49-60
ASSESSMENT OF NILE WATER QUALITY DATA USING EXPLORATORY DATA ANALYSIS AND CLUSTERING OF VARIABLES
Authors: YOUSRY M., AWADALLAH A.G., SALEM T
Number of views: 653
Considerable attention has been given to the monitoring of Nile River water quality and the establishment of a monitoring program carried out twice a year, in low and high flow seasons. A group of tests to quantify the physical, chemical, and biological quality are conducted for each water sample. The main objective of this paper is to assess Nile River water quality data using exploratory and cluster analysis to gain insight of spatial data grouping and relationships between variables. Having an insight of such variability is essential in detecting point source pollution.
Exploratory data analysis (EDA) is undertaken on the 8 years monitored data (2000-2007) during February and August. Comparing the mean value of each variable with the national standards of Law 48 / year 1982, it is found that water quality variables have mean values within allowable limits except COD. Grouping main Nile and its two branches results in a non-homogeneous dataset especially for salinity and biological variables. The large variability of FC, TC, TSS, OP, COD, BOD, NO2 and NO3, depicted through box plots, indicates that the Nile and its branches are exposed to different point sources of pollution. A significant correlation is found between several variables following their expected chemical behavior and exhibits evidence of mutual dependency between water quality variables. Cluster analysis is used to identify a hierarchy of these correlations. The statistical methods undertaken in this research gave a clear picture about the behavior of the variables and the anthropogenic activities in the study area.