Ice break-up on northern rivers presents a unique research challenge because of its dynamic nature and complexity of physical, chemical and biological processes. Rapidly moving ice, ice jamming, flood levels, and enhanced flow velocities can produce significant sediment transport and trace-element fluxes. Systematic sampling for these parameters, however, is rarely conducted because of logistical difficulties. This paper discusses the magnitude and relative significance of sediment and trace-element fluxes during break-up of the Liard River in northern Canada. Historical data for the open-water and stable ice-covered periods are compared to that measured during the 1987 and 1993 break-up events. Analysis reveals that a break-up pulse occurs during this period that if not accounted, can lead to significant underestimation of suspended sediment and trace-element fluxes. More generally, any estimates of annual sediment or trace-elements for northern rivers that do not include data for the critical break-up period must be regarded as being conservative.