Professor University of California Davis, United States
Abstract: Mechanistic models improve understanding of the complex interactions between abiotic and biotic factors impacting riparian cottonwood seedling recruitment. The predictive capabilities of mechanistic models are important for guiding river management and restoration designs, but quantifying the reliability and uncertainty associated with predictions is equally as important. This research investigates hydrophysical processes that impact cottonwood seedling recruitment and presents a methodology that translates the mechanistic model from theoretical exploration to a tool that can be used for its predictive capabilities. Four hydrophysical processes were identified from the literature to be modeled: preparation of germination sites, recession rate, prolonged inundation, and scouring flows. A spatially explicit, mechanistic model was developed to represent the previously conceptualized hydrophysical processes and to predict the potential for seedling recruitment, both for real and design scenarios. The Riparian Seedling Recruitment model is coded in Python-3 and fully integrated into the open-source River Architect software. The model uses two-dimensional (2D) hydrodynamic numerical model results and a daily flow record to assess the recruitment potential for a given topography and substrate.
The lower Yuba River (LYR) in northern California was selected as the pilot site given the availability of high-resolution topographic data, substrate and vegetation mapping, and hydrodynamic numerical model outputs. Using the mean daily flow record from a hydrologic model output under current operational constraints on the LYR and the 2017 topography, 93-years of Populus fremontii seedling recruitment potential was analyzed over ~34 km of the river. For the set of parameters selected for initial analysis, relationships between the hydrophysical processes and the overall recruitment potential indicates that a lack of suitable germination sites is limiting potential seedling recruitment across the LYR as well as within each distinct geomorphic reach when analyzed with 95 years of hydrologic model outputs. Further sensitivity and uncertainty analyses will evaluate the assumptions associated with the model structure and assess the relative importance of individual and sets of parameters. Results from these analyses will facilitate calibration and validation of the model with observation data and evaluate the model’s predictive accuracy before it is used for future applications.
