On the challenges of adopting adaptive Monte Carlo techniques in regional risk assessment

This paper discusses the challenges encountered when adopting adaptive Monte Carlo (MC) techniques in regional risk assessment settings.The MC estimation needs to be established in this case for a very high-dimensional output, corresponding to different locations/assets within the region.The challenges originate from the fact that for the adaptive characteristics of the MC algorithms, a compromising solution needs to be established across the quantities of interest (QoIs) representing these outputs.Two different algorithms are discussed, adaptive importance sampling (AIS) and adaptive multi-fidelity Monte Carlo (AMFMC), for a specific application, the real-time probabilistic predictions for the anticipated surge for landfalling tropical storms/cyclones.These probabilistic predictions are made through an uncertainty quantification process that involves: (i) generating a sufficiently large ensemble of storm scenarios based on the nominal storm advisory and the anticipated forecast errors; (ii) performing high-fidelity numerical simulations to obtain surge predictions for each storm scenario; and (iii) estimating surge statistics of interest by assembling the simulation results.This process is repeated whenever the nominal storm advisory is updated.AIS and AMFMC are considered here to improve the uncertainty propagation efficiency, and reduce number of storm scenarios utilized in the analysis.A direct comparison between these frameworks is established, focusing on the challenges encountered in tuning the algorithm adaptive characteristics to provide probabilistic estimates across a large number of QoIs, corresponding to the surge predictions for different locations within the coastal region of interest.Different compromise solutions are promoted to accommodate the conflicting preferences the QoIs represent.The efficacy of the two frameworks is examined in detail in this setting, comparing the accuracy of idealized implementations (adaptive decisions independently made for each QoI) to the accuracy of practical implementations (single, compromise decision within the MC implementation).The study also showcases the importance of information sharing across storm advisories in real-time probabilistic storm surge predictions and provides guidelines for an efficient adaptive MC formulation in such settings.