A Multi-Model Nonstationary Rainfall-Runoff Modeling Framework: Analysis and Toolbox
We present a framework and toolbox for multi-model (one at a time) nonstationary modeling of rainfall-runoff (RR) transformation. The designed time-varying nature of the five available conceptual RR models in the toolbox allows for modeling processes that are nonstationary in essence. Nonstationary Rainfall-Runoff Toolbox (NRRT) delivers insights about underlying watershed processes through interactive tuning of model parameters to reflect temporal nonstationarities. The toolbox includes a number of performance metrics, along with visual graphics to evaluate the goodness-of-fit of the model simulations. Our analysis shows that the proposed time-varying RR modeling framework successfully captures the nonstationary behavior of the Wights catchment in Australia. A multi-model analysis of this catchment, that has endured deforestation, provides insights on the functionality of different conceptual modules of RR models, and their representation of the real-world.
Sadegh, Mojtaba; AghaKouchak, Amir; Flores, Alejandro; Mallakpour, Iman; and Nikoo, Mohammad Reza. (2019). "A Multi-Model Nonstationary Rainfall-Runoff Modeling Framework: Analysis and Toolbox". Water Resources Management, 33(9), 3011-3024. https://doi.org/10.1007/s11269-019-02283-y