ML17258A725

Tornado & Straight Wind Hazard Probability for Ginna (Brookwood) Nuclear Power Reactor Site,Ny.

ML17258A725
Person / Time
Site:	Ginna
Issue date:	05/31/1980
From:	Macdonald J TEXAS TECH UNIV., LUBBOCK, TX
To:
Shared Package
ML17258A724	List: ML17258A725
References
CON-NRC-04-76-345, CON-NRC-4-76-345, TASK-02-02.A, TASK-2-2.A, TASK-RR NUDOCS 8101270156
Download: ML17258A725 (45)

Text

TORNADO AND STRAIGHT WIND HAZARD PROI3ABILITY for GINNA (l3ROOKWOOD) HUCLEAR POWER P EACTOP SlTE, NEW YORK by James P,. McDonald, P.E.

Institute or Disaster P,esearch

(

TEXAS TCCH UNIOR,SITY Lubboclc, Texas 79409

TORNADO AND STRAIGHT WIND HAZARD PROBABILITY for GINNA (BROOKWOOD) NUCLEAR POWER REACTOR SITE by James R. McDonald, P.E.

Prepared for U.S. Nuclear Reaulatory Commission Site Safety Research Branch Division of Reactor Safety Research May, 1980 Institute for Disaster Research Texas Tech University Lubbock, Texas

FOREMORD Hazard probability assessment for tornadoes and other extreme winds at the Ginna (Brookwood) nuclear power reactor site are presented herein at the request of Robert F. Abbey, Jr., Site Safety Research Branch, Division of Reactor Safety Research, U.S. Nuclear Regulatory Commission.

The work is supported under NRC Contract NRC-04-76-345. Principal Investigator and Project Manager for the Institute for Disaster Research is James R. McDonald; P.E.

I. INTRODUCTION The objective of this report is to assess tornado and straight wind probability hazards at the Ginna (Brookwood) nuclear power reactor site (See Figure 1). The hazard probability analyses are developed using storm records from the geographical region surrounding the site. Ninety-five percent confidence limits on the probabilities are presented to give an indication of the accuracy of the expected hazard probabilities.

The final hazard probability model is presented geographically in figure 6. Mindspeeds corresponding, to selected probability values are summarized in Table 8. The basic data used in the calculations are presented in this report. Derivation of the tornado hazard assessment methodology, the rationale and assumptions are given in McDonald (1980).

Use of the Type I extreme va ue dis~riEution function for straigh~ml hazard assessment is well documented in Simiu and Scanlan (1978).

1

0 O

C)

C4NAOA + NE'8 VCRK Gl NA ROCHESTER, N Y LOCAL REGQJ GLOBAL REGtCN FIGURE 1. LOCAL AND GLOBAL REGIONS FOR GINNA (BROOKMOOD)

C

I I. TORNADO HAZARD PROBABILITY ASSESSMENT A. tlETHODOLOGY The tornado hazard model developed by the Institute for Di.saster Research (IDR) accounts for gradations of damage across the tornado path

,width and along its length (McDonald, 1980). There are four basic steps involved in the methodoloo~:

Determination of an area-intensity relationship in a global region surrounding the site of interest.

(2) Determination of an occurrence-intensity relationship in a local region surrounding the site.

(3) Calculation of the probabilities of a point within the local region experiencing windspeeds in some windspeed interval.

CACOEPY (4) Determination of the probability of windspeeds in the local region exceeding the interval values.

1. Site Ginna (Brookwood) Nuclear Power Reactor Site

2. Coordinates Latitude 43 16'9" N Longitude 77 16'0" kl

3. Area-Intensit Relationshi Global Region Latitude 41 to 44 N Longitude 75 to 80 M Data DAPPLE Tornado Data Tape UT1678 (Fujita, et al., 1979)

Period of Record 1971 - 1978

See Figure 1 for definition of the global region. The region is selected to be as large as possible and still give reasonably homoge-nous conditions for tornado formation. The relatively short period of record is used because the data are more complete and accurate than that collected prior to 1971, especially with regard to tornado damage path characteristics. The area-intensi,ty matrix is shown in Table 1. It gives the number of tornadoes in each corresponding area-intensity classification. From this information, the mean damage path area per F-scale can be obtained.

TABLE 1 AREA-INTENSITY MATRIX Number of Tornadoes*

Area Area Interval FO Fl F2'3 F4 F5 Mean

~smi 2 7 OX 0 0 0 0.316E-02 5 20 5 y0 0 0 0.100E-01 13 3 1 0 0 0.316E-01 9 7 1 y0 0 0.100E-OO 3 4 1 +0 0 0.316E-OO 0% 1 0 0 0 +0 0.100E 01

~0 0 0 1 0 0.316E Ol ON 1 0 0 0 0.100E 02-0 ~0 0 0 0 0.316E 02 0 0 0~ 0 0.100E 03 10 0 0 0LO 0 0 0.316E 03 Totals ll 53 20 3 1 0 "Those. tornadoes outside the dashed lines are considered outliers and have been eliminated from the data set.

Mean Damage Path Area Per F-Scale FO Fl F2 F5 Mean Area, sq mi . 0290 . 0657 . 6054 .1492 3.160 Median Windspeed, mph 56 92.5 135 182 233.5 289.5

, t 4

Ar ea- Intensi ty Functi on Linear regression analysis of the above area-intensity data, based on a long-log plot, yields the following functional relationship:

Log (Area) 3.191 Log V - 7.330 The coefficient of determination is

.865 Area-Intensity Relationship The expected mean area is obtained from Equation (1) above.

Upper and lower bound confidence limits are calculated at the 95 percent level. These values are shown in Table 2.

Figure 2 shows a plot of the area-intensity relationship.

TASLE 2 AREA- INTENSITY RELATIONSHIP WITH 95 PERCENT CONFIDENCE LIMITS FO Fl 'F2 F3 F5 Expected Mean area, ai, sq mi .0178 .0881 .2944 .7638 1.6918 3.3597 Lower Limit ai sq mi .0059 .0298 .0988 .2515 .5431 1.0487 Upper limit a ~ ,

sq mi . 054 . 261 . 877 2. 320 5. 270 10. 763 Median F-scale Mindspeed, mph '56 92.5 1'35 182 233.5 289.5

4. Occurrence-Intensity Relationship Local Region Latitude 42 - 44 0 - o Longitude 76 79 Area = 29,975 6 365

= 14,610 An area of 6,365 sq mi is deducted from the local region because part of the region is in Lake Erie and Canada. There are, of course, no tornadoes recorded over water; the records do no contain any tornadoes reported in Canada. See Figure 1 for definition of local region and its relationship to the site.

100 IO CS CO I

hl 0 LO O.l 95o]o CONFIDENCE INTERVAL O.OI 56 92$ l33.5 I82 2335 289,5 WINDSPEED MPH fIGURE 2. AREA-INTENSITY RELATIONSHIP FOR GINNA (BROOKWOOD)

,( ~

1'

Oata OAPPLE Tornado Oata Tape UT1678 (Fujita, et al., 1979)

Period of Record 1950 to 1978 The records used to not necessarily include every tornado that has occurred in the local region. For one reason or another, some tornadoes go unreported. Because the population density of the local region is fairly high (greater than 250 persons per. sq mi, USNRC, 1979) and because the terrain is such that identifiable paths can be seen should a tornado touch down (damage to structures, trees, fences, or power lines), the number of unreported tornadoes in the region is likely to be less than ten percent. .The number of reported tornadoes in the local region is shown in Table 3.

TABLE 3 NUMBER OF TORNAOOES IN THE LOCAL REGION FO Fl F2 F3 F4 FS Number of Tornadoes 7 10 17 1 Cumulative Number ,35 .28 18 ..., 1 Lower Bound F-Scale Mindspeed, mph 40 73 113 158 207 261 Occurrence- Intensity Function The function used is obtained by per orming a linear regression

'x analysis using the FO and Fl tornadoes and another linear regression analysis using the F2 and F3 tornadoes. There have been no F4 and F5 tornadoes reported in the local region.

'x Linear regression analysis of the data in, Table 3 on a semi-log plot gives the following functional relationships:

v = (45.87)10 < 110 mph)

(2)

= (20.274)10 >

y 110 mph) where y is the cumulative number of tornadoes with windspeeds greater than or equal to x.

II Occurrence-Intensity Relationship The expected number of tornadoes in the 29 year period is obta ned the occurrence-intensity function (Equation 2). Uoper and lower 'rom bound confidence limits are also obtained at the 95 pere nt level.

These values ar then divided by the period of record (29 years) to obtain the number of tornadoes per year for each F-scale classifica-tion X;, which is the needed occurrence-int nsiiy relationship required for the hazard probability assessment. Table 4 lists the values used in the probability calculation. Figure 3 shows a plot of the occurrence-intensity relationship.

TABLE 4 OCCURRENCE-INTENSITY RELATIONSHIP

'PITH 95 PERCENT CONFIDENCE LIMITS F3 Expected -number of tornadoes in inter-val, n 7.00 9.98 '6.93 1.04 0.05 0.002 Lower limit n 2.36 4.74 11.13 Upper limit n 11 .64 15.21 22.72 3.02 0.49 0.085 Expected number of tornadoes per year X 1

0.24 0.34 .0.58 0.04 - 0.002 0.0001 Lower limit A,. 0. 08 0.16 0.38 Upoer limit X,. 0.40 0.52 0.78 0.1 0 0 02 F 0.003 Tornado Hazard Probabilit The tornado hazard probability calculations are performed by computer, although they can easily be done by hand. The expect d hazard probabilities are obtained by using the exoected area-intensity relationship (a ) and the expected occurrence-intensity relationship (X;). Upper and lower limits of hazard probability are obtained by using the upper and lower limit .'ii's and a;'s respectively. The computer orintouts for these calculations are contained in Appendix '.

Table 5 summarizes the tornado ha ard probabilities, and includes the 95 percent confidence limits. The tornado hazard probability model is plotted in Figure 4, Final hazard probability results are sumnarized in Section IV of this report.

TABLE 5 TORNA00 HAZARO PROSASILITIES WITH 95 PERCENT CONFIOENC"- LIMITS Mean Hazard Tornado Windsoeeds, mob Recurrence Probabi 1 i ty Expec:ed Lower Upper Interval Per Year Yalue Limit Limit 10,000 1.0 x 10 <40 <40 100,000 1.0 x 10 68 <40 132 1,000,000 1.0 x 10 135 90 188 10,000,000 1.0 x 10 172 135 247

l 1

0 10000

~ hJ 1000 g m K

>W 100 O

V) (0 p Q (9

K Q

p LLI

~

R5 o e 100 200 ~S~0~50 MPH 'INDSPEED FIGURE 3. OCCURRENCE-INTENSITY RELATIONSHIP FOR GINNA (BROOKHOOD)

~ I i

Cl xO Ix(O hl K

Ix(O C5 2e ~

Cl hl hl >-

UJ O ~

~Z

~o (x(05 U.

0~

0-

~ hl QJ I IO tQ

~

S CL V)

O z~

K (x(0 50 IOO 150 200 250 300-350 MPH 'NDSPEED FIGURE 4. TORNADO HAZARD PROBABILITY MODEL MITH 95 PERCENT CONFIDENCE LIMITS

I P

h

III. STRAIGHT WIND HAZARD ASSESSMENT A. METHODOLOGY A set of annual extreme fastest-mile windspeeds are used to fit a cumulative probability distribution function in order to obtain the straight wind hazard probabilities. The Type I extreme value function generally fits the data well. In view of the studies by Simiu and Fi'lliben (1975), the Type I distribution function is used in lieu of the Type II that was used previously (ANSI, 1972). A detailed description of the methodology is given in Simiu and Scanlan (1 978).

B. CALCULATIONS Annual extreme fastest-mile windspeed data are not available at the power pl.ant.site...The, closest weather .station with the needed data is Rochester, New York, which is located 17 miles east of the site (See Figure 1). Terrain and meteorological conditions are such that the data should be representative of wind conditions at the site.

I extreme value distri bution are taken from Simiu,

- The data and Type Changery-and-Fi.l-l-iben-(1.979)-. The-data--covers-a year period from 1941 to 1977. The mean windspeed is 53.4 mph, while the standard deviation is t~ph 'h~~f annual est'r~e fastest.-mi1.e wiodspeeds for Rochester, New York is given in Table 6. The windspeeds have been adjusted to a standard anemometer height of 10 m.

The expected windspeeds for various mean recurrence intervals, along with 95 percent confidence limits, are given in Table 7. The straight wind hazard probability model is plotted in Figure 5.

12

TABLE 6 ANNUAL EXTREME FASTEST-tlILE MINDSPEEDS AT ROCHESTER, NEM YORK Mindspeed Year ~mh Direction Date 1941 48 M 09/25 1942 49 W 01/02 1943 53 S 03/07 1944 45 M 02/06 1945 51 W 05/22 1946 59 W 02/14 1947 55 M 05/29 1948 51 SW 03/16 1949 55 SW 06/21 1950 65 W 01/14 1951 51 W 01/21 1952 48 W 01/15 1953 58 W 02/21 1954 57 W 03/03 1955 55 M 03/23 1956 60 W 02/25 1957 49 SM .06/29 1958 52 W 06/25 1959 60 SM 01/22 1960 51 M 04/18 1961 W 04/26 1962 49 SW 05/24 1963 54 09/12 1964 ---. 61 M SW 05/09 1965 54 W 10/31 1966 55 SW 01

/31'2/16 1967 62 SW 3968 SW 02/22 1969 48 SW 06/27 1970 46 SW 03/26 1971 53 SM 12/11 1972 55 SW 01/25 1973 61 SW 06/06 1974 49 SM 01/27 1975 61 SW 04/19 1976 50 SW 02/19 1977 50 W 03/30 13

~ I TABLE 7 STRAIGHT WINO HAZARO PROBABILITIES WITH 95 PERCENT CONFIOENCE LIMITS Mean Expected Upper Lower Recurrence Hazard Fastest-Mile Limit Limit Interval Probabilit Windsoeed m h ~mh ~mh 10 1.0 x 10 61 57 20 5.0 x 10 68 59 2.0 x 10 68 73 62 100 1.0 x 10 71 77 200 5.0 x 10 74 81 66

--- 2.0 500-1,000 '.0 x 10 x 10 78 81 90 69 71 10,000 1.0 x 10 --. 90 103 77 100,000 1.0 x 10 100 117 1,000,000 1.0 x 10 110 130 91

~,I I I x 10 C9 Za

~

Q UJ hJ >-

hl I x10 ~

O U.

0z 110' UJ

~ CL M

~ Cl 0~

I x10 50 100 150 200 250 300 350 WNDSPEED MPH FIGURE 5. STRAIGHT WIND 'HAZARO PROBABILITIES WITH 95 PERCENT CONFIOENCE LIMITS

E J IV. MINDSPEED HAZARD PROBABILITY blODEL Mindspeed hazard probability, which includes both tornadoes and straight winds, is the probability of a point. within some defined geographical region experiencing,windspeeds greater than or equal to some threshold value in one year. Tornado hazard probabilities are the same at any point within the defined local region. The Type I extreme value distribution function obtained from data collected at Rochester, New York is used for the straight wind probability hazard assessment at the Ginna (Brookwood) reactor site.

Thus, in effect, Rochester and the reactor site are contained in a common local region.

Tornado windspeeds are referenced to 30 ft above ground level (approx-imately 10 m) and are the maximum horizontal windspeeds. According to Fujita (1971), F-scale windspeeds are fastest-one-quarter mile winds.

However, because of the translational speed of a tornado, winds acting on a structure may be of considerably shorter duration. Because tornado windspeeds are based on appearance of damage, they are consider ed to be effective velocities, which include effects of gust, structure size and structure frequency. For design pruposes, the gust response factor for tornado winds may be taken as unity.

The straight winds are fastest-mile windspeeds which have a variable time duration, depending on the magnitude of the windspeeds. Values are normalized to a 10 m anemometer height. For design purposes, gust response factors greater than unity are appropriate (See ANSI A58.1, 1972).

The tornado and straight wind models are combined in Figure 6 to obtain the final windspeed model. For design or evaluation purposes, one needs to know the type of storm that controls the criteria. For windspeeds less than 109 mph, the straight wind model governs. For windspeeds greater than 16

109 mph, the tornado model governs. In the case of a tornado, the atmospheric pressure change and missiles must be taken into account in addition to the wind effects. Because of this, the union of the two events (tornado and straight winds) is not of particular interest. Table 8 summarizes the final windspeed hazard probabilities.

17

ar P h W ~ ~, ~ \'

I x IO Cl 2 0

O STRAIGH T WINDS I x IO CD Ze ~

UJ LU b

O~

xlO I-O I

~ hJ QJ CQ 0

~~

o~ CD x IO Q. 0 I 9 5 PERCENT CONFlDENCE .

LlMlTS I x IO IOO 200 250 300 350 IJINDSPEEO MPH FIGURE 6. TORNA00 ANO STRAIGHT WINO HAZARO PROBABILITY MOOEL FOR GINNA (BROOKWOOO)

POWER REACTOR SITE, NEW YORK 18

TABLE 8

SUMMARY

OF WINDSPEED HAZARD PROBABILITIES FOR GINNA (BROOKWOOD)

Mean Expected Recurrence Hazard Windspeed Interval Probabi 1 it mph e of Storm

,1.0 x 10 Straight Wind 100 1.0 x 10 71 Straight Mind 1.0 x 10 81 Straight Wind 1,000 1.0 x 10 90 Straight Wind-10,000 1.0 x 10 103 Straiaht Wind 100,000

-. 1.0 x 10 . 135 Tornado

~l= 000 000 1.0 x 10 172 Tornado 10,000,000

t,y I

REFERENCES

1. ANSI, 1972: "Building Code Requirements for Minimum Design Loads in Buildings and Other Structures," A58.1, American National Standards Institute, Inc., New York, New York.

2. Fujita, T. T., 1971: "Proposed Characterization of Tornadoes and Hurricanes by Area and Intensity," SMRP No. 91, The University of Chicago, Chicago, Illinois.

3. Fujita, T. T., Tecson, J. J, and Abbey, R. F., 1979: "Statistics of U. S. Tornadoes Based on the DAPPLE Tornado Taoe," 11th Conference on Sever e Local Storms, Kansas City, iMissouri, October 2-5, 1979, published by American Meteorological Society, Boston, Massachusetts.

4. McDonald, J. R., 1980: "A Methodology for Tornado Hazard Assessment,"

Institute for Disaster Research.,-Texas Tech University, Lubbock, Texas.

5. Simiu, E., Changery, H. J. and Filliben, J. J., 1979: "Extreme Mind-speeds at 129 Stations in the Contiguous United States," NBS Building Science Service 118, National Bureau of Standards, Washington, D.C.

6. Simiv, E. and Scanlan, R. H., 1 978: Mind Effects on Structures, John Wiley and Sons, New York, New York.

7. Simiu, E. and Filliben, J. J., 1975: "Statistical Analysis of Extreme Winds," Technical Note No. 868,. National Bureau of Standards, Washington,

0. C.

8. '.Pertaining S. Nuclear Regulatory Commission, to Nuclear Power Reactor 1979:

Sites, Demographic NUREG-0348, Statistics Office of Nuclear Reactor Regulation, Washington, 0. C.

20

~ Cg APPENDIX A. COMPUTER PRINTOUTS v ".I a'((al'I: I (II! f((I'l(a'(((i( f' vvl(

I'1)nilln'at( I(',4(I(t". I i ala. )(IIII:l(,9 I ~ I",- (.)( c <f

( j I Il)lav I( ((((i II )(II( I! I ll" ala((iFI. L(lldl:I'. I fall f I

~

~

I Iaa) I'I' '! v) "l I.i'ilI- )

0(i(i il (.(a}(l O.i(00  ! . ((((0

~

0, nna) 0. 0() ( ~

0 ~ (af) I J

~ Iva (i hvV! 0.0( 0 Ovnfin 0.000 0 v 0((0 Ov07i'.nhn 0- la'v4 I a".a.' n. z~! 0 0(in

~ 0.600 0. 00(() 0.311) 0('v.

(v 'i.' ( 0..>:In 0. 171 ') . 000 0 000 F O. I I . na000 a.k 0, I',>( 0,',.I 0, I'!0 0.007 0,000 0, f>(17 O.nnn

'v ~

"\ (

~

~

I (~ ..I ra! () ~ '> (0 nv 1(';0 0 ~ i)" > 0.1" 0- vl,(] 0.000

.;(' ill( I ~ I ':: I ~ .I ':((( I ~

II C(II( f )

,I, ~ isi I fi. i.)'):l I! 0(an O.nnn n. Onn (I. 090 i -(i "iI ~ Ca, ~ '!Ov) 0 non ai ~ (>00 0 00!)

~

0 . !' ~ ( ~, ~ ) I ni . 0(( v ~ . 0'.(0 0 f(a!(f

~ n. )nn

~

~i C(a .".ii ( ~ ~ ~ 1 ~ av ~ a"! . (.)On \( (( ' ~

~ ~ (~ ,~ (i ~ v) 0 ()fin (I . 0()fi ii. 0(pa') ~ )i)f'v (i ()a ~ I l ) s ~ (a( 0 v>i) 0

~ (i ~ ~ a<) ~ ) il, (i(ia)

~

' I I ~ i ~ ~ ~ (l,( ~ (i . (if)(l ,((I IIIII(YI'I (ill .(I (' I i I)

.'. ~ I ~ ( '(i ~ I (l ( Iva( I I  !.a I ', I,al( ~ ,!;. I l Oaf.l, il Ilf I i>>( i'l ~ >((l:!ii; a

.~ i 'Ca x.,l(l(! (i l" I ~ va ( ", . I, ) i3

( II' f'I ~ i I'I';il III II II(((I ( r ~(

( i ~~ ~ i I . Ia'.I .vi ~~ '. I'I:()I. (I>> I ",i' I, Ilfl(l,c l-l.v I ijI Ill(i I:( CI

~

( ~ ~

vv I I (I'll;I':rllllli((Ill(i(".I'FII( II! IIIII!(',"iil I,I

ge el s Ill

~ 5 sass s s ~ s I 1 lsl I !I I I I lso II, "' '

()- r I:.I)-I 3,)I ~ ~ ei s) .Sg>1!'1 1~ ~, ~ sos i) e e() ~ i>i>ss 5~ >(p )? ~

~ )s>'( ~,

a, 1 ~ .sc ~ (p, (i(s:1 () a {5('(S a ~ ~ ~ >p I, ~ ~  :)' e) . (>(s() (s ~ ip >5 (s ~ ) oi ~ (s i>>7

~ ) ~

>lao ~>  !). I?'. r>, s)s"pr ~ () . ('s)() I .7('>2 ('. It) I o>l (?. I"!!! () ~ (i(st s.. 5)(a') "s ll i7 i>.AI

~ p ~ ~1

~ e (5, sp ~ os s) l,y ~ ~ (a (IP)',> 'Is"

~ ~

~ I as I so I  ! ~ ~ '~ >(

. I ) "spl s I) 5 ~ i a ~ 1(g 1) ~ i){1(1 a) ~ !1, o ~1 (1 s ~ i()i') 0>p i ~

) ~ i(>l) i) . s)o)s) >1 ~

i iie i) . !)io!i

~, ~

I:>

~

le s;

(~

'>pa"

.() ' (

I) e)'e rp () r ( )s) sl,s I (1 C. e ~ I).>i's,s s.e()(l !) i(!  :),(>1) ) (t ~ ss I s .1>)() ~ 1. i) . i)a')() i) i'1 .' oi ~~ "-/ > ip ~ (p:s~ ~ (J. p ~ ~ I ) .:1si ( I e I 9 e'7:i el 1 5>ai ~ 5 (i a! (:5, ls i ~ o'p 7 ~ sp ~ I ~ "j~ L I .,") I ).'I" s)PP, i o e>'o ~ I I 1 1 ' s>':P. ! ~ 5~ I {)) ' 1 e'> I i;i >( o(e/ I~ ~ 1 1 ". ~ 5 .p) ~ >J, ... I .) si)'e >~ )

p

~ ! s, !) a'p I y s pl (is .! sp ..'>1 pl. ~ ()" !

Ij)s;srsj r( I (',<<.

(el '()ss r f!)I: I ()ssj I: ()S s)f I >5'!Ils')js ~ ') l(sjl I))e.)I. Ia..) (sl" Le'l'is!>I 'I!I:fi I j! ) i: i! '5>I 5 r I'<f!,['s ()rss;.'.;.'I;I')! I ls II ) I ill),'is",lsl I'li II'.I!!I'! '. I ~ s', I >I I ~ >I ><>l >3 ~ ~ > e: ~ s.sfs>> >> ~;

> ~ s ~> ~ ,' >>~ >j> ~ Ig)!'> < >> g

I ~ ! >I>g

!, '>..! I; I I'I I isi' ~ ~ g> ~ ~ ~ () .'.I I ~ ~ !s (I I)>>' ~ I ~, (i ".' ~ )s ~ rpe .>C)>) '4 ~~ r () ~ g) ~ ~ ='> i ()() I 'r! C. ~ ( 1>i"I~ > ~~ I').;i: I": ( > (I. <i: C>>,) ),) 0, e).<g> ~ ~ ) 's. (>I ! I). Cgg'.I ()(: f), Io" gg. s I '.>, <)'3>:. I I '. (h,) ~ ~ ~ ~ . > <I< <), C)f)C> I .()'IO (), <>()'I, se ~ ( .'F.(( 1> e (>.I ~ ~, ~ > j ~ i) ~ <)Cp>1 1 r" ' i) ~ i:O(. l ,) ~ e(i(') ~ ~ ~ Cp(! ~ ~ ~ ~ ~,> se'1,) s'>, f>e)s Ile>II C), (!(>f) ~ > I,I ') e "1>('i, is. <<>Ill> e'> i',> (>e>(> 1>,~I)() .~ g ~, ~ Il )~ i> ~ >i>( ~ r~ ~ '> ~:egg(e I ~ >~ ' ~ se ~ g> ~ ")

.. I. )I >).I g ~,. es'" I (),! ~: p<> < ~, ( I) ~> <!>>t!>!i'I l>>g> .)!!I.'is I)

I  ! pl. ~  !. ~ >>e< , i,i! C) I .>, ~ I(> >> '.!3 I I'; le< f>ltl > !'ll' lg<">et I!I 1 I<!(I I:I! I>l )e >'1 I I ~ ere g s '..I>',I fs(> I .!: I".'re>.I eh I' i I il ~ >=II>)l) >> I I> ls:>>'".I'I'I> It'. 'I!:L>i's!I ~ I 'I ' r" ~ I ~ '>g I'>> 1. II> ~ g\ >'P. >ee;l> e:>)>I ~ >! '> l>f ! ~ >'I'>> g I>!>I>eel!'I 'e ill IIII> ."<'s>l '1 " ~ I>/ I I >!I>t,l<<>>>l>e1<<>l >I'I>!1! > Ilf ! I'I >"'I < I

r. i ~