Analysis of data collected between 1950 and 1995 in the state of India
na indicates that the state experiences an average of nearly 19 tornad
oes each year, which occur on approximately 8 or 9 days during an aver
age year. Analysis of storm dimensions indicates that the annual proba
bility of a point location in Indiana being impacted by a tornado is 1
.4 x 10(-4) per year. The largest number of the more intense tornadoes
(F2 and higher) typically occur from March to June, during the aftern
oon and evening, although tornadoes have been observed in each month o
f the year and each hour of the day. Inter-annual variability of torna
dic events in Indiana is large and may reflect both physical causes (e
.g., the variability of cyclone passages or jet stream location) and r
ecording bias (e.g., increased public awareness during the 1950s). Usi
ng the data set from Indiana, statistical analyses of the temporal and
spatial variability of tornadoes are undertaken. Analyses of tornado
reports by county indicate that, in addition to area and population, s
urface roughness is a significant determinant of event frequency. Anal
yses presented herein also indicate that in contrast to previous studi
es, (1) no statistically significant hebdomadal cycle of tornado repor
ting is evident in the data and (2) the diurnal cycle of tornado recor
ds is evident for all F-scale classes and is not confined largely to t
he weak tornado categories. Analyses of the physical dimensions and F-
scale ratings of tornadoes in Indiana indicate that there has been a d
ecline in reported tornado intensity, width, and length between 1950 a
nd 1995. These changes are in accord with evidence of a decreasing num
ber of reports of strong and violent tornadoes across the United State
s over this period and appears to be principally a result of reporting
biases at the beginning of the record.