The main objective of this chapter is to demonstrate a near-real-time water quality monitoring capability for addressing the concern of Total Organic Carbon (TOC) present in surface waters. TOC is a known precursor of disinfection by-products in drinking water treatment when using chlorine as a disinfectant. One of the most well-known disinfection by-products in this regard is Total Trihalomethanes (TTHMs); TTHMs develop due to the presence of free chlorine residual and TOC simultaneously in drinking water distribution networks when drinking water treatment plants cannot remove TOC from the source water completely. TTHMs is a family of suspected carcinogens and has been related to possible birth defects. The IDFM algorithm described in Chapter 13 will be applied in this chapter to:

create the daily monitoring capability of TOC in Lake Harsha, Ohio, USA that serves as the source for a drinking water treatment plant nearby,

exhibit the use of machine learning technique to estimate spatiotemporal distributions of TOC for monitoring water quality variations in Lake Harsha, verified by 4 statistical indexes, and

compare the monitoring effectiveness with and without the inclusion of data fusion and/or machine learning techniques.

In addition to data fusion and machine learning techniques, the optimal pumping schedule at the intake of a drinking water treatment plant will be discussed for improving adaptive management strategies in the water supply industry.