Online-monitoring Cyclotella sp. concentrations in lake water using an automated fluorescence microscopy method†

Abstract

Algal proliferation in source water presents significant challenges to drinking water utilities owing to the release of taste and odor compounds and algal organic matter associated with the formation of disinfection byproducts. Cyclotella sp., a diatom, can cause undesirable fishy or grassy odors, and its algal organic matter can be a precursor of toxic disinfection byproducts in drinking water. This study aimed to establish an automated method for determining the Cyclotella sp. concentration in surface water. First, an auto-counting method was developed to quantify specific algal concentrations among many other algae based on autofluorescence intensity and morphological characteristics such as dimensions and circularity. The auto-counting of Cyclotella sp. was accurate, with an overall accuracy of 98% and an F1 score of 96%, and the automated and manual counts of Cyclotella sp. were highly correlated (r_s = 0.989). Second, an automated online monitoring system, which can be installed along with a water-containing pipeline and provide concentration measurements within 20 min, was developed by integrating the time-lapse capture module of a fluorescence microscope with a flow cell and pump. Changes in Cyclotella sp. concentrations between 170 and 1661 cells per mL were successfully monitored for 42 h without personnel intervention. The automated method developed in this study can detect rapid changes in algal communities and allow water utilities to take immediate countermeasures against them.