Determining vulnerability and occurrence of residential backflow

Backflow from customer service connections is of concern to water utilities, but few data about the frequency of such occurrences exist and little is known about how to monitor for backflow into the distribution system. For the first time, occurrence statistics for residential backflow have been quantified. As it turns out, backflow is a much more widespread problem than currently believed. Backflow has now been shown to occur in 1.6% of all meter reads and in 5% of homes with backflow‐sensing meters. Using pilot tests to examine the effectiveness of water quality sensors to detect a range of contaminants, the authors found that even though the monitors were sensitive to small changes in quality, they cannot be used to assess the public health effects of such events. This is because wide variations in background distribution system quality overwhelmed the ability of current data processing technology to differentiate between water quality fluctuations caused by backflow from those attributable to normal variability. Backflow‐sensing meters, combined with an effective pressure management program to minimize or eliminate pressure transients (a primary cause of backsiphonage cross‐connections) and integrated into a fixed network automated meter reading/advanced metering infrastructure system, can provide utilities with near real‐time information on backflow events. The analyses conducted in this study show that the use of backflow‐sensing meters is the best available technology for determining the occurrence of residential backflow. Given the benefits shown by these meters, their slightly higher cost ($15‐$20/meter) over regular meters is money well spent.