Identification of relationship between water quality response and land use attributes : case study, Gombak river watershed, Malaysia

Abstract

This study was conducted on Gombak River and its watershed. Gombak River is situated mainly in the Gombak District in Selangor state and its lower zone is situated in the Malaysian Capital Kuala Lumpur. The motivation of this study is to ensure sustainable development, which is essentially has effects on surface water quality as 97 per cent of Malaysia’s water supply is sourced from it; and to find a low-end alternative for water quality monitoring techniques, which are typically expensive and tedious. This will allow decision makers to plan safely to improve the quality of life while maintaining sustainable development. The main aim of this study was to identify a relationship between water quality and land use attributes through selecting the appropriate data generation methods and model building that can enable prediction and forecast of water quality index with high accuracy. This model coupled by the data generation method, could later be utilized as an important extension to major planning software in order to assist planners and decision makers to plan and decide safely for watersheds zones. The thesis examined the water quality of the river focusing on six water quality parameters, these were BOD, COD, AN, SS, pH, and DO in three monitoring stations as well as conducted laboratorial analysis in a fourth one. The data were then analyzed in order to understand its behavior and response to the activities within the watershed. The watershed of Gombak River was redefined using MapInfo from the recently updated coordinates of Gombak River, and the LU/LC data were extracted from the watershed for a discrete 20-years period. LU/LC data focused on five categories, these were build-up area, residential area, commercial area, industrial area and forest area. Methodologically, the thesis used a quantitative method of analysis such as ANN’s Back propagation algorithm in order to examine the correlation and magnitude of relationship between the chosen water quality and LU/LC indicators. The study introduced data generation method for the prediction and forecast of LU/LC data within the watershed. The method used exponential model equation, Lagrange model equation 3rd & fourth degree polynomial fit and saturation growth-rate model in order to generate the required data. In order define the relationship between water quality response and land use attributes, the study introduced the LA-WQI model. This model was developed by associating the appropriate loading factors to a set of sub indices. The loading factors were obtained based on peak factor theorem by assigning weights to different categories of land use indices. The model was applied on the average values of the percentages of land use every year, and was compared with average values of yearly actual and predicted WQI. The findings revealed that the water quality index of the three stations located along the Gombak River was deteriorating from upstream to downstream. Station 18 showed the highest fluctuation during the mentioned period. This is due to increasing number of ongoing development projects within this station. The mean WQI for monitoring stations along Gombak River, puts station 24 as clean river water, station 18 as slightly polluted and station 17 as polluted river water. ANN’s Back propagation algorithm was able to predict WQI with high accuracy at station 18 and 24. This high accuracy was due to the fact that the configured ANN was able to detect a direct relation between sub indices used and the water quality index; in this case the relationship was the DOE-WQI model. On the other hand, when land use parameters were trained, the differences between predicted WQI and actual WQI were slightly higher than that of trained with water quality variables. This is due to the continuous changes in chemistry of water quality variables which doesn’t indicate the actual water quality. However, the training yielded high correlation between average values of predicted and actual WQI. The findings also revealed that as the activities increased throughout the watershed, the values of WQI quality decreased accordingly. The accuracy of prediction of the proposed LA-WQI ranged from 94.3% to 99.3% between Actual DOE-WQI and LA-WQI for station 24 of Gombak River. The same approach was applied for station 18 at Gombak River. The results of predicted WQI obtained using LA-WQI, showed a continuous decrease of water quality. Prediction values of WQI for DOE, ANN and LA-WQI had an accuracy range of 97.9% to 99.5 %. In order to test the sensitivity of data generation method and the LA-WQI model, withheld data were used for the first time for validation. The results indicated high precision in the prediction and forecast of WQI. Generally, there is an appreciable effort to keep Malaysian rivers clean, however in order to achieve an accepted level of sustainability, all developmental schemes within a watershed vicinity should be conducted in accordance with a specific allocation of time and space which follows a certain model. This will help in producing more accurate forecasting of LU/LC sub-indices and hence WQI. Despite the high accuracy attained by the application of LA-WQI model on Gombak River; it has not yet been tested on other rivers due the limitation of the scope of this study. It is recommended that future studies should be able to further test the current model on a regional scale.