The goal of this project is to create new, state-of-the-art, physics-based models that exploit the modern datasets of urban areas commonly stored in Graphical Information Systems (GIS). Sea levels naturally rise and fall each day with the tides and are affected by other factors such as storms. But if mean sea levels are higher overall due to global warming, then the highest high tides will be higher than ever before. Along the Southern California coastline, which is the focus of this study, there are many urbanized harbor areas where the ground height is near sea level. To protect these areas from flooding by high tides and waves, infrastructure including sea walls (also known as bulkheads) and pump stations are often used. And today coastal communities are beginning to ask what else should be done to provide protection for the future. Options to consider include structural measures such as stronger coastal defenses as well as non-structural measures such as policy changes that reduce the value of assets in flood-prone areas. To evaluate these options and enable site-specific flood risk management, flooding simulations are needed and the aim of this project is to create new simulation models that are uniquely tailored to urbanized coastal landscapes. Basic research will be carried out to make physics-based models run more efficiently, to couple flood inundation models with storm drain models that account for subsurface pipe flows, and to account for the impact of waves on coastal flooding. In addition, the project will strive to implement these models within a GIS framework to support infrastructure management.
The PIs will work closely with the City of Newport Beach to ensure that model output delivers the type of information useful for decision making and to educate decision-makers about correct uses and interpretations of model generated data. They will also hold workshops for representatives of other coastal communities in Southern California to disseminate the technological advances in a targeted way and broaden the impact of the work. There is growing awareness of global climate change in California, and public officials are increasingly willing to consider pragmatic steps to address adverse impacts. However, leadership is needed for coastal communities to take cost-effective action. The PIs aim to provide this leadership, using Newport Beach as a case study to show how model generated flood data can be integrated into GIS to support flood risk management. Local governments have invested considerable resources in GIS staff and data, so the models will provide added value to this investment. This project will also advance the science of urban flood inundation modeling, answering questions about the optimal use of new data types, the necessary level of coupling between surface, subsurface and wave models, and strategies for efficient modeling. These advances stand to improve urban flood risk management technology globally. Lastly, this project will include efforts to develop flood modeling skills among under-represented groups of undergraduate engineering students and to recruit talented students for graduate study at UC Irvine.
Summary Recognizing that human populations are increasingly coastal, increasingly urban, exposed to increasingly higher sea levels and an intensified hydrologic cycle, and increasingly in need of personalized information to inform decision-making, this project advances new understanding and improved models to accurately and efficiently predict urban coastal flooding at the parcel scale. This work builds on the wealth of data characterizing urban environments, advanced computing systems and novel hydraulic flood models. Findings include: Finding 1: Bath tub models are biased towards an over-prediction of flood extent. In testing at a southern California site, bath tub models are found to significantly overpredict flood extent and depths compared with hydraulic models. This is attributed to flood events that crest over defenses for short periods of time. These results suggest that bath-tub model projections of the future impacts of sea level rise will be biased by over-prediction unless hydraulic models are used with realistic forcing and parameterized to account for infrastructure. (Gallien et al. 2011, Gallien et al. 2013) Finding 2: Limitations of aerial lidar found for urban flood prediction. The vertical errors in aerial lidar (10-15 cm) are too large to accurately capture the onset of overtopping in defended lowlands of southern California where flooding is typically caused by ca. 1-10 cm exceedances of defenses. Defense heights should therefore be surveyed with an accuracy closer to 1 cm. (Gallien et al. 2011, Gallien et al. 2013) Finding 3: Local data and knowledge required for urban flood prediction. This situation calls for a comprehensive inspection of site conditions and consultation with infrastructure maintenance and operations personnel to better understand system properties and performance history. (Gallien et al. 2011, Gallien et al. 2013) Finding 4: Couple regional and local models to forecast flooding. Regional models account for large scale atmospheric and oceanic processes that elevate local ocean levels and wave heights. In turn, local models respond to regional model output to predict localized water levels, waves, the overtopping of flood defenses, and the spreading of flood water through sewers and over land. This approach places responsibility for regional information at a regional level, and local information at a local level. (Gallien et al. 2013) Finding 5: The overtopping flow rate must be quantified to predict flooding in urbanized lowlands guarded by flood walls and embankments. Overtopping can include a weir-like flow associated with high water levels, or wave overtopping which results in a pulsed flow. Weir-like and wave overtopping are both important mechanisms relative to coastal flooding in southern California, with weir-like flows dominating in protected embayments and waves dominating on open ocean beaches. Empirical overtopping models can be nested into hydraulic flood models to predict spatiotemporal flooding patterns, but more research is needed to improve the accuracies and reduce uncertainties. (Gallien et al. 2011, Gallien et al. 2013) Finding 6: Parcel-scale flood models support structural failure prediction Damage curves based on flow depth, velocity, discharge, force and energy were nested in a parcel-scale flood model to predict structural failure and compare with damages from the 1963 Baldwin Hills dam-break flood. Results show that predicted failures vary considerably across damage curves. The best performing models correctly predict actual damages, but are biased by over-prediction rates approaching 100%. Overall, damage curves based on flow force perform the best. (Gallegos et al. 2012). Finding 7: Parallel computing will overcome the computational demands of parcel-scale flood prediction. Research reveals several options including distributed memory CPU (MPI), shared memory CPU (Open-MP), GPU and hybrid approaches. (Sanders et al. 2010) Finding 8: A new urban flood model formulation with an urban porosity improves model efficiency. This can allow use of a relatively coarse mesh of ca. 10 m which significantly reduces the total number of elements, computational work, and memory requirements compared with more commonly used 2-5 m resolution meshes of urban areas. (Sanders et al. 2008, Schubert and Sanders 2012) Finding 9: Local government GIS may contain valuable data for urban flood modeling. Valuable data for modeling studies includes digital orthophotos, classified aerial lidar data, building footprint polygons, road network polylines, and sewer network polylines and point data. (Schubert et al. 2008, Gallegos et al. 2009, Sanders et al. 2010, Gallien et al. 2011, Gallegos et al. 2012, Schubert and Sanders 2012). References Gallegos, H.A. Schubert, J.E. and Sanders, B.F. (2009) Advances in Water Resources, 32, 1323-1335. Sanders, B.F., Schubert, J.E. and Detwiler, R.L. (2010), Advances in Water Resources, 33, 1456-1467. Gallien, T.W., Schubert, J.E. and Sanders, B.F. (2011) Coastal Engineering, 58(6), 567-577. Schubert, J.E. and Sanders, B.F. (2012) Advances in Water Resources, 41, 49-64. Gallegos, H.A. Schubert, J.E. and Sanders, B.F. (2012) ASCE Journal of Engineering Mechanics, (in press). Gallien, T.W., Barnard, P.L., van Ormondt, M., Foxgrover, A. and Sanders, B.F. (2013) Journal of Coastal Research, (in press).
