distributions, and the errors in forecasts of seasonal melt and shorter-term melt rates also have long tails. Historically, estimating precipitation generally and snow accumulation and melt specifically constitute the most important observational and modeling problems in mountain hydrology. Water resource managers now consider new techniques to measure snow water equivalent and albedo, but adoption of more accurate information on snow accumulation and variables affecting melt rate depends on their conformance with existing forecast procedures. In regions of the world with austere surface infrastructures, remotely sensed information can foster qualitative improvements for users of water.