Multi‐Level Monte Carlo Models for Flood Inundation Uncertainty Quantification

Abstract Flood events are the most commonly occurring natural disaster, with over 5 million properties at risk in UK alone. Changes global climate expected to increase frequency and magnitude of flood events. hazard assessments, using projections as input, guide policy decisions engineering projects reduce impact large return period Probabilistic modeling is required take into account uncertainties model projections. However, dichotomous relationship between probabilistic modeling, computational cost resolution limits applicability such techniques. This paper examines improvements traditional Monte Carlo methods Latin hypercube sampling (LHS) Multi‐level (MLMC) quantify uncertainty extent resulting from input hydrograph uncertainty. The results demonstrate that MLMC a more efficient strategy than current (i.e., Carlo) high outputs produced less time previously possible. novel application technique three Scottish case studies, demonstrating variety river characteristics, domain sizes costs, m grid resulted 99.2% reduction compared up 2.3 times speedup Hypercube Sampling.