ANALYSIS OF WATER-LEVEL--LAKE JACKSON


about the same level. The over-all decline was 3.7 feet in 1952-53
compared to only 2.6 feet in 1962-63; yet, rainfall at Quincy,
Monticello, and Tallahassee was 0.7 feet greater in 1952-53. No
conclusion regarding ground-water outflow can be made based on
this evidence, however, because rainfall at Quincy, Monticello, and
Tallahassee is not necessarily always truly representative of rain-
fall at Lake Jackson. Future observations during a prolonged
drought period may indicate whether ground-water outflow has
changed appreciably, but such observations likewise may not be
conclusive because sinkholes may develop at any time.
 Figures 10 and 11 show that the rise of the lake level is closely
related to areal rainfall. Thus, if yearly evaporation and seepage
are virtually constant, the yearly net lake-level change should be
directly related to yearly rainfall. Data given in table 2 suggest
that such a relation exists but the plot of these data in figure 13
does not expressly define the relation.
 The scatter of the data in figure 13 is not unusual for hydrologic
correlations of this type but part of the data appear to be grouped
in a way that is not independent from the sequence of years. This
type of grouping suggests that the required underlying assump-
tions may not have applied uniformly during the entire period
under consideration. For example, such an effect might result
from a pronounced change in seepage.
 In figure 14 similar data are plotted separately for the individ-
ual rainfall stations involved. The scatter of these plots more
nearly conforms to that which might be due primarily to the non-
representativeness of the rainfall data with respect to Lake
Jackson. Thus, the apparent grouping that results when these data
are averaged may be only coincidental.
 A more adequate measurement of rainfall at Lake Jackson
would be required to evaluate the inconsistencies which abound in
figures 13 and 14. Some scatter still would persist, however,
because the effects of variations in the seasonal distribution of
rainfall are not entirely compensating within the period of a year.
 If, tentatively, the relation between yearly rainfall and net
yearly lake-level change shown in figure 13 can be presumed to be
defined, further speculations can be made regarding the relative
frequency of a given yearly lake-level change. The relative fre-
quency of areal rainfall, as represented by the average of rainfall
at Quincy, Monticello, and Tallahassee, is shown on a cumulative
basis by the graph in figure 15. By means of 'the relation between