A statistical approach for landfill classification in Sri Lanka based on waste characteristics

Abstract

Several parameters have been used to characterize wastes in dumping sites. It is not possible to measure all the parameters and also those parameters are not equally important for management options which are decided on priority basis. In fact some parameters are closely related to other parameters. Therefore, parameters have to be clustered for classifying the wastes in dumping sites. A study was conducted to develop a statistical procedure using available tools to cluster landfills based on landfill waste characteristics. Five waste samples were collected from three landfills in the central province of Sri Lanka to compare the waste characteristics followed by the use of data to elaborate the clustering procedure. Correlation diagrams, principle component and cluster analysis have been applied for eight parameters; Moisture content, Ash content, Unit Volume mass, Lower heating value, Particle density < 2mm, Plasticity index, pH and Electrical Conductivity. Statistical analysis was able to extract two main principle components as primary, and secondary which accounted for 92.1 % of the total variability establishing two main classes which can be further improved to be more representative and precise with the increased number of samples subjected to the analysis.