ML19322E564
| ML19322E564 | |
| Person / Time | |
|---|---|
| Site: | 07000008 |
| Issue date: | 09/30/1977 |
| From: | Fujita T CHICAGO, UNIV. OF, CHICAGO, IL |
| To: | |
| Shared Package | |
| ML19322E562 | List: |
| References | |
| REF-PROJ-M-3 NUDOCS 8003280389 | |
| Download: ML19322E564 (31) | |
Text
.,.
O,'
6 REVIEW OF SEVERE WEATHER METEOROLOGY at BATTELLE MEMORIAL INSTITUTE COLUMBUS, OHIO l
/
by T. Theodore Fujita Professor of Meteorology The University of Chicago l
September 30, 1977 Under Contract No. 31-109-38-373'.
Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439 8003280 b
6-1 Review of Severe Weather Meteorology at Battelle Memorial Institute Columbus. Ohio bY T. Theodore Fujita Professor of Meteorology he University of Chicago 1
INTRODUCTION ne Battelle Memorial Institute, Columbus, Ohio is located just to the west of the Big Darby Creek at 83*15'W and 39 58'N. The elevation of the site, esti-mated from the West Jefferson Quadrangle 7.5 minute topographic map, is 910 ft MSL.
As shown in Figure 1, the overall topography within the 10-mile range is flat, sicping up slightly toward the west, here are no mountains or hills of significant height which might induce orographic upslope or downslope currents.
For most design purposes, the environment of this site may be regarded as being flat.
According to Pautz (1969)(,)the SELS Log reported 18 occurrences of 50 kt 1
and greater windstorms and 20 tornadoes within the one-degree box of latitudes and longitudes which includesthe Battelle site. Itis the purpose of this review to determine the intensity of severe weather events which could affect this location, with return periods ranging between one and ten million years.
Both straight-line winds and tornadoes are regarded as prime phenomena, because no hurricane with significant winds has ever been reported in this area.
(2)
His site is located in Region II of WASH-1300 He esiculated tornado windspeed by five-degree squares for 10-# per year probability is 340 mph.
Pautz, Maurice E. (1969): Severe Local Storm Occurrences, 1955-1967 ESSA Tech. Memo WBTM FCST 12, 1
( ) WASH-1300 by Markee, E. H., Jr., J. G. Beckerley, and K. E. Sanders (1974):
Technical Basis far Interim Regional Tornado Criteria.
U.S. Atomic Energy Commission, Office of Regulation.
i 6-2
\\
\\ b?
P ),
N 3
at
\\
I $
so'o' I
N
(
Q
/t
/
f
',)
SITE sto'ust IO00' l
- / e, O
yo
/
- *' )
900' 10 D//e, \\
Figure 1.
3atte11e Memorial Institute and vicinity. The Institute is 1c.cated just to the west of the Big Darby Creek.
t The elevation of the site is 910 ft MSL. Height contours in i
this map were dr, urn at 100-ft intervals.
l e
e me - -
6-3 2
STRAIGHT-LINE WIhDS Straight-line winds occur more frequently than tornadoes, but their inter-l (I) pretation and evaluation are difficult. " Climatological Data" includes Columbus i
and Dayton, Ohio from which the fastest-mile windspeeds of the year (and month) are available.
Presented in Table 1 are the maximum fastest-mile windspeeds of each year between 1950 -76 at both Columbus and Dayton, Ohio. It should be noted that windspeeds are highly dependent upon the anemometer environment and exposure (including the height) of the instrument.
The mean speed at Columbus is 48.4 mph which is 4.1 mph lower than the 52.5 mph mean speed at Dayton, Ohio. The two stations are 71 miles apart and their elevations are 1115 ft (Columbus) and 1274 ft (Dayton).
Table 1.
Maximum fastest-mile windspeeds in mph by year at Columbus and Dayton, Ohio during the 27-year period, 1950-76. From Climatological Data for these years.
Years 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 columbus 57 44 61 43 49 63 57 48 42 56 Dayton 78 56 62 51 56 61 56 49 56 63 Years 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 columbus 42 40 45 44 54 56 47 54 42 36 Dayton 56 59 39 53 43 41 45 60 43 40 Years 1970 1971 1972 1973 1974 1975 1976 Mean speeds Columbus 56 47 41 42 52 37 51 mph 48.4 mph Dayton 61 56 52 47 47 42 45 mph 52 5 mph
('} Climatological Data. Publication of NOAA, published monthly with an Annual Summary. May be obtained from Environmental Data Service, Nadonal Climatic Center, Federal Building, Asheville, North Carolina 28801
6-4 In order to combine the fastest-mile speeds from these two stations and obtain windspeed probabilities, speeds were normalized by multiplying the following ratios, Mean of Columbus and Dayton 50.45' 1.04 (for Columbus)
=
Mean at Columbus 48.4 and Mean of Columbus and Dayton 50.45 0.96 (for Dayton)
=
Mean at Dayton s2. 5 Windspeeds computed by multiplying each of these ratios by the fastest-mile speeds from each station are called the " normalized fastest-mile windspeeds ".
They are then used in obtaining the statistical results presented in this review.
In central Ohio, the fastest-mile winds of the year.in the cold seasons (late autumn, winter and early spring) are the result of well-developed continental cyclones. The fastest-mile winds in the warm seasons (late spring, summer, and early autumn) are often caused by so-called " straight-line winds " induced by severe thunderstorms.
Table 2 was prepared to show the seasonal variation of the fastest-mile winds of the year by month. April through September are regarded as warm seasons and October through March, as cold seasons.
Table 2.
Normalized fastest-mile windspeeds of the year obtained by making the mean-speed correction. Speeds are tabulated by nonth. 27-year period at Columbus and Dayton, Ohio.
Months Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar 58 58 74 53 43 53 63 48 58 59 65 mph 54 56 50 51 43 59 45 53 59 60 53 50 48 49 58 43 46 57 58 53 39 43 4
47 43 56 53 45 38 42 37 56 49 41 38 53 45 41 43 45 40 37 41 Fa imun 58 58 74 53 43 53 63 48 58 59 65 mph
6-5 Table 3 Probabilities of normalized fastest-mile speeds of the year during warm seasons (April-September), cold sea-sonc (october-March), and all year. Frequencies (Freq.),
cumulative freqancies (Cum.), and probabilities per year (Prob.) are tabulated. 27-year period, 1950-76, at Columbus and Dayton, Ohio.
Normalized Warm Seasons Cold Seasons All Seasons windspeeds Freq. Cum. Prob.
Freq. Cum. Prob.
Freq. Cum. Prob.
37 mph o
25 0.46 yr-'
2 29 0 54 yr-'
2 54 1.00 yr' 38 2
25 0.46 0
27 0 50 2
52 0 96 39 23 0.41 0
27 0 50 1
50 0 93
^
i 40 1
22 0.41 0
27 0 50 1
49 0 91 41 2
21 0 39 1
27 0 50 3
48 0.89 42 1
19 o.35 0
26 0.48 1
45 0.83 43 3
18 0 33 3
26 0.48 6
4 0.81 i
15 0.28 o
23 0.43 1
38
,0 70 45 1
14 0.26 3
23 0.43 4
37 o.69 46 0
13 0.24 1
20 0 37 1
33 0.61 47 0
13 0.24 i
19 o.35 1
32 0 59 48 i
13 0.24 1
18 0 33 2
31 0 57 49 1
12 0.22 1
17 0 31 2
29 0 54 50 2
11 0.20 o
16 0 30 2
27 0 50 51 1
9 0.17 o
16 0 30 1
25 c.46 52 0
8 0.15 0
16 0 30 0
24 0.4 53 3
8 0.15 4
16 0 30 7
24 0.44 54 1
5 0.093 o
12 0.22 1
17 0 31 55 0
4 0.074 o
12 0.22 o
16 0 30 56 1
4 0.074-2 12 0.22 3
16 0 30 57 o
3 0.056 i
io 0.19 1
13 0.24 58 2
3 0.056 3
9 0.17 5
12 0.22 59 o
1 0.019 3
6 0.11 3
7 0.13 60 o
1 0.019 1
3 0.056 1
4 0.074 63 0
1 0.019 1
2 0.037 1
3 0.056 65 0
1.
0.019 1
1 0.019 1
2 0.037 74 1
1 0.019 0
o 0.000 i
i 0.019 l
l
6-6 From a meteorological point of view, one may assume the maximum wind-speeds of a continental cyclone to be 70 to 80 mph, which correspond to the low F 1 scale or low hurricane winds. The maximum possible windspeeds of straight-line winds have not been known very well, but they could reach the 90 to 100 mph range or even higher. The Northern Wisconsin downbursts of the 4th of July,1977 were estimated to be up to low F2 or 110 to 120 mph fastest-1/4 mile speed.
Due to anticipated differences in the nature of winds in warm and cold seasons, their probabilities were first computed separately. 'Ihey were then com-bined into all-season probabilities (see Table 3).
Figure 2 reveals the trend of probabilities given in Table 3. Apparently, the windspeeds in cold seasons tend to saturate at or with 10-* to 10-* year-'
probabilities, while those in warm seasons keep increasing with decreasing proba-bility. These results indicate that 1.0 to 0,1 per year probability is dominated by the winds in cold seasons (continental cyclone origin). A 0.01 per year probability or lower is, however, dominated by the winds in warm seasons, which are induced by severe convective storms.
Table l+.
Frequencies of fastest-nile wind directions by seasons. During warn seasons west-northwesterly wirds of connective origin dominate frequencies; while in cold sea-sons west-southwesterly winds of continental cyclone origin dominate frequencies.
Wind directions E
SE S
SW W
NW N
NE UrJmown Total Warm seasons 0
1 1
5 7
9 1
0 1
25 Cold seasons 0
0 2
12 11 3
1 0
0 29 All seasons 0
1 3
17 18 12 2
0 1
5'+
Directions of the fastest-mile winds in Table 4 also reveal a major difference between warm-and cold-season winds. During the cold seasons, directions are predominatnly from southwest to west, while those in warm seasons are from west to northvest (see Figure 3).
i
6-7
^ 40 50 60 70 80 90 moh 10 m
n q
dD.
yM R.,
s.x lo '
ALL SEASONS
.N 0
s
.s.
N-s
\\
\\
\\
16'
\\
6N N
+3 8
d[#
G 9,,*
wm 10
EOz*
STRAIGHT-LINE WINDS Figure 2.
Probabilities of the speeds of the fastest-mile winds of the year at Columbus and Dayton, Ohio. Based on a 27-year record, 1950-76. Probabilities were estimated by separating the speeds in warn and cold seasons, because the nature of winds is apparently different in these seasons.
i l
From Table 3 I
I i
6-8 WARM SEASONS 8
( APR - SEP)
Running Mean
[
6 4
2 m
O i
k1 i
E SE S
SW W
NW N
NE' E N
cold SEASONS 10 (OCT - M AR) 8 4
d>
6 Running Mean 4
[
2 O
i k1 i
E SE S
SW V/
NW N
NE ' E Figuro 3 Directions of fastest-nile winds of the year at Columbus and Dayton, Ohio during warm and cold seasons. It should be noted that the fastest-mile winds in warm seasons are predominantly from west to northwest while those in cold seasons, from southwest to west. From Table 4
i j
6-9 1
%e probabilities of the occurrence of maximum windspeeds should be defined differently from those of tornadoes, because windspeeds at each station are measured in time domain at a fixed point. Their spatial variations around the anemometei are usually unknown.
Jor tornadoes, the National Weather Service lists all storms based on the best possible information. Tornadoes are listed separately, even if they occur on the same day or even a few minutes later.
He maximum fastest-mile speeds are listed in " Climatological Data " by month and by year. Here is no mention as to how often the maximum speed occurred within one month or one year. He periods of straight-line winds, especially the ones caused by continental cyclones, are long, lasting hours or even days. There will be numerous maxima during such a long period. We should, therefore, define the following terms:
Fastest-mile day
-- the day on which the speed occurred Fastest-mile month -- the month in which the speed occurred Fastest-mile year
-- the year in which the speed occurred Rese are similar to the term Tornado day -- the day in which one or more tornadoes occurred.
In these cases, the number of occurrences within the stated period is not important.
Probability of the fastest-mile year can be computed by Number of years in which specific speed or larger speed occurred p,
Total number of years used in statistics where P.
denotes the occurrence probability per year.
He probabilities in Table 3 were computed by combining the observation years at Columbus (27 years) and Dayton (27 years), thus resulting in statistics of 54 observation years.
Windspeeds of peak gusts are higher than those of fastest mile winds, because the duration of a peak gust is considerably shorter than the period of the fastest-mile wind. In this review, the former is regarded as being 25% longer than the htter. Namely, Peak Gust = 1. 25 Fastest-mile Speed.
6 -10 3.
TORNADO FREQUENCIES FROM NSSFC TAPE he NSSFC Tornado Tape lists 567 tornadoes within 144 miles from Battelle Memorial Institute during the 26-year period, 1950-75 He cumulative frequencies were computed to determine the trend of their increase as a function of the range from the site. As shown in Figure 4, the best-fit parabolic curve is N = 0. 0295 R' where N is the number of tornadoes within range, R, in miles. A gradual decrease in the number of tornadoes at ranges in excess of 100 miles is in part an effect of Lake Erie. Another reason being a decrease in tornado frequencies in the south-eastern sector.
A breakdown of tornado frequencies by year in Table 5 shows relatively low frequencies in the early 1950s when the tornado reporting system by the U.S.
Weather Bureau was in the process of being improved. The maximum frequencies of 74 occurred in 1973 followed by 59 in 1974, which was the year of the April 3-4 super-outbreak.
Table 5.
Frequencies of tornadoes within 144 miles from Battelle Memorial Institute by year. Based on the NSSFC tape, 1950-75 Years 1930 1951 1952 1953 1954 1955 1956 1957 1958 1959 Frequencies 6
5 2
12 17 12 20 14 19 11 Years 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 Frequencies 11 33 5
28 17 59 4
17 33 27 Years 1970 1971 1972 1973 1974 1975 Mean Frequencies 31 21 14 74 59 16 21.8 i
i I
6-1I i
TORNADOES 600 l
l I
I I
I l
+
o o
o 500
- o
)
o N = 0.0295 R*
o o*
i o
i o
o 400 -
a*
O o
o 300 200 O
100 00*
o e*
o*
- 0-1 I
I I
I I
O 20 40 60 80 10 0 120 140 miles i
Figure 4 Cumulative number of tornadoes as a function of the distance from Battelle Memorial Institute. Based on 567 tornadoes in NSSFC tape, 1950-75
6-l2 Tornado frequencies by month reveal that April with 163 tornadoes is the month of the highest frequency. %e lowest frequency month is January, with only 2 tornadoes out of 567 or 0.35% (see Table 6).
Table 6.
Frequencies of tornadoes within 1% miles from Battelle Memorial Institute by month. Based on the NSSFC tape, 1950-75 Months Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total-Frequencies 2
16 31 163 97 95 64 43 23 6
19 8
567 4.
STATISTICS FROM DAPPLE TAPE 01 The DAPPLE Tornado Tape introduced by Fujita (1977) can be used in computing the path lengths of F-category tornadoes within 15 x 15 min sub-boxes of longitudes and latitudes anywhere inside the contiguous United States. An attempt was made, in this review, to make use of the DAPPLE Tape as a new potential tool in assessing the tornado risk at the Battelle site.
U. S. Tornado Map by Fujita and Pearson (1976) ' reveals a significant decrease of tornado frequencies toward the southeast from the Battelle site. He frequency gradient is, more or less, perpendicular to the line connecting Cinctnnati, Gi?io with Cleveland, Ohio.
II)Fujita, T. T. (1977): Tornado Structure for Engineering Applications. To be published as SMRP Research Paper No.153.
I )Fujita, T. T. and A. D. Pearson (1970):
I U.S. Tornadoes, 1930-74. He i
University of Chicago, i
l l
6 -13 A realistic assessment of tornado frequencies can, therefore, be achieved by generating a coordinate system with their axes parallel and perpendicular to the maximum gradient of the frequencies.
The X-Y coordinates in Figure 5 were constructed based on 6e above con-siderations. Ren, the banded areas, A through L, were drawn. Each area is 200 miles long and 20 miles wide, distributed equally on both sides of X a. s.
t
'Ibe total path length within each 15 x 15 min sub-boxes of longitudes and latitudes was computed over the entire area of the bands, A through L (see Figure 6).
Path lengths in miles reveal the largest values of 50 miles (in Indiana) followed by 48 miles (in Michigan), and 44 and 43 miles (in Indiana). Here are a number of sub-boxes with zero path length. As expected, we see more zero numbers to the southeast sector of the Battelle site.
In an attempt to compute the total path lengths in bands A through L, the area of each band was modific.1in such a manner that the band bounda'ry becomes the composite boundaries of 15 x 15 min sub-boxes. He bands reconstructed from the.se sub-boxes are called the modified bands, A through L (see Figure 7).
Naturally the area of a modified band is different from that of the original i
band with approximately 200 x 20 = 4,000 square miles. For most bands, areas were e::panded by adding a sub-box at each end of the original band.
He path length in miles per unit area or the " path-length density " in each modified band was computed from at en @
nm ed band Path-length density =
Area of modified band the unit of which is miles /sq. mile or mile. He unit can be expressed also in
{
any other units with identical dimensions, such as 10-' mile ~' = miles /10,000 sq. miles etc.
Path-length densities in Table 7 are given in 10 mile-' unit. The last column of this table reveals that the path-length density of total tornadoes
( F0 through F5) decreases from 850 to 38 between modified bands A and L.
6 -14 1
I i
i i
86 85 84 83 2
81
~42 t
I
__T_
i y
n O
l l
c i
4 i
- 41 1
+
l 4
l41-
/
I O
I I
D
.- d I
I s 1 j'
N
'I i
e
/
2 in 8
8 Battelle Memorial Institute 4
ji N/
4
'40-n f
I Esp y
l
)
s
/6 I
i
,s
&g
,s
. ' ~ _,
f
-39 i
39-i c',s.)
r O
/
i
's
,__s,_,-
'I
s ~/
's
_I a
i' s
r., '
Q a
s y
e'38 s
38-I
\\
82 85 84 83 81 I
I I
I I
l Figure 5 Bands A through L placed around the Battelle site.
Each band 200-mile long and 20-mile wide, is oriented approxi-mately in a SS*/-?GE direction.
l 6 -15 86 85 84 83 82 81 12
'42 48 18 5
3 I
'i l
4 3
2 10 3
i i4 4 0
0 1
12 14 13 i
i i
i Ol 7 18 4 5
17 Ii 17 15 8 4
13 7
ijl i
4 14 13l14 18 15 21 16 9 4
3 2
6 3
8 I
O 8
i 5
I i
29 35 3
18 0
5 I
i 3
17 5
7 5
4 3
0 1
6 0
l (J
e ii 4 22 1,I l 7
14 7
0 3
16 II 2
3 0
1 12 19 2 I
Ol0 fI 14 22 38 44 25 5
2 6
5 2
8 25 9 10 O
I 8
8 0
I /O l 2 i
i 33 50 4'
7 8,8 4 1
4 7
19 10 9
28 20 5
I 6
6 0
1 0,' O j 43 27 21 21 10 6
3 12 19 0 2
9 30 8 1
5 2
O I
O,.0 8
/
6 17 18 7
28 4 26 14 8
5 5
1 4
i O
O O
2(~0' O s
.r' B
2 4
9*sl8 25 27 10 2J O I
2 O
O O
O!l O
5 t
s.
39
': 3 i \\l 13 21 23 17 II 4
4 2
2' I
( -~,..
- K,,_
'~~
7 36 18 'l6 Gi -O 2
0,O O
7,,,'s l
~,
l O
O O
8 l. 1 1
9
, ~ _ -
a O
O.
O I
N 38 POth Lengths Of Oil TOrnOdoes Figure 6.
Total path length within each 15 x 15 min longitude-latitude sub-box. From the DA??LE tape including 1950-75 tor-nadoes. For DAPPLE tape, refer to Fujita (1977): Tornado Struc-ture for Engineering Applications.
6-16 ll i
________ g e
l j
t I
g I
I I
1 l
i I
i e
l l-i 1
i i
A I
B i
, -t C
I D
E
/
I
~~~
Y F
G
,i t
H r
s j
d I
I K
k d
,/ ~.,.,s,L t
s
'd s
,.,s',-
p
~
i
__,-~,_
'-~,
e
,,~'
r r~ J 1
0 a
i o
/
i t
4
.a.
Figure 7 Modified bands A through L consisting of 15 x 15 l
min longitude-latitude sub-boxes.
i
6-17 Table 7.
The path-length density is defined as the total path length divided by the area which includes the paths.
The area of r:odified band varies fron band to band. See Figure 7 Modified Path-len6th in each band Band area Path-length density in 10" mi" band (F(hF1) (F2+F3) (F4+F5)
(sq. ni)
(F0+F1) (F2+F3) (F4+F5) (Total)
A 100 214 58 4333 231 494 134 859 3
53 158 34 4114 129 384 83 596 c
52 119 44 5030 103 237 87 427 D
24 55 47 3904 61 141 120 322 E
26 65 82 3665 71 177 224 472 F
24 111 70 4624 52 240 151 443 G
80 112 28 4620 173 242 61 476 H
85 46 16 4406 193 104 36 333 I
12 32 17 4640 26 69 37 132 J
22 55 0
4186 53 131 0
184 K
5 11 0
4668 11 24 0
35 L
8 9
0 4440 18 20 0
38 i
f This dramatic decrease is shown in graphical form in Figure 8.
'Ibe data points given with painted circles show an appreciable scatter, probably due to the short statistical period, time variations of tornado activities, and other unknown causes. Nevertheless, a smooth line sloping down from A to L represents the general trend of the path-length density applicable to the Battelle site and vicinity.
A breakdown of the path-length density into three-category tornadoes results naturally in large scatters (see Table 7). Nonetheless, we are able to draw a smooth curve for each category tornado (see Figure 9).
Table 8 shows the path-length densities at the Battelle site obtained by this smoothing method. Results indicate that strong ( F 2 + F 3 ) tornadoes were the largest in density, 0.02 mi" (mi/sq. mi), followed by both weak and violent category tornadoes, each with 0.01 mi" density. These path-length densities, thus obtained, can immediately be used in computing tornado probabilities using the DAPPLE Method.
6 -18 Path -length Density in mi/lo,ooo sq mi O
800 SITE I
l I
600 g
I
.e 8
410 mi/10,000 sq mi 400 I
e l
e 200 e
e o
A B
C D
E F
G H
I J
K L
Figure 8.
Path-length density (miles per 10,000 sq. mile unit) defined as the total path length within a codified band divided by the band area. Fath-length density in this figure includes all tornadoes, FO through F5 From Table 7
!i
6 -19 FO + FI 300 StTE
~
200 l
g
'O mi/i0,000sq mi 100 e
e!
O A
B C
D E
F G
H I
J K
L F2 + F3 400 SITE l
300 l
i.
200 200 ml/10,000sq mi i
100 I
I 9
O
- ~
A B
C D
E F
G H
I J
K L
3 F4 + F5 SITE 200
.i g
100 g
100 mi/lo,000 sq mi 1
0 A
B C
D E
F G
H I
J K
C Figure 9 Path-length densities (miles per 10,000 sq. mile unit) of three-category tornadoes averaged over the area of j
each band, A through L.
From Table 7 l
6 -20 Table 8.
Path-length density (ni/10,0.00 sq mi or 10 '
mile-' ) of three-categor-/ tornadoes, Weak (F0+F1), Strong (F2+F3), and Violent (Fb?3) applicable to the Battelle Site.
Weak Strong Violent All tornadoes 110 200 100 410 5 WINDSPEED PROBABILITIES OF TORNADOES The DAPPLE (Damage Area Per Path Length) METHOD developed by Abbey and Fujita (1975) is capable of computing tornado probabilities as a function of the F-scale damage categories, which can be converted into windspeeds (see Table 9).
Using the DAPPLE METHOD, the area of specific windspeed can be computed by the product, Windspeed Area = Path Length X DAPPLE,
where path length denotes that of specific F-scale tornadoes and the DAPPLE values vary with F scale.
Table 9 Ranges of F-scale windspeeds arxi their weighted.
mean values. (Refer to Abbey, Robert (1977): Risk proba-bilities associated with tornado windspeeds. Proc. of Symp.
on Tornadoes, Assessment of knowledge and implications for
)
man.)
F Scale F0 F1 F2 F3 F4 F5 Range of Windspeed 40-72 73-112 113-157 158-206 207-260 261-318 mph Weighted mean speed 59 92 131 177 227 276
- 0) Abbey, R. F. and T. T. Fujita (1975):Use of tornado path lengths and gradation of damage to assess tornado intensity probabilities. Preprint of 9th Conf. on Severe Local Storms, 286-293.
i 6-21 Since F-scale assessments assume an accuracy of one scale, tornadoes are classified into three categories: WEAK ( F 0 + F 1 ), STRONG (F 2 + F 3),
and VIOLENT (F 4 + F5). DAPPLE values were then obtained based on the 147 tornadoes of April 3-4,1974 Super-outbreak tornadoes (see Table 10).
At the present time, DAPPLE values are being updated by adding new survey data. Since we do not expect to survey a large number of F5 tornadoes in the near future, we will have to use these DAPPLE values for our immediate solutions.
Table 10. DAPPLE, in miles, as a function of tornado windspeed. Unit in sq. mile per path mile. DAPPLE values are given for 3 categories of tornadoes. From Abbey, R. F.
and T. T. Fujita (1975): Use of tornado path lengths and gradations of damage to assess tornado intenalty probabil-ities. Preprint of 9th conf. on severe Local storms. 286-293 Windspeeds Violent (F4k5) strong (F2&3)
Weak (F0&i) 50 mph 0 51 mile 0.43 mile 0.074 mile 100 0.14 0.062 0.0028 150 0.036 0.c098 0.000052 200 0.0081 0.0012 0.000000 250 0.0016 0.000087 0.000000 300 0.00023 0.000000 0.000000 m
350 0.000016 0.000000 0.000000 Now the path-length density in Table 8 and the DAPPLE values in Table 10 can be combined into a product which may be called the " area density ",
area density =
unit = s where A denotes the araa in which the area of a specific windspeed is located.
For example, a 0.03 square-mile area located inside a 10 square-mile area results in 0.003 area density Naturally, the area density decreases with windspeed.
The areas of three-category tornadoes in Table 11 are given as square miles of wfudspeed area per 10,000 square miles or 10. Values were computed from Tables 8 and 10 i
,I
6-22 Table 11. Areas of three-category tornadoes during a 26-year period, 1950-75 applicable to the Battelle Site.
Values are in sq mi/10,000 sq mi or 10'*.
Windspeeds Violent Strong Weak All tornadoes 50 mph 51 86 8.14 145 1 100 14 12.4 0 308 26.7 150 3.6 1 96 0.0057 5.6 200 0.81 0.24 0.0000 1.05 250 0.16 0.0174 0.0000 0.18 300 0.023 0.0000 0.0000 0.023 350 0.0016 0.0000 0.0000 0.0016 The windspeed probabilities are now computed from en ens W Probability = Years of Statistics unit Year
'Ihe total number of years used in these statistics is 26 years,1950-75 The probability of any windspeed can be obtained by simply dividing the area density of specific windspeed by 26 years.
Shown in Table 12 are probabilities per year of various windspeeds at 50 mph intervals. Values indicate that 10 year probability occurs when wind-speeds are between 250 and 300 mph, l
Table 12. Probabilities (year-' ) of various wirdspeeds by all tornadoes at the Battelle Site.
Windspeeds Probabilities (yr)
50 mph 5 58 x 10
- 100 1.03 x 10
150 2.15 x 10
200 4.04 x 10
250 6.92 x 10-#
300 8.85 x 10
350 6.15 x 10"
6-23 6
SUMMARY
AND CONCLUSIONS -
Results of the foregoing computations of windspeed probabilities are summarized in Figure 10 which includes three curves applicable to the Battelle site. The fourth curve represents the peak-gust probabilities from the SELS Log. The three curves are (A)
Probability of fastest-mile year (B)
Probability of fastest-mile year (C)
Tornado probability These curves reveal that the speeds of straight-line winds are higher than tornado winds when the probability is greater than about 10-* per year.
Table 13 gives the maximum windspeeds corresponding to curves ( A),
(B ), and (C ).
Values for the retum periods of one to 10-million years are tabu-lated. It is seen that tornadoes dominate the windspeeds when the probability decreases below one in 100,000 years.
Three-category probabilities are used in this site analysis:
High probability
-- 10-* per year Low probability
-- 10-' per year Remote probability -- 10 per year It should be noted that 10 ' per year is used in determining the design-basis tornadoes
~
for nuclear power plants in WASH-1300 Table 13 Maximus windspeeds expected at the Battelle Memorial Institute Site as a function of the probability per year. (A)--fastest-mile speeds of straight-line winds.
(3) --6ust speeds computed as 125 of A.
(C)- tornado windspeeds based on 1950-75 data.
Probabilities Return periods
- dindspeeds in miles per bour (A)
(3)
(C) 10 1 year 38 mph 48 mph 10-' per year 10 61 76 10-8 100 72 90 10-'
1,000 86 108 10" 10,000 101 126 102 mph 10-5 100,000 113 141 171 10-*
1,000,000 125 156 237 l
10*'
10,000,000 297 I
6-24
- 10. o 50 10 0 15 0 200 250 300 mph M
\\u lo-'
'T n
i b
i j
\\
PEAK GUST (1955-72)48 States in SELS LOG
-m
\\
.g
\\
lo G M M
\\%,
O
\\
m t-wm N
B 10
A
\\,
\\C im
\\s
~
\\
10 '
\\
\\
\\
-s
\\
10
\\
'\\
N o,3 C
9
~>,#
10-*
G lo #
1 To 10
- Figure 10. Probabilities of straight-line and tornado winds.
The probability of strai ht-line winds (gust) is higher than 6
tornado winds when windspeeds are less than 130 mph. Peak gusts from SEIS IDG are also shown in this figure for comparison pur-poses.
6 -25 For secondary or short-life structures, one may wish to use the maximum windspeeds corresponding to the low-or even high-probability category defined above.
The storm characteristics in Table 14 were computed for these three proba-bilities. If one wishes to protect a structure against the high-probability (10)
windspeed, a 108-mph gust of straight-line wind is to be used as design criteria.
Tornadoes may be disregarded for such a high-probability case.
Table 14 Characteristics of storms corresponding to three probability categories, remote (10), low (10**),
and high (10 per year). Air density at the Institute site, 910-ft IEL, is assumed 1.2 kg/m8 and radius of maximum wind,150m for computational purposes.
Probabilities (per year)
Remote (10'#)
Low (10)
High(10 )
Storm types Tornado Tornado Straight wind l
l Maximum total speed ('aph) 297 237 108 Translational speed (mph) 59 47 Max. tangential speed (mph',
238 fM Total press. drop mb) 135 8 86.6 1) 1 97 1.26 Rate of pr. drop (
see) 23 9 12.1 l
(psisee) 0 35 0.18 i I The translational speed, T, was assumed to be T = 0. 20 V.,
where V., is the maximum total speed ecpected to occur at the radius of maximum i
wind, r.. The radius of maximum wind or that of tornado 's outer core was assumed to be
- r. = 150 m.
l.
i
6 -26 Under the assumption of the combined Rankine vortex and cyclostrophic wind equation, the total pressure drop was computed from p V.'
P.
=
where P at the 910-ft MSL Institute site was assumed to be 1.2 kg/m'
'Ihe rate of pressure drop was computed from p V,'
and dr = Tdt
=
or dP T
a
? y*
dt T.
which is identical to Equation (3 ) of WASH-1300 It is recommended that the team of structure analysts determine the proper probability category to be applied to each structure or portions of structure of the Battelle Memorial Institute. Table 14 will, then, be used to determine the characteristics of the design-basis storm.
V L/
T. Theodore Fujita 5727 South Maryland Avenue Chicago, Illinois 60637 l
l
APPENDIX LET OF VIOLENT TORNADOES WITHIN 140-MILE RANGE OF BATTELLE MEMORIAL INSTITUTE
'Ihe DAPPLE Tape (1950-75) includes 16 violent (F 4 & 5 ) tornadoes which affected the area within a 140-mile range of the site. 'Ihese tornadoes are listed in Table 15 and shown in Figure 11 in this appendix.
Table 15 Violent tornadoes (F 4 & 5) within 140-mile range of the Battelle Memorial Institute.
16 tornadoes of this category are included in DAPPLE tape. Tornado with
- was surveyed by Fujita.
No.
Year Month Day Name of tornadoes F scale Path length 1
1953 Jun 8
Clevelard Tornado, OH 4
100 miles 2
1961 Apr 25 Shelbyville Tornado, OH 4
65 3
1%5 Apr 11
- Grafton Tornado, IN 4
25 4
1%5 Apr 11
- Kokomo Tornado, IN 5
75 5
1968 Apr 23-Newtonsville Tornado, OH 4
25 6
1968 Apr 23 Ripley Tornado, KY-OH 4
50 7
1%9 Aug 9
Cincinnati Tornado. OH 5
22 8
1972 May 14 InM an= polis Tornado, IN 4
29 9
1974 Apr 3
- 3 ear Branch Tornado, IN 4
28 10 1974 APr 3
- Frankfort Tornado, KY 4
36 11 1974 Apr 3
- Hamburg Tornado, IN 4
37 12 1974 Apr 3
- Kennard Tornado, IN 4
20 13 1974 Apr 3
- Madison Tornado, IN 4
38 14 1974 Apr 3
- Parker, Tornado, IN 4
22 15 1974 Apr 3
- Sayler Park T., IN-KY-OH 5
21 16 1974 Apr 3
- Xenia Tornado, OH 5
32 i
Ik x.
1 I
'I NO.J
.l F5, APR il. 65 l l
NO.4 g
j
!I F5,* APR 3,74 l
\\*'
NO./4
-)
.g1 BATTELLE MEM. INSE NOl.8 /NOIE NO./6 l
es,O l'
t F5, APR 3,74 qA,
/02 I
/
/
M ' t **
NO.//
NO.3 S
l F4, APR 23,68
['
39 sk NO.7
- S* 5 9s F5, AUG 9,69 NO.6 jy )
F4, APR 23,68
/ lg7 g
f(o NO.9 ?
h
.I
- /
NO 13
,/-'^g
_f
\\
3.
/
y F4, APR 3,74 L
0 NO.10 gs
(
s Figure i1. Paths of violent tomadoes within 140 miles of the Battelle site. The Xenia tonado (F5) of April 3,1974 was the closest to the site.
(k i.
3.._.__.-.-.-
\\
/
F4, Jun 8,5 l
g,65 Y ' NO,3 I
i FS, APR ll,65 l i
NO.4 g
!\\
. FS, APR 3,74
\\
f
/
NO./4 l BATTELLE MEM. INST fF4, APR 3,74
.j
- f JNO./2 NO.16 g,e i
F5, APR 3,74 g s,#*Na2 l
,)
Nai, $'+
l n,./
NO.s S
]
F4, APR 23,68
$+
4 gS 'T
}
, APR 23,68 h
g
/5 AU 9 69
- O j
NO./J Q
/
p\\
F4, APR 3,74 0
NO.10 g5
}(.
A Figure i1. Paths of violent tornadoes within 140 niles of the Battelle site. The Xenia tornado (F5) of April 3,1974 was the closest to the site, i
k of
- - - - 3_._._._._.._._.-
1 I
1
/
- y. - a;5b-g es l
NO.J I
FS, APR ll,65 l l
Nu.4 g
'A' gI i
F5,jAPR 3,74
!k s
44
/
\\*"
NO./4 l BATTELLE MEM. INSE
/q $#
l j
h
, AP.9 3,74 l
V0.12
.c NO./6 f
g,6\\
l g
F5, APR 3,74 f
9 p?
4 NO.2 l
g
./
I j
9.D'1**
NO.fi NO.5 l
F4, APR 23,68 F4,AP 23,68
/S
/NO.9 *>
h 6
gh
j.J
'w
~
\\
Y NO /J ur ~ y I
)
/
F4, APR 3,74 y'
\\g No./0 3
i j
\\
\\
Figure iI. Paths of violent tornadoes within 140 miles of the Battelle site. The Xenia tornado (F5) of April 3.1974 was the closest to the site.
,