Multiphysics urban flood models are commonly used for urban infrastructure development planning and evaluating risk due to climate change and sea level rise. However, these integrated flood models rely on several parameters that are hard to measure directly, and the resulting uncertainty in model prediction needs to be quantified, often without observable data. As a part of the Urban Flooding Open Knowledge Network (UFOKN) project, in this study we quantify parametric uncertainty in urban flood models. UFOKN incorporates flood model predictions in combination with machine learning, data and computer science, epidemiology, socioeconomics, and transportation and electrical engineering to minimize economic and human losses from future urban flooding in the United States. As a case study, we choose the Interconnected Channel and Pond Routing (ICPR) numerical model to simulate flooding in the city of Minneapolis in response to the design storms (e.g., 100-year rainfall). Through a sensitivity study, we reduce the number of uncertain model parameters to the Manning’s roughness coefficient and vertical hydraulic conductivity of soil, and construct the distributions of these parameters using open databases. We employ the multilevel Monte Carlo (MLMC) method that combines a small number of high-resolution ICPR simulations with a larger number of low-resolution simulations to reduce the computational cost of computing the key statistics of the quantities of interest describing the urban flooding. Our results show that the uncertainty in the flood predictions (as described by the coefficient of variation of the flood water depth) is distributed highly non-uniformly in the urban area with the coefficient of variation exceeding 0.5 limited to a relatively few computational elements in the ICPR model. Our results demonstrate that urban flood models such as ICPR can provide reliable flood predictions and can be used for a targeted data acquisition to further reduce the parametric uncertainty.

Physics of two-phase flows in heterogeneous rocks plays an important role in many applications such as oil and gas migration and geological sequestration of carbon dioxide. Although current pore-scale models are used to compute macroscopic properties required in reservoir simulators, most work is limited to small sample size and homogeneous rocks. There is a need for pore-scale modeling approaches that can accurately represent the 3D complex pore structure and heterogeneity of real media. Pore network modeling simplifies the geometry and flow equations at pore-scale, but can provide characteristic curves in capillary-dominated systems on fairly large samples with huge saving on computational costs compare to direct numerical simulation methods. However, there are limitations for attaining a large representative pore network for heterogeneous cores, namely the technical limits on sample size to discern void space and computational limits on network extraction algorithms. To address these issues, we propose a novel stochastic pore network stitching method in combination with network generation to provide large-enough representative pore network for a core. Our approach proposes to use micro-CT images of various reservoir rock cores in different resolutions to characterize the pore structure. Few signature parts of the core are selected and their corresponding void space and equivalent pore network are extracted. The space between pore networks is filled by using a stochastic network generator that utilizes statistics of all signature networks and a layered stitching method that glues networks based on their average properties. The output is a large network that can be used in any pore network solver. We are focusing on flow properties via quasi-static pore network modeling solver to obtain absolute permeability, relative permeability, and capillary pressure curves. Since the method is stochastic, the workflow should be run on many realizations and final results yield both average and variability of the derived properties. We have tested the developed method on various generated and extracted networks, and we have extended the stitching method to 3D heterogeneous samples.