QUATERNARY GEOSCIENCE AND FLOOD HAZARD

Quaternary geology and geomorphology have greatly aided quantitative assessment of geologic hazards. Recent work includes paleohydrological reconstructions from floodplain deposits, and documentation of the extent of tsunamis and storm surges. However, the sedimentary record is often discontinuous, and the needs for accurate risk assessment for a 7.1+ billion global population are increasing. Flood protection or population resettlement may cost billions of dollars; millions of inhabitants are at risk; standard risk evaluation methods are sometimes insufficiently informed or based on out-of-date science. Examples: Flood time series commonly include diverse genetic populations of events (e.g. New England tropical storms and spring snowmelt floods), yet hazard maps based on standard flood frequency analysis do not incorporate this knowledge. Non-stationarity of flood probability distributions, widely-regarded as the reality, is not yet incorporated within standard numerical procedures. Flood hazard along many sediment-rich rivers is increasing, due to decreasing capacities of channels and reservoirs, but risk methodologies do not include this information from geoscience. The 2008 Kosi River, India, and the 2010 Indus River, Pakistan flood catastrophes were caused by levee breaching (avulsion) at relatively modest flood flows: neither flood frequency analysis nor standard flood hazard mapping techniques could have predicted the severity of events. Yet the switching behavior of the Kosi channel and the historical avulsions along the Indus are both well documented. Finally, model- and topography-based prediction for Sandy's (2012) storm surge extent correlates closely with published post-surge maps. The post-surge maps are, however, based on topography-guided interpolation; they must agree with surge model predictions due to the methodology involved in producing both. Remote sensing, in contrast, shows numerous areas of disagreement with both the predictive modeling and the post-event "mapping”. Remote sensing can provide an objective record: the actual reality against which future hazard should be evaluated. Advanced surface process modeling also provides a path towards improved, geoscience-based prediction of flood inundation and channel processes, including avulsion.