1.1 A Brief History of Floodplain Management
For millennia, people have attempted to protect inhabited areas from flooding and to deliver water to areas that lacked sufficient supply. There is evidence that the first major hydraulic structure, a masonry dam across the Nile River, located about 14 miles (23 km) south of present-day Cairo, Egypt, was built around 4000 B.C. (Rouse and Ince, 1963). The increased water levels upstream of this structure enabled the diversion of flows through excavated canals to irrigate the arid lands near the Nile.
Major dams across the other great rivers in the Middle East are known to have been built earlier than 3000 B.C. by the Egyptians and Babylonians and dams and irrigation works were being constructed in China earlier than 1000 B.C. (Rouse and Ince, 1963). The Marib Dam in present-day Yemen operated for more than 1,400 years before failing in 550 A.D. (Morris and Wiggert, 1972), due to lack of maintenance. What was truly incredible about these structures and those built for several thousand years thereafter was that their designs were based solely on trial and error and the practical experiences of the builders. No hydraulic analysis was ever performed. These structures most likely lasted as long as they did, without any engineering analysis, because they were grossly overdesigned. Today, engineers try to avoid overdesigning structures because of the unnecessary costs and land use involved.
The Roman aqueducts built in approximately 100 A.D. are often cited as outstanding examples of hydraulic structures; and yet, the Romans had no insights into the relationships between slope, velocity, and discharge (Herschel, 1913). In fact, their writings indicate that they believed cross-sectional area was the main variable that determined the discharge; increasing or decreasing the slope was apparently not understood as affecting the discharge capacity of the aqueduct. Figure 1.1 shows one of the few intact remains of the Roman aqueducts. Some of these aqueducts were still in service long after the fall of the Roman Empire.
Figure 1.1 Roman aqueduct.
The first uniform flow formula for calculating channel design velocity was formulated in 1768 by the French engineer Antoine Chézy and was used for the design of a water-supply canal for Paris, France (Rouse and Ince, 1963). More than 100 years later, an Irishman, Robert Manning, modified the Chezy equation, and the four main equations (continuity, energy, momentum, and Manning) for floodplain hydraulic analysis were then established. These same equations are still used today. However, even into the beginning of the twentieth century, hydraulic-structure design generally continued to reflect practical engineering experience rather than hard computations using the four fundamental equations.
The first 30 years of the twentieth century saw significant progress in floodplain hydraulic analysis. In addition to channel design computations with Manning's equation, laboratory model studies in Europe demonstrated the applicability of physical models in riverine studies. Research with physical models was often performed in direct response to hydraulic problems encountered in the field and their use also became more common for addressing hydraulic questions that were analytically indeterminate.
Following the disastrous flood on the Lower Mississippi in 1927, the U.S. Army Corps of Engineers (USACE) founded the Waterways Experiment Station (WES) in Vicksburg, Mississippi to support hydraulic studies for the Lower Mississippi and, later, around the country. In the United States, floodplain hydraulic analysis with physical modeling was pioneered by WES, with a physical model of most of the Mississippi River Basin constructed on a 200 acre (81 ha) site near Clinton, Mississippi, shown in Figure 1.2. Much of the levee construction along the Mississippi River was based on the results of the design water levels simulated with the Mississippi Basin Model (MBM) during the 1950s and 1960s.
Figure 1.2 A physical model of the Mississippi Basin, looking downstream from south of St. Louis, Missouri. The end of the model (Baton Rouge, LA) is near the water tower in the background.
By World War II, analytical hand computations to calculate water surface profiles routinely used the continuity, energy, momentum, and Manning's equations. However, hand computations were time-consuming and engineers often spent days or weeks completing a water surface profile analysis for a reach of river.
By the 1960s, the first simple, automated procedures were developed to make water surface profile computations less painful to the hydraulic engineer. Early programs used geometric data for selected cross sections of the river and floodplain and required analysis of bridge effects to be performed by hand outside the program. A great improvement was the development and initial release of a FORTRAN version of the USACE's Hydrologic Engineering Center (HEC) program "Backwater, Any Cross Section" in 1966. This program was revised, expanded, and rereleased in 1968 as HEC-2, Water Surface Profiles.
With the release of HEC-2, subcritical and supercritical flow profiles incorporating bridge and levee effects and other modeling concerns could now be analyzed in a straightforward manner within one program. Similar programs were developed in the 1970s and 1980s by different U.S. agencies, including WSP2 by the Natural Resources Conservation Service (formerly the Soil Conservation Service), WSPRO by the U.S. Geological Survey (USGS), and E431/J635 by the USGS.
Of all the river hydraulics models, HEC-2 was the most widely applied. HEC-2 was one of the very first open channel hydraulics programs available and could incorporate bridge and culvert analyses and other hydraulics modeling components. Even more important, the program was well documented and supported by the USACE's Hydrologic Engineering Center.
By the 1980s, certain methods and procedures in HEC-2 did not use the most-accepted routines for some computations, especially for bridge and culvert calculations. When the 1990s arrived, the program was still largely based on the mainframe computers of the 1970s and, although HEC-2 had been converted to run on personal computers by 1984, the data input and output were still based on punch card format. HEC-2 did not incorporate easy-to-use templates for input, as is common today with personal computers. The HEC began the development of a replacement program in 1991, with the maiden release of HEC-RAS (River Analysis System) Version 1.0 in 1995.
Updates to HEC-RAS have been periodically released and the product is still being actively developed. Among the many improvements since the initial version of HEC-RAS are channel modification analysis, mixed-flow capabilities, bridge scour analysis, WSPRO bridge analysis procedures, ice jam hydraulics, lateral and inline weir analysis, hydraulic simulation of gated structures, modeling of changes in Manning's n in the vertical direction, and geographic information system (GIS) integration capabilities. HEC-RAS will likely be the prime computational tool used over the next few decades for river hydraulics work, especially with the inclusion of full unsteady flow analysis capabilities in 2001. Future additions to the program will include greatly improved hydraulic-design features, such as the computation of riprap (rock revetment) requirements and sediment transport, scour, and deposition analysis.
HEC-RAS is the most widely used floodplain hydraulics model in the world (Wurbs and James, 2002) and is emphasized in this book as the primary tool for performing floodplain modeling. HEC-RAS and most other floodplain hydraulic programs use steady, gradually varied flow computation procedures, which are also featured in this book. While most floodplain modeling studies can be adequately addressed with steady-flow techniques, HEC-RAS can also simulate unsteady flow situations (covered in Chapter 14). Chapter 3 discusses steady versus unsteady flow simulations and the types of situations they are appropriate for.
