Investigating the Potential Environmental and Operational Impacts of the Desalination Plants in Egypt

Abstract

Water scarcity has emerged as a critical global challenge, particularly for countries situated in arid and semi-arid regions. Desalination has become the most promising non-traditional solution for securing a reliable and sustainable supply of freshwater. Egypt has increasingly invested in the construction of seawater desalination plants along its coastlines, particularly within new coastal cities. This study presents a new contribution, as it may be considered as the first attempt to conduct a large-scale study across Egypt, encompassing water quality datasets from seawater reverse osmosis (SWRO) desalination plants. This work offers a comprehensive assessment by evaluating three key water streams (feed water, brine, and permeate). The main objective of this study is to investigate the environmental, water quality status, and operational performance aspects of eight full-scale SWRO desalination plants along the Red Sea and Mediterranean coasts. Hence, 288 water samples were analyzed to identify different physical, chemical, and biological parameters during 2022. Firstly, statistical analysis and water quality indices approaches were applied to evaluate the various water quality states based on the available international and Egyptian guidelines. Most seawater properties met permissible limits, except for elevated turbidity, oils and greases, and total bacteria in one plant, and elevated iron, manganese, and silica in another. The findings highlight the influence of intake type and plant location on operational efficiency and environmental risk mitigation. Secondly, three saturation indicators (LSI, S&DSI, and RSI) were applied to determine the potential for scaling or corrosion in different types of water in plants. Feedwater generally showed scale-forming yet non-corrosive characteristics, while brine exhibited conditions ranging from balanced to scale-forming with varying corrosion risks. Permeate was consistently undersaturated and corrosive. pH emerged as the most sensitive driver of scaling behavior, whereas low calcium concentrations intensified corrosivity. Thirdly, this study introduces a machine learning-based framework to predict and classify the drinking WQI score and water quality class (WQC) for plants. The study integrates real plant operational data with multiple supervised algorithms to enhance predictive accuracy and operational decision-making. Multiple linear regression produced the best predictive accuracy, while XGBoost achieved the highest classification performance. SHAP analysis identified total bacteria, TDS, sodium, chloride, and residual chlorine as the most influential parameters. Finally, this research provides a quantitative framework designed to measure the contribution of thesis outputs across relevant Sustainable Development Goals (SDGs).