Spatio-Temporal Evaluation and Quantification of Pollutant Source Contribution in Little Akaki River, Ethiopia: Conjunctive Application of Factor Analysis and Multivariate Receptor Model
More details
Hide details
Arba Minch University, AWTI, Arba Minch, Ethiopia
Rostock University, Department of Water Management, Rostock, Germany
Water Development Commission, MoWIE, Addis Ababa, Ethiopia
Zelalem Abera Angello   

Water Management, Rostock University, Justus Von Liebig Weg, 18059, Rostock, Germany
Submission date: 2019-12-26
Final revision date: 2020-03-15
Acceptance date: 2020-03-16
Online publication date: 2020-07-24
Publication date: 2020-10-05
Pol. J. Environ. Stud. 2021;30(1):23–34
Little Akaki River (LAR) is among the heavily polluted urban rivers in Ethiopia. A bimonthly physico-chemical and heavy metals water quality analysis was conducted aimed at assessing the spatio-temporal characteristics and quantifying sources contributing to the pollution during dry and wet season at 22 montoring stations. Accordingly, laboratory analysis results indicated that most of the constituents deviated from the national and international guideline limits and the river is critically polluted at the middle and downstream segment. Factor Analysis (FA) was used to qualitatively determine the possible sources contributing to the pollution of LAR where three factors are identified that determine the pollution level during wet and dry season explaining 79.26 % and 79.47 % of the total variance respectively. Agricultural and urban runoff (nonpoint pollution sources), industrial and domestic waste are the three dominant factors that contribute to pollution in LAR. On the other hand, pollution sources of heavy metals in LAR are mostly dominated by industrial release whereas urban washouts from garages and automobile oil spills are other possible sources. Cluster Analysis spatially grouped all 22 monitoring stations into four and three clusters during the dry and wet season respectively. USEPA’s receptor model, UNMIX, was used to quantify the composition and contribution of LAR constituents. The model predicted quite well with a minimum Signal to Noise ratio (S/N) of 2.71 and 2.16>2 and R2 of 0.91 and 0.88>0.8 for the dry and wet season respectively. The UNMIX model effectively predicted the water quality source composition with a model predicted to measured ratio (P:M) of 1.04 and 1.16 during the dry season and wet season with an average percentage error of 1.38 % and 17.13 % respectively. LAR water quality management approach incorporating all the three pollution sources could be feasible.