VISCOUS: A Variance‐Based Sensitivity Analysis Using Copulas for Efficient Identification of Dominant Hydrological Processes

Abstract

Introduction On the High Computational Cost Incurred by the Sampling-Based Global Sensitivity AnalysisWith the rapid increase in computational capability, an increasing number of distributed and semi-distributed process-based hydrologic models have been introduced to simulate the quantity and quality of water on a range of spatiotemporal scales. The ever-growing complexity of these models is driven by the

Toward SALib 2.0: Advancing the accessibility and interpretability of global sensitivity analyses

“…Beyond accessibility and understandability, it is intended that new and emerging SA approaches be targeted for inclusion, particularly if they prove capable of more efficient sampling and analyses. These include variogrambased SA methods (e.g., Razavi et al, 2019); meta-methods which combine multiple approaches (e.g., variance and distribution-based metrics; Baroni and Francke, 2020); incorporating adaptive sampling approaches (e.g., Steiner et al, 2019); and further expansion of moment-independent, adaptive, and iterative analysis approaches (Cuntz et al, 2015;Sheikholeslami et al, 2021). Improved accessibility and interpretability unlock opportunities for new research and applications that may have previously been unviable.…”

Challenges in Hydrologic‐Land Surface Modeling of Permafrost Signatures—A Canadian Perspective

“…However, the high dimensionality of LSMs and the non‐linearity of their response require a large number of samples, with high computational cost, especially as we also explore different meteorological forcing sets. To reduce the computational burden, a sampling strategy was used that conveys the maximum information from the model‐output space with a minimal sample size (Sheikholeslami et al., 2021), based on the semi‐structured parameter sampling scheme of STAR. Thus, the STAR‐based samples used for sensitivity analysis are also used to propagate the uncertainty of parameters to the simulated permafrost characteristics and to study their identifiability.…”

The delusive accuracy of global irrigation water withdrawal estimates

“…Apart from Monte-Carlo approaches (Fig. 1 c), cost-effective methods such as cheaper-to-run emulators or convergence monitoring allow a much better exploration of the uncertain space than OAT or model ensembles 22 . The outcome will be IWW estimates whose range better matches our knowledge gaps.…”