|
<br />
<br />
<br />26
<br />
<br />ROGER A, PIEU<E. JR
<br />
<br />l Prototype: the Scientific Assessment and Strategy Team (SAST)
<br />n the aftermath of the 1993 floods the White House established a Scientific As-
<br />essment and Strategy Team (SAST) of scientists from various agencies 'to provide
<br />.cientific advice and assistance to officials responsible for making decisions with
<br />espect to flood recovery in the Upper Mississippi River Basin' (SAST. 1994,
<br />J. x.iii). One of the responsibilities given to the SAST was to 'organize the infor-
<br />nation in existing databases 10 aid in the near-term and long-term decision-making
<br />xocess' (SAST, 1994, p. 232). In its review of existing data on the floodplain the
<br />5AST found that 'some data vitally imponant to making informed management de-
<br />.;isions on lhe floodplain were not readily available, or were not unifo.rmly acquired
<br />throughoullhe floodplains or river basin' (SAST, 1995, p, \3),
<br />As a prototype, the SAST has begun to collect, document, and distribute flood-
<br />plain data for the Upper Mississippi Ri....er Basin. Some of the data collected by
<br />the Team includes: hydrology, land use/cover, soil, topography, vegetation, flood,
<br />agriculture, infrastructure, climate, biology/ecology, reservoir, etc. Data has been
<br />collected from federal, regional, state, al1d local organizations: If the SAST is to
<br />contribute broadly to the U,S. flood problem by helping to fill the 'data gap' that
<br />it identified, then it must include societal data (e.g., demographics) in its mapping
<br />efforts. Further, its efforts must be evaluated from the s!<lndpoint of whether or
<br />not the information it t.:ollel;ls is uS<lb\c by decision makers. If thc Tet\ln'~ work is
<br />judged by decision makers to be useful, then the SAST is a model that ought to be
<br />emulated for other U.S, river basins.
<br />
<br />2.6. DATA ON FLOOD CASUALTIES IS A PROXY FOR FLOOD RISK
<br />
<br />Due to the lack of systematic data on th~ number of people a( risk to floods, trends
<br />in flood casualties, for which relatively systematic data is available, are sometimes
<br />used as a proxy for trends in population at risk. An assumption underlying many
<br />such analyses is that a rise in flood-related casualties is indicative of a rise in the
<br />number of people at risk to flood events, Unfortunately, at least three confounding
<br />factors limit the use of trends in flood casualties as a proxy for trends in the gross
<br />number of people who are vulnerable to floods.
<br />First, many flood-related deaths are concentrated in single extreme events, like
<br />a hurricane or a severe flash flood. Second, society has taken many steps to reduce
<br />its level of exposure, with mixed results. This means that a moving baseline of ex-
<br />posure underlies any record of flood-related casualties. Consequently, there may be
<br />a number of trends within a trend record of flood casualties (e.g:, level of exposure,
<br />success and failures of mitigation efforts, etc.), Finally, the data on flood casualties
<br />is generally not perceived to be accurate enough to lead to definitive causal conclu-
<br />sions (Richards, 1995, personal communication). The longest continuous record
<br />of flood casualty data is that of the National Weather Service (J 903-present),
<br />
<br />* Data and information on the ongoing status of the SAST can be accessed on the World Wide
<br />Web al hHp:lledcwww.cr.usgs.gov/SAST.home.html.
<br />
<br />NJNE FALLACIES OF FlOODS
<br />
<br />427
<br />
<br />600 Deaths
<br />500
<br />400 I
<br />300
<br />200
<br />100
<br />0
<br />
<br />
<br />M h - ~ ~ M ~ _ ~ m M
<br />g 8 ~ ~ m ~ ~ ~ ~ M ~ ~ ~ ~ m ~ ~ ~ ~ ~ ~ ~ -
<br />_ ~ _ _ .... ~ ~ ~ ~ m m m m m m m m m ~
<br />Figure I. Flood fatalities in the United States: 1903-1~4 -lbY-Y" - d-fi - - - - -
<br />ran ve-year average).
<br />
<br />However, there arc diITcrr.::nt soun;t:s of uutu whl'ch have d',rl' . I b I
<br />R de' . crr.:n num crs e g
<br />e ross data In FIFM!F (1992) nnd Wood (1994)). For these reasons, tre~d d~l~
<br />~n flood-r~lated cas~altles does not lend much, insight into broader .questions of
<br />acto,rs which underlie (rends on VUlnerability to floods.
<br />F.l~ur~ 1 shows the data kept by the National Weather Service on flood-related
<br />~alaJll1eS 10 the United Slales from 1903-1994.. The data shows a downward trend
<br />In flood~relat.ed deaths since the early 19705, but aJso an increased frequency of
<br />year~ WIth high de~ths. .Fi~ur~ 2 shows the trend of ilood.related deaths over a
<br />movmg 25-year penod begmmng with 1927 (j,e,. sum of 1903-1927) and ending
<br />m, 1994 (I.e., sum of 1970-1994).t At this time scale. the more recem period COn.
<br />~ms more ~eaths (Wood (l ~93), llsi?g a different dataset finds a similar trend).
<br />owever,. thIS data must be Viewed With caution, as it may be possible that part of
<br />the trend IS due to better accounting in the more recent years, Of the annual deaths
<br />~elated to floods, 80--90 pe~cent ar: caused by flash floods and 40 percent of these
<br />are related to ~tream crossmg or highway fatalities' (Zevin, 1994, p. 1267).
<br />In sum, aVailable data indicates that flood-related deaths have increased in re.
<br />cem d~ades. However, because of the nature of the data, little can be said with
<br />authonty ~bout what the trend of increased deaths means from the standpoint of
<br />people al nsk to floods.
<br />
<br />* Data. is kepI by 'water year' which rllns from October J through September 30 the foJlowin
<br />ye~r. For mstance, W~ler Year 199.6 slarted on October I, 1995 and ended September 30. 1996. g
<br />A 25-year movmg average IS used because it is the approximate period between the m t
<br />extreme /load events (as measured by economic impacts), e.g., 1903, 1927, 1951, 1972, 1993. os
<br />
|