Extended Data Fig. 1: Discharge and water temperature seasonality on the Koyukuk River (Alaska) and theoretical predictions for the timing of riverbank erosion.
a, Discharge climatology for the Koyukuk River at Hughes (66.04696° N, 154.26097° W) based on data from the USGS streamflow station during the period 1962–1981 (Extended Data Fig. 7a). Note that 1 ft3/s is equal to approximately 0.028 m3/s. Discharge peaks during the spring freshet in late May to early June. Some years have a second discharge peak associated with August rains (Extended Data Figs. 3 and 7a). The Koyukuk River maintains very low discharge from late October to mid-May, when the surface of the river is frozen. b, Average water temperature time series from the USGS gauge at Pilot Station on the Yukon River (61.93369° N, 162.88293° W). The USGS gauge at Hughes does not record water temperature, which is why we rely on the Pilot Station temperature record. However, comparison of water temperatures measured by HOBO loggers deployed on the Koyukuk River near Huslia during the summers of 2022 and 2023 show that the water temperature at Pilot Station is a good proxy for the water temperature on the Koyukuk River. Water temperatures approach 0 °C during the river-ice ‘break-up’ and ‘freeze-up’ periods, and peak in mid-July, at a time when the water discharge approaches its summertime low (a). c–e, Theoretical predictions for the sub-seasonal patterns of riverbank erosion under the endmember scenarios that erosion is controlled by: ice gouging during break-up69,70,71,72 (c), the thawing of pore-ice in frozen bank sediments14,26,27,73 (d) and the ability for flowing river water to entrain bank sediment14,36,37,38 (e). The time series in c is an illustrative cartoon. The break-up period in May is probably the time of greatest erosive action from ice21, although the freeze-up period in October can proceed in fits and starts, during which thin ice lenses flow downstream and could erode thawed riverbanks. The uncertainty envelopes in d and e propagate the discharge and water temperature variability in a and b using Monte Carlo simulations.
