The original article proposed a sampling method for detecting continuous
contamination sources, a method that does not require elaborate water
quality data or real-time sensors. The method has practical merit for
many water utilities but the authors’ work has some limitations not
discussed in the article that we comment on here. First, the authors use
terms like “real-world system” to refer to a study done with a model,
which may mislead readers. Second, the method is advertised as simple
but actually requires some advanced tools and steps that are not fully
described and that may not be available to practitioners. Third,
determination of flow directions is not always explicit and usually
requires a hydraulic model, which the authors happened to have, but may
not be available for every system. Fourth, the authors do not adequately
describe the hydraulic model; their figures show networks that differ
from the provided models and may have differing lengths or flow
directions that affect the analysis. Finally, the study data are not
provided as per the Data Availability Statement, but doing so could help
overcome these limitations and make the method more useful.