An examination of the monthly average salinities of the Hooghly waters for the months of March, April and May for the entire period for which data are available, has revealed that the increasing trends of salinity are statistically significant.
The Control Chart Method of analysing the data for their 35-year homogeneity has, however, shown that the recent years’ data conform to a different range of variations. The increasing long-term average rates of year to year salinity in the months of April and May respectively, have appreciated by 198 and 305 per cent for Garden Reach over their pre-1951 rates, and 243 and 373 per cent for Cossipore. By analogy, it may be inferred that higher increases have occurred at Palta. For want of Palta data for the earlier years, a different approach was considered expedient.
Alternatively, when correlation studies were made directly of the monthly average salinities of (i) Garden Reach and Cossipore, and (ii) Garden Reach and Palta, separately for the two durations of the years 1927-49 and 1950-61 the slopes of the mean relation lines of the latter periods data were invariably found steeper, the increases being statistically significant. Obviously, the mean relations law for the 35-year period did not conform to a single straight line form.
The examination of more copious data of the daily salinities of the six separate years all separately corroborated the inadequacies of the straight line relationships. The behaviour of logarithmically transformed data of the daily salinities of the same six years, however, established the adequacy of the logarithmic linear fits.
On extending the logarithmically linear form of law for evaluating the mean relation between the monthly average salinities at Garden Reach and Cossipore, the slopes for each of the three months data for both durations of the years, were found to be not distinct from each other. Covariance analyses made by holding the Garden Reach salinities constant showed that the forms of the month to month relations and their expressions not significantly different, and that the data for the later period of years corresponded to a different datum.
Accordingly, two logarithmically linear laws were found sufficient to explain the mean monthly salinity relationships obtaining between Garden Reach and Cossipore with a unified, value of the slope for both the groups of the years. In terms of the original units, the mean relationships obtained were—
C=0.399G1.179—0.113 (G<6) for 1927—49
C=0.459G1.179—4.248 (G<8) for 1950—61
where G and G stand, respectively, for average monthly salinities observed (in parts per thousand) at Cossipore and Garden Reach. The corresponding expression for the related variations law between the average monthly salinities at Garden Reach and Palta was:
P=0.010G2.046—0.057 (G<8) for 1945—61
The greater-than-unity values of the unified indexes indicate the higher increases in the differential effects to be anticipated at upstream sites.
