A Hadoop cloud-based surrogate modelling framework for approximating complex hydrological models

Abstract Hydrological simulation has long been a challenge because of the computationally intensive and expensive nature complex hydrological models. In this paper, surrogate modelling (SM) framework is presented based on Hadoop cloud for approximating The substantial model runs required by design experiment (DOE) SM were solved using cloud. Polynomial chaos expansion (PCE) was fitted verified high-fidelity DOE then used as case study to investigate approximation capability in Soil Water Assessment Tool (SWAT) with regard accuracy, fidelity, efficiency. experiments, reduced computation time approximately 86% when global sensitivity analysis. PCE achieved results equivalent those standard Monte Carlo approach, flow variance coefficient determination 0.92. Moreover, proved be reliable approach but significantly more efficient. proposed greatly decreases computational costs through computing modelling, making it ideal optimization.