North-Central Section - 42nd Annual Meeting (24–25 April 2008)

Paper No. 1
Presentation Time: 1:20 PM

IDENTIFYING POTENTIAL FACTORS CONTRIBUTING TO LEVEE FAILURE ON THE MISSISSIPPI RIVER


FLOR, Andrew, Geology, Southern Illinois Univ, 1259 Lincoln Drive, Mailcode 4324, Carbondale, IL 62901, REMO, Jonathan, Environmental Resources and Policy Program, Southern Illinois Univ, Carbondale, IL 62901-4623 and PINTER, Nicholas, Geology Dept, Southern Illinois Univ, 1259 Lincoln Drive, Carbondale, IL 62901-4324, flojo@siu.edu

A logistical regression model is being constructed to identify potential causal mechanisms contributing to historic levee failures on the Mississippi River between St. Louis, MO and Memphis, TN. Predictor variables were extracted from an extensive Geographic Information System (GIS) database for 84 historical levee failures and for an equal number of randomly selected non-failure sites. Site characteristics were derived from DEM data, surficial geologic maps, and borehole logs which were used to map landforms, fluvial stratigraphic, and geomorphic relationships (e.g., abandoned channels, initiation and inflection points of meander bends, highly permeable strata, etc.) impacting levee failure. Land-cover data from historic maps in our database were used to identify the riparian corridor between the river channel and levees and its width, type, and location. Other potential explanatory variables came from a river engineering database, and these data were used to identify levee construction irregularities or discontinuities (e.g., high-angle junctions segments, repaired levees, scour holes). Application of logistical regression to historical levee failures allows us to determine the predictor variables most strongly correlated with levee failures.

In addition, modern levee elevations were measured between St. Louis and Cairo, IL using dual-frequency GPS equipment (post-processed kinematic surveys accurate to <10 cm). We compared the current measured elevations with both the design elevations as well as measured elevations from 1994 to determine whether these levees have settled (e.g., degradation of organic-rich soils, excessive loading of soils due to enlarged levee bases, sand boil development) or if they have been raised above the design flood specifications. Levee elevations have changed significantly at some sites between 1994 and 2007. Locally changes ranged from decreases of up to 1-3 m to increases in levee elevation of up to 1 m, and other areas of little to no change. Many of these regions of height change coincide with past levee failures or towns on the floodplain. Despite the local degradation of levees at some sites, elevations of levees within the surveyed districts generally still meet or exceed their design protection level.