ABSTRACT
The physicochcmical characteristics of water in rivers of middle Galma basin were monitored at twelve stations during a period of two years. These stations were located at various points on river Shika, river Kubanni and river Galma. The parameters measured were pH, electrical conductivity, temperature, total dissolved solids, nitrates, phosphates, sulphates, dissolved oxygen, COD and BOD. The data collected was subjected to multivariate statistical analysis techniques (Principal component analysis, Factor analysis, Cluster analysis and Discriminant analysis). Results obtained showed that principal component and factor analysis were able to isolate factors that could be used to effectively monitor the water quality of the basin. Total solids, Chemical Oxygen Demand, Nitrate and Phosphates (in that order), were significant in the analysis, which proved that agricultural activities were a major source of pollution in the area during the wet season and domestic wastes during the dry season periods. Cluster analysis clearly delineated the seasons and sampling stations. The stations on river Kubanni and its tributaries and also the stations on river Shika were near perfectly grouped. Cluster analysis was also applied to the months, where they were clearly divided into rainy and dry months. Discriminant analysis was applied to these clusters to extract stations that were more useful in terms of representative data as well as months that were inappropriately classified. The output results from Discriminant analysis were able to predict all these in a very convincing manner. From the overall results obtained, sampling could be limited to the identified stations S8 and S9 which are Gaskiya Bridge and Tudun Wada Bridge for Kubanni River and station S10 which is located near the Zazzau ginnery on the River Shika. This could provide representative results for the rivers concerned. It therefore showed that multivariate statistical analysis can be used for temporal and spatial data interpretation in the middle Galma basin.