December 2007
·
22 Reads
A state-wide digital elevation model (DEM; www.atlas.lsu.edu) based on 1999-2002 LiDAR data is widely used in assessing present and future flooding potential in New Orleans and southeast Louisiana due to storm surge. Although the data were acquired during a time when official vertical controls had been deemed unreliable due to subsidence, the DEM continues to be used for operational modeling and planning. To test its viability, an accuracy assessment of the DEM was performed in 2007 using a statewide real-time kinematic GPS system based on NOAA-sanctioned continuously operating reference stations. Sampling was focused on built structures, i.e., levees, floodwall, and roads, features considered to control surge flow in the low-lying coast. Over 100,000 points were measured and compared to 5X5 m DEM pixels. The vertical accuracy of test points was determined to be +/-10cm (0.3 ft). It is claimed that 90% of the DEM is to accurate to +/-15 cm (0.5 ft).The study had the added benefit of providing a snapshot of the progress being made in augmenting the regional hurricane protection system following the 2005 storms. The study shows that only 40 percent of DEM samples pass the accuracy test, i.e., are +/- 0.8 ft of the true elevation. Where the DEM is too low, it is likely due to: levee augmentation, floodwalls are too narrow to be detected by LiDAR, new levees, and a "levee crown bias", i.e., sampled DEM pixel includes levee slope areas. Where the DEM is too high, the causes can be traced to two factors, inaccurate vertical controls established prior to LiDAR acquisition, and to a lesser degree, post-acquisition subsidence. The DEM south and east of New Orleans overestimates levee elevations by 0-1 m. We conclude, therefore that the state-wide digital elevation model (DEM) is unreliable and inadequate to support present-day surge modeling.