Field testing of three bedload samplers' efficiency in a gravel-bed river, Spitsbergen

A comparative study of the three bedload traps was accomplished to determine the effectiveness differences between the devices. The research was carried out in two transverse sections of the proglacial, gravel-bedded Scott River (SW Svalbard) using three devices to perform direct measurements of bedload transport in the river channels. Several samplers are used for field measurements, but none of them has gained widespread acceptance as a standard so far. The paper compares transport rates collected in three bedload samplers that differ in structure and functionality: i) a portable Helley-Smith, pressure-difference bedload sampler (H-S), ii) an anchored River Bedload Trap (RBT), and iii) the portable sampler used by the Polish Hydrological Services (PIHM type C). All three samplers are constructed using the same components: a frame made of thin stainless steel sheet and a nylon bag with the same (2 mm) mesh size (except PIHM). Measurements were conducted within three consecutive days in two cross sections, in the same vertical profiles and at the same time (at intervals of 30 and 60 min). Measurements of bedload transport and water velocity were performed simultaneously. In order to determine the effectiveness and representativeness of short-term measurements using particular devices, the masses of the samples were converted into a specific rate of bedload transport and compared to the results of long-term continuous measurement (24 h measurement cycle). The results confirm that both specific transport indicators and the mass of the samples collected by the tested devices show a significant variation. Generally, the highest mass and specific transport indicators were obtained using the H-S sampler. The RBT produced moderate transport rates, while the PIHM sampler gave the lowest rates. The bedload transport rate in a profile estimated on the short-term measurements was compared to results of the continuous measurements conducted in the same cross sections. In the 7 cases we obtained underestimation (to max. 10 times) and in the next 7 cases - overestimation (to max. 114 times). Only in four cases were the results of short-term measurements similar to the continuous ones. The results obtained from the RBT sampler show the smallest variability among all tested devices.