Environment and Ecology Research Vol. 11(1), pp. 165 - 182
DOI: 10.13189/eer.2023.110112
Reprint (PDF) (1435Kb)

Principal Component Analysis Approach in Klang River Water Quality Index Modelling

Norhafizah Ghazali , Zalina Mohd Ali ^*
Department of Mathematical Sciences, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor, Malaysia

ABSTRACT

Water Quality Index (WQI) has been widely used in monitoring river water quality to get the information of the river water rapidly and easily. The basic measures of WQI development are variable selections, weight determination, development of sub-index functions and weighted sub-index aggregation. In Malaysia, the Department of Environment (DOE) uses the WQI formula which is calculated using an aggregation function based on six variables, namely: Dissolved Oxygen (DO), Biochemical Oxygen Demand (BOD), Chemical Oxygen Demand (COD), Ammoniacal Nitrogen (AN), Suspended Solids (SS) and potential of Hydrogen (pH) to indicate the general status of river water quality in Malaysia. However, the weighting of sub-indices determined by DOE and used in development of WQI is subjective in nature as it is published based on expert opinion. Besides, the natural inter-relationships between variables are also less applied. The relationships between these variables are important and should be considered due to the unique, complex, and diverse dimensional characteristics of the river. Therefore, based on the principal of the relationship that exists between river water quality variables, Principal Component Analysis (PCA) approach is appropriate to meet these requirements. Hence, in this study, three new models were introduced through modification of the loading calculation from PCA which was proposed for the recalculation of the Klang River IKA. The new calculation approach is through the PCA-WQI model with a) normalization of the square of component loading, b) normalization of component loading and c) normalization of the absolute value of the component loading. Comparisons are also made using the PCA-WQI approach commonly used by other researchers. The results showed that the PCA-WQI model using the normalization of component is the best model through a value of 65.54% of the training data (2008-2014) can be classified in the same general river status group as DOE while 73.60% was recorded for the test data set (2015-2018). Comparison was made with other existing PCs methods. The PCA-WQI new model also demonstrates an inversely relationship with DOE-WQI through the different score values with the existing PCA-WQI method which implies that the lower the value of PCA-WQI, the better the water quality.

KEYWORDS
Principal Component Analysis (PCA), Water Quality Index (WQI), Klang River

Cite This Paper in IEEE or APA Citation Styles
(a). IEEE Format:
[1] Norhafizah Ghazali , Zalina Mohd Ali , "Principal Component Analysis Approach in Klang River Water Quality Index Modelling," Environment and Ecology Research, Vol. 11, No. 1, pp. 165 - 182, 2023. DOI: 10.13189/eer.2023.110112.

(b). APA Format:
Norhafizah Ghazali , Zalina Mohd Ali (2023). Principal Component Analysis Approach in Klang River Water Quality Index Modelling. Environment and Ecology Research, 11(1), 165 - 182. DOI: 10.13189/eer.2023.110112.