Using river ice time series (MODIS-AQUA) to identify and predict ice hazards

Ice jams are stationary accumulations of ice and as such may occur wherever below zero temperatures persist for an extended duration. Ice jams may develop as ice becomes arrested and further accumulates and alters the river flow. While the vast majority of ice jams are harmless and resolve on their own, severe cases may cause significant flooding on short time scales. Much effort has been made to develop physical and statistical models to better predict ice jams, but the ability to make accurate predictions or transfer the models to other sites have been mixed. One limitation of the models is that that river ice observations, which directly relate to ice jams, are either not available or sparse in space or time. Satellite based approaches to estimate locations, and to a lesser extent the amount of ice, may provide this information at large spatial scales with a mean revisit time of a few days. In this work we explore the how satellite ice observations may be used to better model and predict ice jams at Harrisburg, PA.