ML19291B867
| ML19291B867 | |
| Person / Time | |
|---|---|
| Issue date: | 10/05/1979 |
| From: | Abbey R, Macdonald J NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES), TEXAS TECH UNIV., LUBBOCK, TX |
| To: | |
| Shared Package | |
| ML19291B868 | List: |
| References | |
| NUDOCS 7911140084 | |
| Download: ML19291B867 (10) | |
Text
r-e.
lith CONFERENCE ON SEVERE LOCAL STORMS OCTOBER ?-5, 1979, KANSAS CITY, MISSOURI available from American Meteorological Society, Boston, Massachusetts COMPARISON OF 1EE NSSFC AND DAPPLE 7013 ADO DATA TAPES James R. Mcdonald Institute for Disaster Research D
1D
~
~
P Texas Tech University l h' '
J 1.ubbock, Texas 79409 2
3 Robert F. Abbey, Jr.
Office of Nuclear Regulatory Research W
U.S. Nuclear Regulatory Commission Washington, D.C.
20555 s
/
1.
INTRODUCTION cant awancement w'as ma e with the advent of the Fujita-Pearson rating system (Fujitam 1971; Records of United States tornado occurrences Fujita and Pearson, 1973). Since 1971, tornadoes and their destructive effects date back to the have routinely been tabulated using the FPP ratings.
late 19th century. Various types of records Attempts have also been made to assign an FPP or have been maintained since the U.S. Weather at least an F rating to a11 tornadoes reported Bureau be::en tabulation of tornado statistics since 1916 (by the University of Chicago) and in 1916. Compilation and utilization of these since 1950 (by the National Severe Storms Forecast data can be found in the earlier studies by Center). Utilization of these dets sets have Wolford (1960), Pauta, ed. (1969) and Court been described by Fujita (1978); Tecson, et al.,
(1970). Additionally, Asp (1963) documents (1979); Kelly, et al., (197t}, Schae f er, et al.,
the history of tornado observations and data (1979); and Abbey (1976).
sources. Statistical manipulations of this Tornado information has been assembled and data are found throughout the formal and informal filed on computer data tapes. These two tapas literature. Abbey (1976) briefly addressed will be referred to herein as the NSSFC and the some of these endeavors in his treatment of 1 APPLE tornado data tapes. (DAPPLE stands for tornado occurrences, characteristics, and Jamage Area fer Path Length and was coined by
'~
intensity ratings.
Abbey and Fujita (1975) as part of their hazard With the increased awareness of and model me'hodology.]
importance assigned to quantifying the risk one is exposed to, increased attention has been 2.
DESCRIPTIONS OF 'niE DATA TAPES focused on dev loping and validating herard rohnbility models, l.ar ge l y the result of ths N basic data source for both tapes is
.uclear industry and the desire of r.mula m,
Storin Data, a publication of the U.S. Department agencies to minimite the risk to the public of Comerce, National Oceanic and Atmospheric arising from severe natural phenomena, a range Administration, Environmental Data Service, of tornado hazard probability models have been National Climatic Center, Asheville, N.C. The created. The objective of these efforts is to basic difference between the NSSFC and DAPFLE be able to assess realistically and to extrapolate date tapes largely stems from the manner in with confidence the windspeeds associated with which the incomplete data is treated.
relatively lov occurrence probabilities; the Storm Dats contains systematic records of j
windspeed associated with a 1 x 10 per year var.tous types of severe storms, including tornadoes mbability of occurrence has been set as the and funnels alof t.
Specifically, the information im value for many nuclear power plants.
on tornadoes includes:
tornado hazard probability models have been (1) state, coun'y, community iitically reviewed tf Abbey and Fujita (1975)
(2) year, month, day, time and Abbey (1976).
(3) path length, path width (average)
In order to perform a tornado hazard assess-(4) deaths, injuries ment for any given region of interest, a consistent (5) property and crop damage and high-quality data set is desirable. When (6) narrative description of damage extrapolations are made from a 30 year data The data are assembled by state and published record to recurrence intervals of ten million
- monthly, years, the quality of the data set becoa.as very important. The type of information required is often not available. The existing records tabulate time, date, and location of tornado occurrence, but often do not record intensity (windspeed) and the damage path area. A signifi-
- Ob
p--
The NSSFC tornado data tape contains reports en intensity rating or a path dimension. Thu s, t he
.f tornadoes since 1950, the format of the tepe NS5FC data tape has reasonable credibility for is set uo to include the following information those tornadoes that are rated, but the data are for each tornado occurrence incomplete. The amount of "incomple t eness" was tes (1) year, month, date, time significantly from one geographical region to (2) latitude and longitude of beginning and another.
termination points of the tornado path Examination of the DAPPl.E tape shows that an
( 3) type of path, per cent on the ground, FPP rating has been assigned to all reported stora type and rotational sense tornadoes since 1916. Under Dr. Fujita's tutel*S*.
(4) path length and mean path width all avellable sources from his personal file have (5) deaths, injuries, and damage classifi-been Jtilised; additionally, his vast in-Jepth cation experience in damage surveys has been used to fill
- 16) states and counties affected in gaps in the tornado data record. Where little (7) Fujita Pearson scale ratings or no information on a particular tornado occurrence i s k nown, save for the f act that is occurred, a Dr. T. Theodore Fujits and his staff at t h,e value of FO was assigned to the tornado. The University of Chicago have casembled the DAPPLE assumption employed is that had the tornado been tornado data tape.
Fujita (197M produced the more severe, more would have been reported on it.
first map of tornado tracks coded according to When the length was reported as short, unknown, an intensity classification. Although developed or merely touchdown, a value of one mile (P =0) e g
in several stages, the current version contains was assigned. Unknown path widthe vere also I
reports on tornadoes from 1916 to the present.
assigned a value corresponding to P = 0.
y U.S. Weather Bureau records and Storm Data list-ings were the basic data sources. The history 4.
BASIS FOR COMPARISON and detailed descriptions of this data record can be found in Te u n, et al., (1979). Prior A comparison of the two tapes is presented to assembling data on the tape, all known tornado for an arbitrarily sele 6ted 5-degree square paths were hand plotted on large scale maps of located east of the Rocky Mountains. The results the United States, are indicative of what is found for other slailer Because the DAPPLE method of tornado hazard
- regions, assessment is based on path length, segments of The formats of the two tapes are dif ferent each tornado path are associated with 15-minute and do not readily tend themselves to direct subboxes defined by latitude and longitude. Each comparison.
Thus, the data on the NSSFC tape subbox is identified by the latituda and longi-was first converted to the DAPPLE tape format.
tude of the northwest corner of the one-degree The following comparisons are presented:
Marsden square that contains the subbox. The (1) number of tornadoes per year pe* F-scale subboxes are numbered like the elements in a (2) number of tornadoes per month per F-scale 4 x 4 matrix.
(3) number of tornadoes per two-hour tire The DAPPLE tape contains:
periods (touchdowa) per F-scale (1) year, moath, date, time (2) F scale rating The number of tornadoes per year is grapht-(3) deaths and injuries cally shown in Figure 1 and listed in Table 1.
(4) Pearson path width rating The number of tornadoes on the NSSFC tape in (5) For each 15-minute subbox that contains comprised of rated and unrated tornadoes. Table 3 a portion of the tornado path, the indicates the means and standard deviations of the f ollowing information is listed:
numbe r of tornadoes per year. For all comparisons (a) latitude and longitude of the north-reported herein, the period of record is 1959-west corner of the one-degree square 1974 inclusive.
(b) location of the subbox Table 2 presenta the distriuution of tornadoes (c) path length by month as a function of F-scale; whereas Table 4 (d) character of the tornado path demonstrates the variation of tornadoes according (e) direction toward which the tornado to their initial touchdown time and F-scale.
la moving Twenty-seven per cent of the reported tornadoes in this region during this 16 year period are not 3.
BACKCROUND ON DATA AQUISITION rated. These comparisons are provided for your perusal and are not meant to be definitive at Much of the data listed above for both tapes this time. Detailed examinations as to the reasons is not available in the basic data sources.
behind the more significant differences will be Under the guidance of NSSFC personnel (Mlen D.
the subject of a more extensive report scheduled Pearson and Joseph T. Schaefer), newspaper for completion in early 1980.
accounts and other literature sources have been Clearly, both tapes have their limitations.
t ulled to obtain additional information on Use of the NSSFC data without correction for Fcific tornado occurrences and characteristics.
newe red t ernadoes may lead to a result that is it, for example, sufficient inf ormation un st J not truly rtpicsontative of the hazard probability an Storm Data or one of the other sources to assign of a given region. On the other hand, there is an F-scale rating to a particular tornado, then some uncertainty as to tra accuracy of the ratings it would be done. Where inf ormation was not of some tornadoes on the DAPPLE t ae, the consistency available to permit such a determination, those of the ratings assigned notwithstanding Careful undefined parameters are lef t blank on the NSSFC and rational judgements must be employed to tornado data tape. This ef fort has been completed ascertain which tornado data set will yield the for all contiguous states back to 1950. As a most acceptable results for the region under result of this philosophical approach, about consideration.
M t of the tornadoes on the tape are not assigned
_1 m
- 0)l 2
r-
,50 44 0
- Rated Tornadoes
' NSSFC L30 l
- Unra ted Tornadoes 12 0 uO DAFPLE-All Tornadoes Cated 10v ti-30
?
80 5
70 60 50 5
1__
R__ 40
\\
I s
s j
-- 0 30
+
h-h 2
G
_R
\\
- 20 h
g y
S
\\
y i
bh I I I k If k
~
h
~"
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 NUMBER OF TORNADOES PER YEAR FIGURE 1 NUMBER OF TORNADOES PER 2 TABU 1ATED IN NSSFC AND DAPPLE DATA TAPES TABLE 2 MEANS AND STANDARD DEVI ATI )NS OF NUMBER OF TORNADOES PER YEAR, 1959-1974 NSSFC D.' Pl;E At t tortuJoes mean 44.13 40.82 a.d.
32.62 24.91 All rated tornadoes mean 26.88 40.81 s.d.
28.39 24.91 s
^N 7 336
p.
TABLE 1. NUMBER OF TORNADOES PER YEAR PER F-SCALF, 1959-1974 Y EAR FO F1 F2 F3 F4 F5
'OTAL UNR Hg 32 3g 3g 3g 3p 3
g N only 1959 0 4 0 9 2 9 1 10 1 0 0 0 24 32 20 1960 0 3 3 16 1 8 0 8 1 0 0 0 31 35 26 1961 0 1 0 9 3 11 3
7 2 1 0 0 42 29 34 1962 2 3 4 6 0 5 1
2 0 0 0 0 13 16 6
1963 1 2 1 4 10 19 4 i3 0 4 0 0 47 42 31 1964 0 3 5 5 8 9 1
6 3 0 0 0 29 23 12 1965 0 3 5 11 5 9 0
3 0 0 0 0 25 26 15 1966 2 4 7 1 1 8 2
0 0 0 0 16 15 5
1967 2 7 12 27 9 19 1
4 0 0 0 0 50 57 26 gy 1968 0 11 10 3 1 4 1 4 1 0 0 0 25 22 12 P
9 1 4 0 9 5 9 1 5 1 1 0 0 19 28 11 a 170 s 5 5 20 20 9 18 4
6 0 0 0 0 50 49 12 1971 5 9 18 20 20 21 8
2 1 4 g2 63 58 9
1972 0 1 16 16 12 8 2 4 2 2 00 40 31 6
1973 11 11 52 55 24 21 0 0 0 0 0 0 91 87 4
1974 4 28 27 26 20 15 i 22 17 19 15 2 2 141 103 47 TOTAL 33 99 182 237 130 193 50 93 33 27 2 4 706 653 276 (N denotes NSSFC data record; D denotes DAPPLE data record)
TABLE 3. NUMBER OF TORNADOES PER MONTH PER F-SCALE, 1959-1974 MONTH F0 Fl F2 F3 F4 F5 TOTAL UNR HE EE EE EE fD ED g
p N only JAN 0 0 4 9 5 5 1 5 1 1 0 0 16 20 5
~
FEB 1 3 4 7 3 9 1 4 0 1 0 0 22 24 13 MAR 2 8 8 17 18 26 5 15 3 0 0
0' 73 66 37 APR 7 11 60 61 49 47 32 38 25 22 2 4 240 183 62 MAY 10 28 38 47 22 38 6 10 3 3 0 0 120 -
126 41 J UN 8 15 34 37 10 18 0 10 1 0 0 0 81 80 28 JUL 3 8 15 15 5 16 1 2 0 0 0 0 36 41 12 AUG 1 23 4 12 2 7 1 3 0 0 0 0 45 45 37 SEP O 2 3 6 2 4 0 3 0 0 0 0 14 15 9
OCT 1 1 2 4 0 4 0 2 0 0 0 0 10 11 7
NOV O O 5 5 3 10 1 1 -
0 0 0 0 19 16 10 DEC 0 0 5 17 11 9 2 0 0 0 0 0 30 26 12 TOTAL 33 99 182 237 130 193 50 93 33 27 2 4 706 653 276
[N Jenotes NSSFC data record; D denotes DAPPLE data recordj
)
e-TABLE 4. NUMBER OF TORNADOES PER 2-HOUR 71HE PERIODS PER F-SCALE, 1959-1974
[ time of :nitial touchdown normalized to CST)
TIME FO Fl F2 F3 F4 F5 TOTAL iTNR HE EE 3 -E E E' EE EE E
g N only 0- 159 3 5 5 2 2 6 1 2 0 0 0 0 24 15 13 200- 359 0 3 9 8 14 19 3 6 0 0 0 0 40 36 14 400 559 1 3 0 9 4.5 0 2 0 0 0 0 19 19 14 600 759 0 1 7 4 1 4 0 1 0 0 0 0 13 10 5
800- 959 1 2 6 10 4 4 1 2 1 0 0 0 22 18 9
1000-1159 1 3 9 12 7 7 2 5 2 1 0 0 29 28 8
1200-1359 3 14 25 23 12 15 6 14 2 1 1 0 64 67 15 1400-1559 6 19 20 14 19 27 9 9 16 8 1 2 125 98 54
,eA+
1600-1759 6 17 27 37 24 37 9 18 4 3 0 0 117 112 47 1800-1959 5 12 42 43 15 24 12 12 3 9 0 1 122 101 45 e
2000-7159 6 10 18 22 14 17 4 7 3 4 0 1 77 61 32 2200-2359 1 2 14 17 14 16 3 8 2 1 t0 0 54 44 20 not specified 0 8 0 17 0 I?
0 7 0 0 0 0 0
44 0
TOTAL 33 99 182 237 130 193 50 93 33 27 2 4 706 653 276
[N denotes NSSFC data record; D denotes DAPPLE data record!
[UNR denotes unrated tornadoes appearing in NSSFC tape only!
In nuclear power plant criteria, the low acceptable
- Fujita, T.T.,
1971: Proposed characterization of level of risk reflects the uncertainty in the tornadoes and hurricanes by area and intensity.
data set.
Acceptable levels of risk for other SMR P R e s. Paper No. 91, Univ, of Chicago, 111.
applications of tornado resistant design should 1978: Work'..ook of tornadoes and high winds, also. account for these uncertainties.
SVR P Res. Paper No. 165, Univ, of Chicago, Ill.
._ _ _, 1976: U.S. Tornadoes, 1936-1974. map 5.
ACKNOWLEDCEMENTS eva11able from the Univ, of Chicaeo. Ill.
, and A.D. Pearson, 1973: Repults of F P P The authors are indebted to Dr. T. T. Fujita, classification of 1971 and 1972 tornadoes.
The University of Chicago, for the many hours Preprint s, Eighth Conf. on Severe Local Storms, spent in fruitful discussions of tornado damage, Denver, CO, October 15-17, 142-145. AMS intensity ratings, and data assessments; we
- Kelly, D.L.,
et al., 1978: An augmented tornado would also like to give Dr. J. T. Schaefer, NSSFC/
climarology. Hon. Wea. Rev., 106, 8, 1172-1183.
TDU, a lagnisppe of his very own for his very Pautz, ed., 1969: Severe local storm occurrences willing cooperation in furnishing us with the 1955-1967. ESSA Tech. Memo. WBm FCST 12,
'tu of & teorological Operations, Silver
.is t up-to-date data on the NSSFC tape. This wm L
.as partially supported under contrac t hkC-64 spaiug, MD.
J45 between the Texas Tech University and the
- Schaefer, J.T., et al., 1979: Tornado track Of fice of Nuclear Regulatory Research, U.S.
characteristics and hazard probabilities.
Nuclear Regulatory Commission.
Fifth Int. Conf. on Wind Engineering, Colorado State University, Fort Collins, CO., July 8-14, 6.
REFERENCES II-1 !!-1-15.
Teccon, et al., 1979: Statistics of U.S.
tornadoes Abbey, R.F.,
1976: Risk probabilities associated based on the DAPPLE (Damage Area Per Path Length) with tornado windspeeds. Proceedings, Symposium tornado tape. Preprints, Eleventh Conf. on
% Tornadoes: Assessment of Knowledge and Severe Local Storms, Kansas City, MO, October 2-5, m.ications for Man, Texas Tech University,
.AMS.
l ubbock, Texas, June 22-24, 177-236.
- Wolford, L.V., 1960: Tornado occurrences in the
, and T.T. Fujita, 1975: Use of tornado United States. Tech. Paper No. 20, revised, U.S.
path lengths and gradations of damage to Weather Bureau, Washington, D.C.
assess tornado intensity probabilities. Preprints, Ninth Conf, on Severe Local Storms, Norman, OK, This report was prepared as an account of work October 21-23, 286-293. (AMS) sponsored by an agency of the United States Covern-Asp, M.O., 1963: History of tornado observations ment. Neither the United States Covernment nor any and data sources. Key to Meteor. Records Doe, agency thereof, or any of their employees, makes any No. 3.131, U.S. Weather Boreau, kash., D.C.
warranty, expressed or implied, or assumes any legal Court, A., 1970 Tornado incidence maps. ESSA liability or responsibility for any third party's Tech. Memo. ERLTM-NSSL 49, National Severe use, of any information, apparatus, product, or pro-Storma Laboratory, Norman, Ok.
cess disclosed in this report, or represents that its use by mch third party would not infringe privately P
owned ri.,ata. The views expressed in this paper are go not necessarily those of the U.S. Nuclear Regulatory Commission.
m7 3%
o %Q wq o i
x
.