against actual drought and flood situation inde-
pendently. It is noticed that 23 years fall in the
category of normal and nine years each fall in
¯ood and drought years, respectively. The average
weighted all India seasonal rainfall has been taken
to be 84.9 cm, standard deviation is 8.5 cms and
coefficient of variation (C.V.) is 10%. Thus in a
year when the average seasonal rainfall equal or
exceeds 93.4 cms, it would be a flood year and
when it equals or below 76.4 cms, it would be
come a drought year.
Table 1 shows the years of normal, flood and
drought years over India. Further correlation of
each year seasonal rainfall with the rainfall for
each of the 68 stations was calculated and a test of
significance at 5% level was carried out using
students ``t'' distribution test. (Very large samples
of data are considered in the present study. It is
presumable that the density is normally populated
and independent of the significant parameters
mean ( ) and variance ( 2 ). Hence the student ``t''
distribution for simplicity is applied in place of
Fisher-Z-Transform with the implicit indepen-
dence assumption). Only those stations qualifying
the significance test have been considered for this
study (in normal, flood and drought years). The
correlation of individual station rainfall with all
India seasonal rainfall is carried out to identity the
homogeneity associated with the nature of rainfall
activity of individual stations with the rainfall
character of the all India seasonal rainfall. (It is
not necessary that the correlation observed at each
station in the network will be strongly correlated
with those at neighbors. For example, between
Bombay and Pune there is 200 km distance but
rainfall ratio is 3:1, it is only due to orography).