Suspended sediment in a high-Arctic river: An appraisal of flux estimation methods.

Abstract

Quantifying fluxes of water, sediment and dissolved compounds through Arctic rivers is important for linking the glacial, terrestrial and marine ecosystems and to quantify the impact of a warming climate. The quantification of fluxes is not trivial. This study uses a 8-years data set (2005-2012) of daily measurements from the high-Artic Zackenberg River in Northeast Greenland to estimate annual suspended sediment fluxes based on four commonly used methods: M1) is the discharge weighted mean and uses direct measurements, while M2-M4) are one uncorrected and two bias corrected rating curves extrapolating a continuous concentration trace from measured values. All methods are tested on complete and reduced datasets. The average annual runoff in the period 2005-2012 was 190±25mio·m3y-1. The different estimation methods gave a range of average annual suspended sediment fluxes between 43,000±10,000ty-1 and 61,000±16,000ty-1. Extreme events with high discharges had a mean duration of 1day. The average suspended sediment flux during extreme events was 17,000±5000ty-1, which constitutes a year-to-year variation of 20-37% of the total annual flux. The most accurate sampling strategy was bi-daily sampling together with a sampling frequency of 2h during extreme events. The most consistent estimation method was an uncorrected rating curve of bi-daily measurements (M2), combined with a linear interpolation of extreme event fluxes. Sampling can be reduced to every fourth day, with both method-agreements and accuracies <±10%, using 7year averages. The specific annual method-agreements were <±10% for all years and the specific annual accuracies <±20% for 6years out of 7. The rating curves were less sensitive to day-to-day variations in the measured suspended sediment concentrations. The discharge weighted mean was not recommended in the high-Arctic Zackenberg River, unless sampling was done bi-daily, every day and events sampled high-frequently.