Reproducing dynamic sediment-discharge relationships driven by time-varying sediment availability in cold basins: the sediment-availability-transport model

Ongoing climate warming has significantly accelerated glacier-snow-permafrost melting and expanded the erodible landscapes in cold environments. The subsequently amplified sediment availability leads to the highly time-varying suspended sediment concentration (SSC) and discharge (Q) relationships. However, the sediment rating curve (SSC=a×Qb with a and b as fitting parameters) fails to capture the dynamic SSC-Q relationships and the widespread SSC-Q hysteretic patterns. To bridge this gap, we propose a Sediment-Availability-Transport (SAT) model to simulate the long-term evolution and dynamics in suspended sediment transport by integrating the sediment availability coupled by thermal processes, fluvial processes and long-term storage exhaustion into traditional rating curves. In the SAT-model, increased sediment sources from glacier-snow-permafrost erosion are represented by changes in the basin temperature, showing an exponential amplification of SSC when basin temperature increases and thermally-controlled sediment sources are activated. These thermal processes are found to be best captured by the eight-day average temperature (with one-week antecedent conditions). Enhanced fluvial erosion by the elevated water supply from rainfall and meltwater is represented by an intensification of runoff, which results in a linear amplification of SSC when runoff surge strengthens and erosion processes are enhanced by flushing and channel scouring. Such fluvial processes are found to be best captured by the two-day discharge increase. With the support of multi-decadal daily SSC and Q in-situ observations (1985-2017), the SAT-model can be validated for the permafrost-dominated Tuotuohe basin on Tibetan Plateau. Results show that sediment rating curves for Tuotuohe display significant inter-annual variations. The higher parameter-b in a warming and wetting climate confirms the increased sediment availability due to the expanded erodible landscapes and gullying-enhanced connectivity between channels and slopes. Through capturing such time-varying sediment availability, the SAT-model can robustly reproduce the long-term evolution, seasonality, and various event-scale hysteresis of SSC, including clockwise, counter-clockwise, figure-eight, counter-figure-eight, and more complex hysteresis loops. Overall, the SAT-model can explain over 75% of long-term SSC variance, outperforming the sediment rating curve approach by 20%. The performance of the SAT-model is stable, even under an abrupt hydroclimate change, with a Nash-Sutcliffe efficiency coefficient ranging from 0.76 to 0.71 during the calibration and verification periods respectively. The SAT-model not only advances understanding of sediment transport mechanisms coupled by thermal/fluvial erosion processes, but also provides a ready-to-use model to simulate and project future sediment loads for other cold basins.