A Sensitivity Analysis Approach to Identifying Drivers of Streamflow Hysteresis

ABSTRACT This study reveals the critical interplay between streamflow hysteresis and local hydrogeomorphic conditions. Hysteresis is known to be influenced by characteristics such as bed slope, roughness, and wave intensity, but there has been no comprehensive study of the conditions facilitating hysteresis. Our 1D HEC‐RAS study was conducted to highlight the hysteresis response to local hydro‐morphological changes acting in isolation or combination. Wave intensity, backwater condition, bed slopes (0.0001 < S 0 < 0.001), and roughness (0.02 < n < 0.2) in simple and compound channels were gradually varied from a base scenario of the Illinois River at Henry, IL. A Random Forest analysis revealed that hysteresis is controlled 28% by wave intensity, 23% by bed slope, 16% by roughness, and 15% by backwater. We found that bed slope and backwater are the most persistent controls on the expression of hysteresis, with mild bed slopes and high backwater consistently generating strong hysteresis signals. We also explored the impact of various discharge estimation techniques on modeled streamflow, demonstrating how flow boundary conditions that account for hysteresis produce a more accurate response in the modeled reach compared to those provided by conventional methods. In understanding the sensitivity of streamflow hysteresis and its drivers, we narrow the gap between our evolving knowledge of flow dynamics and strategies for monitoring, modeling, and forecasting rivers under natural unsteady conditions.