Flood Risk and Climate Change: The Contributions of Remote Sensing

Abstract:

Since the mid-1970s, satellite observation has gathered an exceptionally valuable but largely un-harvested record of flood inundation world-wide. Commencing in late 1999, the two MODIS sensors also obtained daily surveillance of all of the Earth’s surface waters. These data are analogous to the record of earthquake seismicity provided by seismographic stations; they provide the only objective characterization of many extreme, damaging flood events. This information should be deployed to its maximum utility in defining areas of flood risk. In the developing nations, the remote sensing archive provides the immediate opportunity, without hydrological data infrastructure, to directly identify hazardous land areas. As well, satellite passive microwave radiometry, commencing with near-daily global coverage in 1998, has the ability to characterize at-a-site flood hydrographs. When combined with the satellite record of mapped inundation, this allows exceedance probabilities to be placed on observed inundation limits. The coupled data set can then be used to validate predictive flood modeling. As climate changes, flood statistics change. Yet hazard evaluation has for many decades proceeded using assumed stationarity of flood frequency distributions. New floods-of-record at any location thereby present a dilemma to policy makers and to hydrologists: immediately include the new extreme flood in the flow series, and thus increase the size of the regulatory floodplain, or use the pre-flood flow records to label the exceptional new event as, for example, “the 1000 yr flood”. The remote sensing record also includes defended floodplains where levees have failed, sometimes even during relatively common floods. We can use the powerful observations provided by remote sensing to confront the old probability estimates directly: by arguing that the recent observed record of inundation from actual floods must take priority in guiding public policy.