ML19341A573
| ML19341A573 | |
| Person / Time | |
|---|---|
| Site: | La Crosse File:Dairyland Power Cooperative icon.png |
| Issue date: | 05/31/1980 |
| From: | Mcdonald J TEXAS TECH UNIV., LUBBOCK, TX |
| To: | |
| Shared Package | |
| ML19341A570 | List: |
| References | |
| CON-NRC-04-76-345, CON-NRC-4-76-345, TASK-02-02.A, TASK-2-2.A, TASK-RR NUDOCS 8101270141 | |
| Download: ML19341A573 (26) | |
Text
.-.
O l
i TORNADO AND STRAIGHT WIND HAZARD
~
PROBABILITY fer i
LA CROSSE NUCLEAR FOWER REACTOR SITE. WISCCNSIN by Jernes R.McDcacid. P.E.
4 ns:::u:e for 7sas::er Researca TEXAS 'ECF UN VERSFY l
_ubbock, Texas 79409 810127n g )
=_.
TORNADO AND STRAIGHT WIND HAZARD PRCBASILITY 4
for LA CROSSE NUCLEAR PCWER REACTCR SITE, WISCONSIN by James R. Mcdonald, P.E.
J 4
Prepared for i
U.S. Nuclear Regulatory Commission Site Safety Research 3 ranch I
Division of Reactor Safety Research l
May,1980 Institute.for Disaster Research
[
Texas Tech University Lubbock, Texas '
i l
l:
+
f c
9 4
FCREWARD Ha:ard probability assessment for tornadoes and other extreme winds at the la Crosse nuclear pcwer 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 werk is supportedunderNRCContractNRC-Cd-75-345.
Principal Investicator and Project Manager for the Institute for Disaster Research is James R.
McConald, P.E.
l' 1
3 A
I.
INTR 000CTICN The objective of this recort is to assess tornado and straight wind probability hazards at the La Crosse nuclear pcwer reactor site (See Fig.1). The ha:ard probability analyses are develeted 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 ha:ard probability mcdel is presented gra:hically in Figure 6.
Winaspeeds correspcnding to selected probability values are su:rari:ed in Table 8.
The basic data used in the calculations are presented in tnis report.
Derivation of the tornado hazard assessment methcdology, the rationale and assumptions are given in McConald (1980).
Use of the Type I extreme value distribution function for straight wind ha:ard assessment is well documented in Simiu and Scanlan (1978).
1
i
.=m h
p.-
./L A.
./
wwescs b
1
./
N.
N.3%.
\\.N
~
)
I LA CRCSSE 7
- c.m
(
D_.
MACISCN WEATHER STATICN s. T _.
LCCAL
.q REGICN
'\\'s i
l
,)
l u sos C-LCEAL I
(f /
REG CN
(
>i)
FIGURE 1.
LCCAL AND GLCBAL REGICNS.:CR LA CRCSSE 2
D*9D
- ]D 3'Th w e))
o Ju. S l iru a II.
TORNACO HAZARD PRCSABILITY ASSESSMENT A.
METHCCCLCGY The tornado hazard xdel developed ::y the Institute fer Disaster Research (ICR) accounts for gradatiens of da:nage across the tornado path wicth and along its length (McConald, 1980).
There are fcur basic steps invcived in the methodology:
(1) Cetermination of an area-intensity relaticnship in a giocal region surrcunding the site cf interest.
(2) Ceterminatien of an occurrence-intensity relationship in a local region surrounding the site.
(3) Calculation of the procabilities of a ;cin: within :ne local region experiencing winds:eeds in seme winespeed interval.
(1) Cetermination of the probability of winds:eeds in the lccal regicn exceeding the intervai values.
3.
CALCULATICNS 1.
Site La Crosse Nuclear Pcwer Reactor Site 2.
Cccedinates Latittce 43 33' 36" N Longitude 91 13' 42" W 3.
Area-Intensity Relaticnshic G1 coal Region U
Latitude 41 to 46 N 0
U Longitudc. 89 to 91 W Cata CAPcLE Tornado Data Tace UT1678 (Fujita, et al.,1979)
Period of Record 1971 - 1973 3
W b
p g
,1 0 N D gLf h
. $.N
=
See Figure 1 for definitien of the gicbal regicn. The region is selected c be as large as ;cssible and still give reascnaoly herx:ge-ncus c:nditiens for ::rnace formation. The relatively snce: :erice of recerc is used because the data are mere ccmcleta anc accurate thac that collected price :: 1971, especially witn regar: to tornado danage patn characteristics.
The area-intensity matrix is shcwn in Table 1.
It gives the numcer Of :Ornacces in each corres:cnding area-intensity classifica:icn.
Frem :nis infer-atien, -he mean damage path area ;er F-scale can be Obtained.
TABLE 1 AR5A-INTENS TY MATF.IX Numcer of Ternacces' Area Mean area
.1 F2
- 3
- 5 Isn mi1 Interval 70 0
17 5
6\\
0 0
0 0.315E-02 1
13 22 5
NO O
O 0.100E-01 N
2 14 24 9
2N 0
0
- 0. 315E-01 3
5 30 16 6
0 0.100E-C0 4
0 18 15 3
0 N, 0
0.315E-00 5
Oh 7
12 2
0 Q
0.100E 01 5
0 N0 4
5 2
0 0.315E 01 7
0 0N 1
2 1
0 0.lCOE 02 3
0 0
\\ Q 0
0 0
0.316E 02 g
0 0
ON O
0 0
0.1005 03 N
10 0
0 0
NO O
O 0.315E 03 Totals 49 106 59 20 3
1
'These tornadces outside the dashed lines are c:nsidered cu liers and have been eliminated frem the data set.
Mean Damage Fa n Area ?er F-Scale i
70
.=1 F2 F3
.=a F5 Mean Area, sq mi 0.0230
.1573 1.0570 1.9706 5.4300 ' 3.16C0 Median Wincs:eed, m:n 56 92.5 135 182 233.5 239.5 1
i Area-Intensity Function Linear regression analysis of the accve area-intensity data, based en a icng-leg pict, yielcs the fellcwing functicnal relationsnip:
Leg (Area) = 4.05 Lcq Y - 3.71 (1)
The ccefficient of determinatien is o
r = 0.99 Area-Intensity Relationship The expected mean area is cetained frem 5;uatien (1) abcVe.
Upcer and icwer bcund confidence limits are calculated at the 55 ;ercent level. These values are shcwn in Table 2.
Figure 2 shews a ple cf the area-intensi y relationshic.
TABLE 2 AREA-!ITEllSITY RELaTICtlSHI: WITH 95 PERCE.'1T C0:4FICE lCE LIMITS F0 FT F2 F3 F4 F5 Ex::ected Mean area, a, sq mi
.0228
.1738
.3023 2.5863 7.3605 17.5622 g
Lcwer Limit a,
9 sq mi
.0149
.1135
.5236 1.7502 4.7843 11.3874 Ucper limit a$,
sq mi
.035
.266 1.229 4.123 11.323 27.C85 Median F-scale Windsceed, mph 56 92.5 135
'182 233.5 239.5 4
Occurrence-Intensity Relaticnship Lccal Regicn U
U Latitude 42 to AS U
U Lengitude 90 to 93 Area = 31,206 sq mi.
See Figure I for definition of local region and-its relationship to the site.
5
- C0
/
/
f:
f
- o 9
l
~
'/ /I l /
s/
I
.o
/ (.
m I
/
/
F l
t.o 4
/I I
/ >/
/
r
/
/
/
/
/
./
ei
/
/
/
/
/
/
95% CONFIDENCE l
/'
/
INTERVAL J/
o o.cl 56 923 13 3.5 18 2 233.5 289.5 WINDSPEED MPH FIGURE 2.
AREA-INTE.'ISITY RELATIONSHIP FOR LA CROSSE 5
DAPPLE Tornado Cata Tape UT1673 (Fujita, et al.,1979)
Period of Record 1950 to 1968 The records used do not necessarily include every tornado that has occurred in the local region.
For one reason or another, scme tornadoes go unreported.
The population density within a 50-mile radius of the site is 41 persons per square mile (USNRC,1979).
This relatively low population 6nsity, coupled with a short sight distance because of the rolling ccuntryside, tends ::warc an increase in the number of unreported tar 2adoes. This trend is partially offset because the terrain is such that identifiable paths can ce seen shculd a tornado touch dcwn (damage to structures, trees, fences, or cower poles). The numcer of unreported tornadoes in the local region during the reporting period is likely to be less : nan 25 percent. - Neglecting One unrescrted tornadoes gives results that are sligntly uncenservative. The numcer of reported tornadoes in tne lccal region is shown in Table 3.
TABLE 3 NUMBER OF TCRNAC0ES IN THE LOCAL REGION F0 F1 F2 F3 F4 F5 Number of Tornadoes 70 125 82 20 3
0 Cumulative Number 300 230 105 23 3
0 Lower Scund F-Scale Windspeed, mph 40 73 113 158 207 261 Cccurrence-Intensity Func:f on The function used is obtained by performing a linear regression analysis using the F0 and F1 tornadoes, and another linear regression analysis using the F2 and F6 tornadoes.
7
.- ~,
b Linear regressien analysis of the data in Table 3 :n a semi-leg plct gives the fellcwing functional relationships:
y = (114.00)l0-0.0035x (x < 100 mch)
(2) 00.0165x y = (3086.30)l0 (x 1 100 mch) wnere y is the :umulative nu:rter of tornadces with winds:eeds greater than or e<;ual to x.
Occurrence-Intensity Relaticnship The expected nurrter of tornacces in the 29 year period is cbtained from the occurrence-intensity function (Equa:icn 2). Uccer and icwer beuna c:nfidence limits are also cbtainec at the 95 percen: level.
These values are then divided by the perice of record (29 years) to cbtain the nuccer of tornacces per year for each F-scale classifica-ti:n lj, which is the needed cc:urrence-intensity relaticnshic recuired for the ha:ard probacility assessment.
Table 4 lists the values used in the probability calcula:icn. Figure 3 snews a plct of the ec:urrence-intensity rela:icnship.
TABLE 4 CCCURRENCE-INTENSITY RELATIONSHIP WITH 95 PERCENT CONFIDENCE LIMITS F0 F1 F2
~3 FA
~5 Ex:ected numcer of tornadoes in inter-val, d 70.00 118.08 91.57 17.17 2.77 0.41 Lower limit 6 55.64 101.49 75.94 9.28 Uccer limit 6 84.36 134.66 107.20 25.06 6.02 1.67 Expected numcer of
- rnadoes per year.\\g 2.41 4.07 3.16 0.59 0.10 0.01 Lcwer limit A 1.92 3.50 2.62 0.32 g
U:per limit A 2.91 4.64 3.70 0.86 0.21 0.06 4
5.
Tora. ado Ha:ard Probability The ::rnade hazard probability calcula-icns are ;erformed by ccmouter, ai:ncugh they can easily be dene by hand. The ex:ected hazard :recaci'.i ies are Octainec by using ne ex;:ected area-intensf:y relatiensnia (15) and the expected cc:urrence-intensity relationshic (1j).
U::er and 1:wer limits Of na:ard Or cability are Obtained by using One u cer anc icwer limit Aj's and aj's rescectively. The c:m: uter print:uts f:r thesa :alculations are c:ntained in ?ccendix A.
8
- CCCO g
g C
i W
W C.
I r.O O
E W:CCO 3D l
J
'~~_'N
'~~~,,Th g
l 30.J l00 N
9
\\l mm
\\
gs N
%' Ns
<E s
s&
Z io 8
N l 's H O
\\
N E
N N
o N
u.y
\\
g O
t
.o w
@y 0
50 10 0 15 0 2C0 250 3C0 350 WINDSPEED MPH FIGURE 3.
CCCURRENCE-INTENSITY RELATIONSHIP FOR LA CROSSE l
l 9
10 Table 5 su=ari:es tne tornado ha:ard probabilities, and includes the 95 percent confidence limits. The tornado ha:ard pmbability mcdel is pictted in Figure A.
Final ha:ard probability results are sumari:ed in Section IV cf this recort.
TABLE 5 TORNACO HA7_ARD PROSABILIT!ES WITH 95 PERCENT CCNFIDENCE LI:1ITS Mean Hazard Tornade 'dinds:eeds, m:h Recurrence Probabili ty Expec:ec Lower Uccer Interval 7er Year Value Limit Limi 10,0C0 1.0 x 10~#
93 61 132 100,000 1.0 x 10~I 174 146 215 1,CC0,000 1.0 x 10-6 237 200 283 10,0C0,0C0 1.0 x 10~7 300 261 364 10
PROBABILITY OF EXCEEDING TilRESHOLD WINDSPEED IN ONE YEAR 4
M M
i 5,
5, 5,
5, 5,
N on on 84 (d
g il
/
/
m
/
/
/
ma as 1
/
/
ms z85
// _/
- c. x o
ah;
/
99 o
/
~'
m 9
/
nu e
s mg y
[_
CM g
/
i i5 i=
E o
-s~
o I
/
M
/ /
/
~
B 8
/
/
/
/
x Y
n,/
/
O
/
un (71 (M
8 e
/
3>
O
I 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 ha:ard probabilities.
The Type I extreme value function generally fits the data well.
In view of the studies by Simiu and Filliben (1975), the Type I distribution function is used in lieu of the Type II that was used previously (ANSI, 1972). A detailed descriptien of the methodology is given in Simiu and Scanlan (1973).
3.
CALCULATIONS Annual extreme fastest-mile windspeed data are not available at the power reactor site. The c*osest weather station with the needed data is Madison, Wisconsin, wnich is located 98 miles southeast of the site (See Figure 1). Terrain and meteorological conditions are such that the data should be representative of wind conditions as the site.
The data are taken from Simiu, Changery and Fillicen (1979) and covers the 31-year period 1947 to 1977. The sample mean is 55.55, and the standard deviation is 10.50.
Statistical tests indicate that the Type I extreme value distribution does not fit the Madison data as well as some other locations within the United States (the tail length parameter y is 45 rather than infinity as required for a true Type I distri bution). However, because the Type II distribution predicts windspeed values at low probability levels that exceed the physical cnaracteristics of the wind, the Type I distribution function is recem-mended for straignt wind ha:ard probability assessment at this site.
12-
The set of annual extreme fastest-mile windspeeds for Madiscn, Wisconsin are given in Table 5, along with the date and direction. The windspeeds have been adjusted to a standard anemcmeter heignt of 10 m.
TABLE 6 ANNUAL EXTREME FASTEST-MILE WINDSpEEDS AT MADISON, WISCONSIN Windspeed Year mch Direction Cate 1917 76 SW C4/05 1948 64 SW 12/C5 1949 68 SW 12/11 1950 30 SW C5/05 1951 75 SW l0/30 1952 48 SW 01/15 1953 66.
5 05/21 1954 73 SW 03/25 1955 51 S4 11/16 1956 49 W
05/12 1957 58 W
07/C8 1958 45 SW 11/18 1959 47 W
04/05 1960 52 SW 11/15 1961 44 5
11/02 1962 52 W
06/17 1963 56 N
06/07 1964 59 NW 07/27 1965 54 SW 06/27 1966 48 S
C4/19 1967 49 SW 04/14 1968 63 W
06/10 1969 43 SW 10/09 1970 65 NW 07/30:
1971 46 SW 02/27 1972 48 NW
.08/14 1973 53
.NW 07/09 1974 43 SW C4/12 l
1975
.50 SW 01/11 1976 50 N
07/30 1977 50
'a 07/16 L
The expected windspeeds for various mean recurrence intervals, along with 95 percent confidence limits, are given in Table 7.
The straight winc ha:ard probability medel is plotted in Figure 5.
13
TABLE 7 STRAIGHT '4INC HAZARD ?RC2 ABILITIES
'4ITH 95 PERCENT CONFICENCE LIMITS Mean Expected U:per Lcwer
.astes:-Mile Limit Limit Recurrence Ha:ard Interval Probability
'dindsceed. men men men 10 1.0 x 10'I 70 78 63 20 5.0 x 10-2 77 36 57 50 2.0 x 10-2 85 37 72 100 1.0 x 10-2 91 105 76 200 5.0 x 10~3 97 113 30 50 0 2.0 x 10~3 105 124 25 1,000 1.0 x 10~3 111 132 39 10,C00 1.0 x 10~4 131 159 102 1C0,000 1.0 x 10~3 1 51 186 115 1,000,000 1.0 x 6 1 71 213 128
~14
i PROBABILITY OF EXCEEDING TilRESHOLD WINDSPEED IN ONE YEAR M
as M
M M
5, 5,
5, 5,
5, u
u.
O a
a
/
- o
/
s us y
/
E '".
g
/
sis z
/
o s
--/
~
t
.o @
(A o
un ~'
~U f
u in g ITl 7
sg m
/
ei o o
ix
/
u,
~'
/
ns
/
E e
9 -
5 so 0$
/
flC
~
=
c;2 8
5a L!"
m 2
o O
P M
O O
+
D Y D
O' b
dj e JL S.1 m
IV. WINDSPEED HA72RD ?R09 ABILITY MCCEL Windspeed ha:ard pr:bability, which includes both ::rnadoes and straight winds, is the pre: ability Of a ;cint within scme definec geogra:hical region ex;eriencing wincs;eecs greatar than er equal to scme threshcid value in one year.
Tornado ha:ard probabilities are the same at any ;cint within the :efined lccal regien. The Type I extreme value distribution function catsined fr m da:a collected at Madison, Wisc:nsin is used for the straight winc :rocabili y ha:ard assessment at the la Crosse reactor site. Thus, in effect, Macison anc the reac:ce site are centained in a : mmon local region.
Tornacc winds;eeds are referenced to 30 f t above grcund level (acprox-ima:ely 10 m) and are the maximum heri: ental windspeeds.
According to Fujita (1971), F-scale windspeeds are fastest-cne-quarter mile winds.
However, because of the translational speed of a Ornado, winds acting on a structure may be of considerably shorter duration. Because tornado windsceeds are based on appearance of damage, they are : nsidered to be '
effective velccities, which include effects of gust, structure si:e and structure frecuency. For design pur;oses, the gust response fac Or for crnacc winds may be taken as unity.
The straign: winds are fastest-mile windspeecs which have a variable
- ime buration, ce:ending On the magnitude of the windspeeds. Values are ner.alized :c a 10 m anememeter heign:.
For design pur;cses, gust res;cnse fac:crs greater than unity are 3perepriate (See ANSI A58.1, 1972).
The tornado and straign: wind mcdels are ccmcined in Figure 5 to obtain
- ne #inal windspeed mcdei.
For design er evaluaticn purposes, one needs-o 'now ne type of s:crt tha c:ntrols the :riteria.
For windspeeds less nan I M m:n, tne straign: wind model governs.
For windspeeds greater than l
e 16
qw q
y g6 6
Ju.
3 144 mph, the tornado medel governs.
In the case of a tornado, the atics:heric pressure change and missiles must be taken into ac: unt in addition to the wind effects. Because of this, tne unien of the two events (tornado and straight winds) is not of particular interest.
Table _
8 sunnari:es the final windspeed ha:ard probacilities.
A
/
4 7
~
g--
i r 10'
\\
\\
STRAGHT
\\
'w:NDS
\\
\\
2
\\
\\
S i,:e a
\\ \\x B
\\
\\
eI
\\
s H
i::63
\\
\\
g ZE
\\
o w g
\\
W>
h, 4
iro gg
\\
wg
\\
bz
\\ \\ \\
l's 3
i, jg5 bo
\\
N!
\\
80 2
-6
\\
cr 1 x 10 Q3 95 PERCENT TcanaccES
\\
k CONFIDENCE
\\
g UMITS
\\
K I :10- 7 50 10 0 15 0 200 250 300 350 WINDSPEED MPH FIGURE 5.
WINDSPEED :4ZARD PRCBABILITY MCDEL WITH 95 PERCENT CONFIDENCE LIMITS
o g
TABLE 3
SUMMARY
OF WINDSPEED HADRD PRCBASILIT!ES 3CR LA CROSSE Mean Expected Recurrence Ha:ard Windspeed
- nterial Probability moh Tyoe of Storm 1.0 x 10-I 70 Strafgnt Wind 10 s
100 1.0 x 10 '
91 Straign: Wind 1.0 x 10-3 111 Straight Wind 1,CCC 10,C00 1.0 x 10 1 31 Straignt Wind 100,CCO 1.0 x 10 174 Torr. ado
-6 1,0CO,CCO 1.0 x 10 237 Tornado 1.0 x 10-7 300 Tornado 10,0C0,000 i
19
~
- 2 :WA _
REFERENCES 1.
ANSI, 1972:
"Suilding Ccde Requirements for Minimum Design Leads in Buildings and Other Structures," AS3.1, American Nati:nal Standards Instituta, Inc., New Ycrk, New York.
2.
Fujita, T. T.,1971:
"Procesed Characterization of Tornadoes and Hurricanes by Area and Intensity," SMRP No. 91, The University of Chicago, Chi:ago, Illinois.
3.
Fujita, i. i., iecson, J. J, and Abbey, R.
.,1979:. " Statistics of.
U. S. Tornadces Based en the CAPPLE Tcrnado Ta:e," lith Conference en Severe L: cal Storms, Xensas City, Missouri, Cet:ber 2-5, 1979,
- ublished by American Meteorological Society, 5cs
- :n, Massachusetts.
4 Mc0cnald, J. R.,1980:
"A Methcdol:gy for Tcrnado Ha:ard Assessment,"
Institute for Disas:ar Research, Texas Tech University, Lubbock, Texas.
5.
Simiu, E., Changery, M. J. and. illiben, J. J.,1979:
" Extreme Wind-s:eeds at 129 Stations in the Con igueus United States," N35 Building Science Service 118, Ma:icnal Bureau cf Standards, Washingt:n, D.C.
5.
Simiu, E. and Scanlan, R. H.,1973: Wind Effects en Structures, Jchn Wiley and Sens, 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, D. C.
3.
U. S. Nuclear Regulatory Ccmmission,1979: Cemcgra: hic Statistics Pertaining to Nuclear Pcwer Reac:cr Sites, NUREG-03a3, Office of Nuclear Reac cr Regulaticn, Washington, O. C.
l
[
I r
l l
l l
t l
20
- i..
",.r g
.,.n..* :=..,
4*.
s.
. ?Af*C*O.
,-..=
4
. w
.3 2*
.J.
= ~..
i.,.. *~
g.g l3 e.n a..; g..,
- 2 7
- g..4
%..*I7.
,T."*
m a.
g g w
w I*
3~*
- 4 2
ow*
T*
m
'd 6==
XZLZ%.
1 1 A.*
- s' 3
%. 1..K. *..
A. a* '. ?..\\
=.,...
g
. w
%)
- ~
- C O. O.P.O N N
- 7. 2. *.. :-
2.*.
. =
l
.x,
=. =. =...... -
. t v,, o.
c w
w..,
Z w w w..
idJ
~.
g m.
4 K.-g
. m
. w w w.
- x. '
. f N. ". %
w ww.
=. w w
w.
e b
N.
g
- ". m ; =
w.
.w..
.. w
..C,..
.. g,-.
.e y
7
=..
a
=
..g
- - 2
- E*
., s,,' L'
,,~ T.,..
1 ' *=*'. g
. i.=
'
- 7 N *4 0 w
...0 e..* i. a *.
- =
T
. w w.
w e.
. 1 a.
g y
=
g 3
- x. : 4' s.. p %.
~
. ;- 4
- *. *..T.
= ;;.
- . e. O. >. %.
i
- s '*.* = -
j
=
O?"
?.
w*
J 2
.?
g..
a..
.e 4
. a 21 -
e 3--
e'
- *2
.~.,
~*
=.
y.
.>e%*%
s.
3.
42
- . *. v..:O t.s. @.*
O
..", e * "J O O 9
7 L.
- * ** O. =.
I,* *.
=~
1..*=.
=. *
= * ' > * *. *
- g C - :* *.' ;;.: ;.
.=- -
.a 2
7
- 4 J...
?
- T
? -
2--
I *
- .? : 2. 1 i.
- Z... 3 *.
- a.. 1 e
... e
.e X-iY Jf 11%Y a7 s.
2 v...;
. e.. -..
4.
.. -. _ =
w
.'.3.
7 A. %' *...
g
. O. O...
O. 4.. 4 9..*
O C O C.*. C C * *. 0 *. ~.
- =*.
g.
N.
. : k. - *.
.. ? :- r..~
C.s.
. 2 w
.C
't 2
=
4
.s
=.
1
.;;; 9 4 : -
- e. *
+-
s..
..s.
- r w = =~
C.,.**'.<
- 5
- g
,2,.
.l
,=r.
s.
(
=
.e r.
g
=.-
=
=
w 3
+ =.
z a *:
..g
.~.,; J,
.. ~.
g....
-Q: 4..;
- z. *%
.**P. ; 1..:
- 6
= *
..d.
.1.
sY:.
2
.~
- f.
.St
=
OO&
O.*"**..
0,.
g
.c-
-.a.
a;.
M T,
.I f.
f Ap r
I
'. G t
lu 7.-
4 s.
l s
. S 0. : 33%
.~
1
- n. r..
.. =.
2
- 1., 7..." "
.1r.
-;. '
- K ". ".
..- =m s.
.3""
"2 O ;.,. e
.=2
- 2. ~
2...=
1 : : it.
3
- 3...
0
. +
3
. 3
- g
.,o O.
.. e s..s
_3.
., - e
.I.
- O C * ""
3JJ%. * *
. *. * =* =
0
- j % *=. O4 000>*.
- 7
?.*
3 3..
C. C. C. 'J. C. 0 OCOO3C C O 3 0 0 'J 09 s.
C 3..>.* #.
- 3*.
3 s,...
4*1
- e
..1*.
- 3 o.
C
. D.
O. *.
.h
. O..
..e O.
=
=.
.x.
r.o..,
m c
.,...... - -pT
. e
.,.s L.
g
.s...
w 6 e.
="
'g.
3 E
a.
.. =.
=
.g 1
= "
7 2
.E s.
e.
.s.
g
=.,.
n g
7 g
b
.