FLOOD09459

• County: Statewide
• Basin: Statewide
• Title: Nine Fallacies of Floods
• Date: 1/1/1999
• Prepared By: National Center for Atmospheric Research
• Floodplain - Doc Type: Educational/Technical/Reference Information

-'-'" _. .__.'~-'_.__..

420

ROGER A. PIELKE, Jft

2.3. LEVEES 'PREVENT' DAMAGES

The United States, one flood expert has commented, has had an 'undyin.g love af-
fair' with levees (Tobin, 1995), Another flood expert has called the late-eighteenth!
early nineteenth century policy of sole reliance on levees to control ,floods 'perni-
cious beyond belief' (Wright, 1996, p. 247). A.levee is a stru:ture bull~ lO keep w~-
ter from inundating an area that would otherwise be flooded. The United States IS
home to approximately 25,000 miles of levees - or enough to encircle the world at
the equator (FIFMTF. 1992). The U,S. Army Corps of Enginee~s has been respon-
sible for building more than 10,500 miles of levees. Levees bu.ilt by other fed~ral
and state agencies, as well as by individuals, make up the remamder. T~e fundIng,
design, construction, and repair of levees by the federal government IS overs~en
by five federal agencies: USACE. Agriculture, National Resources C?nservatlOn
Service, t Federal Emergency Management Agency (FEMA), Economic Devel.op-
ment Administration, and HUD (Housing and Urban Development). The mu~ttple
agencies and their differing requirements and responsibilities 'confuses' the Issue
of levee systems (Tobin, 1995, p. 364). Local agencies and privately-owned levees

further complicate the issue.
According to Tobin (1995, p. 365) levees can serve to actuall~ increase the
potential for flood losses. He calls this the 'levee effect' and defines II as follows:

Once [a levee] has been constructed, however, the structure may'generate a
false sense of security to the extent that floodplain inhabitants perceive that all
flooding has been eliminated. With the incentive to take precautions removed,
few residents will be prepared for remedial action in the event of future floods.
Even more costly, however, this false sense of security can also lead to greater
development in the so.called sufe areas, thus adding to the properly placed
at risk. .. when the levee does fail, the increase in development can actually
raise losses even higher lhan if no levee syslem had been constructed in the

first place.
The 'levee effect' can increase society's vulnerability to floods in two ways: by
creating a sense of complacency, which can act to reduce p~eparedn~ss and by
creating incentives to build structures in areas subject to floodlOg. ToblO observes
that a comprehensive study of the 'levee effect' has yet to be undertaken, .
The fallacy that 'levees prevent damages' is a subset of a much broader Issue
about the role of structural measures in flood mitigation. Structural approaches to
flood control are characterized by attempts to keep high water from inundating pub-
lic and private prope~y and thus causing damage. Since the I 940s. be~inning. with
the pioneering work of Gilbert White, floodplain mana~ement has mcre~slOgly
included 'nonstruclural' approaches that emphasize adjusting human behaVior (see.

. Both levees and f100dwalls are defined as a fonn of 'dike' - a general term for 'longitudinal
structures thaI serve to retain water' (FlfMTF, 1992, p. 12-27). This paper uses Ihe more common
term '\e"ee' to refer to aU IYpes of levee.~, lloodwalh, and dikes.
t Fonnerly the Soil Conservation Service.

NINE FALLACIES OF FLOODS

421

FIFMTF (1992). Chapters 12 and 15 for discussion), Issues of structural and com.
plementing nonstructural approaches to flood mitigation go well beyond the scope
of this paper. It is worth noting, however, that in spite of widespread agreement
among flood experts that structural measures alone cannot adequately address the
nation's flood problem, the proper role of structural measures has yet to be ade-
quately defined. Lack of understanding of the 'levee effect' is a prime example
of how an incomplete understanding of the role of structural and nonstructural
measures in flood mitigation limits what can be authoritatively said with respect 10
the flood problem.

2.4. FLOOD FORECASTS ARE UNIVERSALLY AVAILABLE

The Unified National Program for Floodplain Management report asserted that
'regional warnings for floods and hurricanes are becoming universally available'
(FIFMTF. 1994. p, 16), This stands in stark contrast to Changnon (1996. p, 309)
who stated with respect to the 1993 Midwest flood that 'the flood condition pro-
nouncements of the National Weather Service (NWS) and those of the Corps of
Engineers for the 51. Louis area wen~ often inaccurate and sometimes controver-
sial' (d. Gruntfest and Pollack, 1994). As a consequence, 'these forecasts and
pronouncements likely worked to the detriment of the strategies used to fight the
flood, shipping alternatives, and other flood-impacted endeavors' (p. 27). Further-
more, even the Weather Service's self evaluation found 'substantial opportunities
for improvements' (DOC, 1994, p. 2-1). More recently, residents of the Red River
of the North basin have criticized the flood forecasts issued during the spring 1997
floods (see, for example, Glassheim, 1997). In short, in spite of the progress made
over the last century in weather and flood forecasting, there remains considerable
potential for improvements in flood forecasts. Further. there also remains a vast
potential for improved use of those forecasts by decision makers (Pielke, 1999;
Changnon, 1996; cf. Gruntfest and Pollack, 1994),
One potential area for improvemeQt of flood forecasting lies with continuing
improvements in 'Quantitative Precipitation Forecasting' (QPF), which is defined
as 'the forecasting of the amount of rain or water equivalent of frozen precipi[Ution
that will fall over a particular area over a given period' (Schwein, 1996, "po I). Such
improvements have been described as 'steady, albeit slow' (Krzysztofowicz, 1995,
p. 1143). The goal of using QPF in the flood forecasting process is to provide a
longer lead time for flash flood and river stage forecasts. Of course, the success of
this tool depends on large part upon the ability to forecast reliably the amount of
precipitntion that will full over a particular watershed. To date. QPF has not been
fully implemented. Results have been mixed in experimental uses of QPF with
both improvements and degradation in forecasts. One review cites several studies
of the use of QPF that show benefits to decisions associated with flood warning,
reservoir control, and commercial navigation (Krzysztofowicz, 1995). According
to Schwein (1996) a key to beneficial use of QPF in flood forecasting will be to train