Using Real-time Sensing of Surrogate Water Quality Parameters to Characterize Microbial Dynamics of Urban Waterway

Real-time time sensing and control of urban stormwater systems using "smart" online sensor and actuator networks is an inexpensive and versatile alternative to traditional large-scale, expensive engineering interventions. Yet, while initial focus of intelligent water infrastructure was on volumetric control (e.g. combined-sewer overflow reduction), imperative to the functioning and health of an urban watershed is its underlying water quality, with nonpoint source pollution (e.g. urban stormwater runoff). However, real-time sensing and control of water quality is a far more difficult undertaking due to (i) the lack of sensors able to immediately measure the most critical water quality parameters (e.g. fecal indicator bacteria) and (ii) the computational complexity to successfully model nonpoint source pollution dynamics.

This research aims to develop a real-time predictive model of local microbial dynamics at four locations in the Chicago Area Waterway System via in-situ, adaptive measurements of surrogate water quality parameters. Two types of data will be used to build this model: Real-time microbial data collected adaptively via a custom-built wireless sensor system deployed at each location that consist of wireless cellular sensor nodes (built based on the Open-Storm.org open source sensor node design) connected to: A YSI EXO3 multiparameter sonde with four sensor probes capable of measuring dissolved oxygen, pH, oxidation-reduction potential, conductivity, and fluorescent dissolved organic matter water quality parameters in real-time, A Matbotix ultrasonic depth sensor able to measure water level variation, and A Teledyne ISCO 6712 Portable Automated Sampler able to collect water samples on demand for later microbial analysis. Historical data of the Chicago Area Waterway System, which includes two sets of water quality data collected by The Metropolitan Water Reclamation District of Chicago: Continous Dissolved Oxygen Monitoring Program data collected hourly over a decade from over 25 locations. Ambient Water Quality Monitoring Program data collected monthly for over a decade also from over 25 different locations.

The initial predictive model developed from this data, along with the details of the sensor systems deployed will be presented.