FLOOD08791

• County: Statewide
• Basin: Statewide
• Title: ANCOLD Guidelines on Risk Assessment
• Date: 8/1/1998
• Prepared By: ANCOLD Working Group on Risk Assessment
• Floodplain - Doc Type: Educational/Technical/Reference Information

reflection, because there are only 26 data cases; and situations, that in reality are very complex, are
represented by only three parameters (Population at Risk - PAR, Warning Time - WT and Force). A
proper understanding of the D-M approach'requires a careful study of the data cases on which the
procedure is based, as reported by the US Bureau of Reclamation (1989). Warning Time requires great
care in estimation, since it depends on many uncertain factors, such as the availability of formal Dam
Safety Emergency Plans and Warning/Evacuation Plans through interaction of the owner, the emergency
authority and the community, the reliability of warning hardware, response of emergency services
personnel and status of public communications systems, typically under emergency conditions (severe
flood or earthquake). Users of the D-M procedure need to be aware that the High Force and Low Force
curves cross over at a Warning Time of about 1.7 hours. Also, for low Warning Time and low Population
at Risk, the approximate formula given by D-M is seriously in error and the accurate formulae should be
used for such cases. There is a need for an experienced person to make a sanity check of the predicted
loss of life result, to ensure that it is realistic in terms of the circumstances of a specific case. There is a
view that the D-M result should be taken as a base estimate and then adjusted upwards or downwards
according to the particular circumstances of the case at hand, taking account of factors as set out in
Section 3-1 of US Bureau of Reclamation (1989). The merits of such an approach will be discussed in the
new guidelines. In the new guidelines there will be advice on interpretation of results where predicted
loss of life is less than one. There will also be guidance on how to deal with variations in PAR over time
of day and season of the year.

Table 1

Confidence Limits for the DeKay and McClelland Data Cases

Dam Fa~ufe/Fiash Fklod Events POp..llation rswamin (WTl FlocCil'1gforce- Actual lOss Pre d 'oss 01 tife ( ,
2.1r1sk Estimated Used for lutness index ollifa Mean c-, nff enceintetv"l
(PARI orecictina lOl {See nole 1) flOLl Lowerlimir Uooer limit
I AI: h , un PA,198 22CO . 0 0 9 I 0 100
Austin TX 1981 "' , , " 9 1 , 7
nHilI A I 161::iQO 1 , 5 . I 0
. 8earwallow NC 19i~ , 0 0 I . I 0 ,
5 ; Thomasan CO 1 c;76 <1.n 0:< I ". 5' . '62
, Slack Hi~s SD 1~7? 170 <1.0 .5 I . 174 I 10 25
7 uff,,~f"':r"''''< WI/ 197 <I I I" 7 6 1 7'
, Bu c:hv Hill Pond CT 1982 '00 .3 0 0 I 0 6
9 CentraH WA 1991 IS . , I 0 20
10 Denver .1965 10000 2.33-4.0 17 0 I I I 0 2'
11 K " ; M 1977 "'0 <11' ....". 1 20 . I . HO""O
12 Kellev Barn A.19n 250 <0.5 n~25 1 31 I 2 170
13 L rfAnPA1977 '50 0 0 I 4() 4() 12'
" Lawn " 1982 5JlOO .0.1. 0 10'
I' 09' "K<> MA tQh:::! 1 "" I ? 71
16 little De9r r elo: UT 1 50 0 0 I I I 0 10
17 Maloa set Fran 1 59 0000 -n n , . , '06 I 2 "14'6
" Mot'! an P r~ CT 1963 lOCO 0 on- 0 . 0 61
I N rlt1ern NJ 1 >? , n I .
2 ha ..sice OH 1990 88' 0 , 2' 127 I 9 6
2. S~2va Italv 19>1o; 0 I ?7n 64 I 5 2"-
22 Swift aM Two Medi ,ne ams. MT 1964 250 <1.5 0.7 , 2' , I 0 '6
2' Te!on 10 1 7 Dam throu n Wilford\ 00 <1. 0_75 I 7 - 25 I 2 -"5
2' Teton 10 1 7 IFlelbur fa Am<>rican Falls O1oon >0 2.25 . . I 0 '7
25 T;:.xas Hill "I 1978 2070 <1,5 "7 I 2 .. 25 I 2 '"
25 Veaa De Tera. Soa:n. 1959 500 0 0 1 ISO 89 I 7 387
.'\lO~= ;
, F1codillQrercaIUII1&SSi~l-tBro(epr9senl$'lowFOtCao'(ie.IlcOdwarersar':ikaoly10berelal;,.ely$/lallQwafld$l(lw).
Flooo; lon:etul"",ss inee. .. I reo,est!nrs "Hen Force. lie, IToOd walers are Iiht. ro ~ deeD ~ <r...rtn.

~

. contributions to Individual Risk should be computed for every failure scenario as one or other of the
products given in the fourth paragraph under "MAIN FEAl1JRES OF THE 1994 CRITERIA" above.
The total Individual Risk is then easily computed.
. it is now apparent from experience with risk assessment, that it will usually not be possible to identify a
single individual who is most at risk, but it will be possible to identify a group who are clearly most at
risk. Typically there will be a dwelling, or several dwellings or even a small community where all