Design and Evaluation of a Low-Cost Water Quality Monitoring System

Thesis Abstract

This study addresses the persistent gap in timely, affordable water quality monitoring for rural and peri-urban communities by developing and evaluating a low-cost, sensor-based monitoring system capable of real-time data collection, transmission, and analysis. The problem stems from reliance on centralized laboratories and expensive multi-parameter sondes, which limit spatial coverage, increase response times, and constrain routine surveillance in resource-limited settings. The aim is to design, implement, and evaluate a scalable low-cost water quality monitoring system (WQMS) that integrates affordable sensors, open-source hardware, and cloud-based analytics to deliver reliable measurements of key water quality indicators. Specific objectives are to (i) identify a minimal yet essential set of water quality parameters (pH, temperature, turbidity, dissolved oxygen, electrical conductivity, and nitrate) suitable for low-cost deployment; (ii) develop a modular hardware-software architecture enabling plug-and-play sensor modules, remote data transmission, and user-friendly dashboards; (iii) validate sensor performance against reference laboratory analyses using a two-stage field trial across three water bodies with diverse contamination profiles, comprising a total of 180 sampled events over six months; (iv) implement robust data processing workflows including calibration, quality control, anomaly detection via regression-based calibration models and comparison with EPA/WHO guideline thresholds; and (v) assess system usability, maintenance requirements, and cost-effectiveness to determine scalability for municipal and community-led monitoring programs. The methodology adopts an explanatory sequential mixed-methods design. The quantitative phase involves a field evaluation with 60 paired sensor-laboratory samples per site (total n=180) collected biweekly from freshwater streams and a local irrigation reservoir. Laboratory analyses follow standard methods (APHA) for pH, nitrate (ion chromatography), turbidity (nephelometric), dissolved oxygen ( Winkler titration/modern probes), conductivity, and temperature. Data collection employs low-cost sensor nodes (Arduino-based) connected via GSM/LPWAN to a cloud platform, featuring automated calibration routines and data integrity checks. Statistical analysis uses linear and multivariate regression to develop calibration models, Bland-Altman plots for agreement assessment, and ANOVA to compare performance across sites and seasons. The qualitative component comprises semi-structured interviews with 12 stakeholders (water managers, community monitors, technicians) to evaluate usability, maintenance burden, and perceived reliability, analyzed thematically using a framework aligned with Technology Acceptance Model and Diffusion of Innovations. Ethical considerations include informed consent, data privacy, and mitigation of potential environmental impacts during sampling. Anticipated findings indicate that the low-cost WQMS can achieve measurement accuracy within ±5% for pH, ±10% for nitrate, and robust detection of turbidity and dissolved oxygen trends, with calibration models adapted to site-specific conditions. Data completeness and uptime are expected to meet minimum thresholds (?90% daily data return) after initial deployment, with acceptable agreement to laboratory results (limits of agreement within predefined bounds) and consistent performance across diverse hydrological settings. The study is expected to demonstrate that the system provides timely alerts when parameter thresholds are breached, enabling rapid management responses and community engagement. The contribution to knowledge includes (i) a validated design blueprint for an affordable, extensible WQMS that bridges the gap between laboratory and field deployment; (ii) a open-source software framework and modular hardware architecture that can be adapted to different parameter sets and regional needs; (iii) empirical evidence on the reliability and cost-effectiveness of low-cost sensors in real-world environments and across seasonal variability; and (iv) a practical implementation roadmap for scaling to municipal or community-based monitoring networks. The conclusion anticipates that a carefully calibrated, modular low-cost WQMS can provide scientifically credible, locally actionable water quality information, supporting proactive water resource management in settings with limited budgets. Recommendations include expanding parameter coverage to include microbial indicators where feasible, integrating edge-computing for on-device anomaly detection, establishing standard operating procedures for calibration and maintenance, and pursuing policy frameworks that recognize community-driven data as a supplement to official monitoring programs.