Monitoring, modeling, and an image processing algorithm to quantify river ice formation processes on the Red and Assiniboine Rivers

Abstract

Ice formation processes have been monitored on the Red and Assiniboine Rivers, Manitoba by establishing an extensive network of hydrometric equipment and trail cameras. Although much of the study area is often subject to mechanical freeze-up jams, thermally driven border ice growth is typically observed just upstream of the Assiniboine’s confluence with the Red River. Unlike mechanical border ice growth, thermal border ice forms slowly and is better suited for data acquisition through time-lapse photography. Using photogrammetry and an automated digital image processing algorithm developed for trail camera imagery, one-dimensional border ice growth rates have been quantified. The presented combination of tools and techniques provide high frequency data that is a safe and economic alternative to more traditional site surveys. Processed time-lapse photos have been used to support the enhancement of a two-dimensional numerical model for thermal border ice growth.