A Computationally Efficient Method for Estimating Multi‐Model Process Sensitivity Index

Identification of important processes a hydrologic system is critical for improving process-based modeling. To identify while jointly considering parametric and model uncertainty, Dai et al. (2017), https://doi.org/10.1002/2016WR019715, developed multi-model process sensitivity index. Numerical evaluation the index using brute force Monte Carlo (MC) simulation computationally expensive, because it requires nested structure parameter sampling number simulations on order (N being samples). reduce computational cost, we develop new method (here denoted as quasi-MC brevity) that uses triple sets samples (generated sequence) to remove in theoretically rigorous way. The reduces from 2N. performance assessed against MC approach recent binning by through two synthetic cases groundwater flow solute transport Due its theoretical foundation, overcomes limitations imposed inherently empirical nature method. We find outperforms both terms requirements aspects, thus strengthening potential assessment indices subject various sources uncertainty.