Measuring the effectiveness of an integrated algae bloom, T&O, and cyanotoxin monitoring program Open Access

Water quality aesthetics are usually not associated with health risks, so they are often viewed as being of little importance. However, aesthetic issues, i.e., discoloration, particulates, tastes and odors, etc., can quickly undermine consumer confidence in a water provider. Consumers relate poor aesthetics to poor overall quality, so it is important for water providers to proactively monitor water quality and address issues as quickly and efficiently as possible. This article will describe an integrated monitoring program developed by the City of Wichita Falls (CWF) Cypress Environmental Laboratory (CEL) for algal blooms, taste and odor (T&O) compounds, and cyanotoxins, and discuss the program's effectiveness over the last 7 years.

The importance of aesthetic T&O issues in the United States (US) had a slow beginning. The first record of a T&O event in the US comes from Boston, Massachusetts in 1854, and was described as a ‘peculiar cucumber odor’ (Horsford & Jackson 1855). Methods for water quality analysis were not standardized until the Report of Committee on Standard Methods of Water Analysis (1905) was published, which recognized that T&O issues could be caused by both ‘microorganisms and algae,’ and stated ‘the odors are usually connected with some organic growths or with sewage contamination, or both.’ Interest grew in the 1940s–1950s, and hundreds of works have been published, including early understanding of actinomycetes as a biogenic source of T&O (Silvey et al. 1950), laboratory culture of T&O-producers (Silvey & Roach 1959), and early gas chromatography (GC) studies to detect T&O compounds (Silvey et al. 1968).

A turning point occurred in the 1980s when national (within the US) and international collaboration began with funding from the Water Research Foundation (WRF) to develop reproducible methods to describe and monitor the T&O quality of drinking water, and to find better ways to manage and control drinking water flavor (Burlingame et al. 2022b). Since that time, international efforts through the WRF, the American Water Works Association (AWWA), and the International Water Association (IWA) have continued to support researchers and utilities. This work ranges from the development of closed-loop stripping with gas chromatography-mass spectrometry (CLSA GC-MS) (McGuire et al. 1981), understanding the T&O of chlorine and chloramine species (Krasner & Barrett 1984), the development of the Flavor Profile Method at the Metropolitan Water District of Southern California (Krasner et al. 1985), development of the T&O Wheel (AWWARF 1987), the development of solid-phase microextraction (SPME) for volatile organics in the early 1990s (Weggler et al. 2020), advances in treatment of T&O (AWWARF 1995), identification of decaying/septic T&O compounds (Khiari et al. 1997), understanding T&O in distribution systems (Khiari et al. 2002), and the addition of SPME to the 21st edition of Standard Methods in 2005, as SM 6040D.

Much of this foundational work can be found reviewed by Burlingame et al. (2022a). Recent work moves beyond questions of identification and treatment, to understand how T&O should fit into United States Environmental Protection Agency (USEPA) Secondary Maximum Contaminant Levels (SMCLs) (Dietrich & Burlingame 2015), how humans can act as sensors for early detection of T&O compounds (Burlingame et al. 2017), and how utilities can use this knowledge to proactively address customer complaints (Dietrich & Burlingame 2021; Adams et al. 2023b).

The CWF, TX (www.wichitafallstx.gov) has a surface water system that consists of four surface water sources, including Lake Arrowhead, Lake Kickapoo, and the Lake Kemp/Lake Diversion system, two water treatment facilities, including three conventional and one advanced treatment plant at Cypress Water Treatment Facility (CWTF), and two conventional treatment plants at Jasper Water Treatment Facility (JWTF), and its water distribution system. Historically, the system has experienced regular T&O cycles in surface water, but little was implemented to detect and treat events until after 2016. In response to a historic drought, the CWF built the world's largest direct potable reuse (DPR) system, reusing the city's wastewater effluent (with further treatment at the CWTP) for 50% of its drinking water supply. The DPR was operated from July 2014 to July 2015, and was decommissioned in 2015 after historic rainfall (Nix et al. 2020). No T&O complaints were received during the DPR operation.

The water system experienced two extreme T&O events in February and August 2016 when the CWF was back to relying solely on surface water. Hundreds of customer complaints were received in a two-week period in August 2016. To mitigate this problem, the CEL designed and implemented a comprehensive monitoring program using sondes, traditional threshold odor number/flavor profile analysis (TON/FPA) testing, a single quadrupole GC-MS system, flow-imaging microscopy (FIM), and quantitative polymerase chain reaction (qPCR) (Adams et al. 2018). Existing water system oxidation and adsorption processes were evaluated and maximized to increase effectiveness. This program has been in place for over 7 years and has detected and mitigated 16 T&O events.

CEL staff had been performing basic T&O-related analyses for years, but they did not provide a complete picture of the overall water quality of the surface water. Analyses included TON testing (SM 2150B) and algae identification and enumeration by light microscopy. To develop a more robust program, CEL needed to answer the following:

• Are algae or cyanobacteria present in source water?

• What quantity of algae or cyanobacteria are present?

• Can the organisms present produce T&O compounds or cyanotoxins?

• What is the concentration of T&O compounds or cyanotoxins?

• What is the trigger to know if there is a problem and how is it mitigated?

To gather real-time data, YSI EXO1 multi-parameter sondes are used to monitor temperature, dissolved oxygen (DO), pH, chlorophyll a, and phycocyanin pigments. Lake intakes are profiled monthly, at every foot of depth, to determine if the reservoir is stratifying. Pigment concentrations are used to determine the likelihood of higher cell counts of cyanobacteria vs. algae, because while both produce chlorophyll a, it is mostly cyanobacteria that produce phycocyanin, e.g., exceptions being cryptomonads and red algae (Adams et al. 2022a). Trending the diurnal variations indicates the extent of biological activity in the water, because pH and DO can increase as photosynthesis outpaces cellular respiration, producing more DO and decreasing available dissolved carbon dioxide (CO₂), which serves as a good indicator of blooms (Smith 2019; Adams et al. 2022b).

Just as no single analytical chemistry method can detect all chemical contaminants in water under any given set of conditions, no single sensory method can provide all the answers to T&O questions. Sensory analysis often detects the presence of T&O compounds at lower concentrations than can be detected by analytical instrumentation, e.g., many halogenated anisoles have an odor threshold concentration (OTC) of <1 ng/L, while they will not be detected until around 5 ng/L by single quadrupole GC-MS. TON (SM 2150B) testing is performed to determine the magnitude of a T&O event, while a modified FPA (SM 2170) is performed to determine what type of odor is present (SM 2022). When T&O events occur and the compound/source is not known, samples should be collected throughout the treatment process, from source, through each unit process in the treatment plant, and into the distribution system. Samples can be screened by the following sensory method to determine where the T&O originates, which provides direction on how to mitigate the event even if the compound is not identified.

Samples are dechlorinated with 200 μL of a 5,000 mg/L ascorbic acid (C₆H₈O₆) stock solution to 100 mL of sample. Samples are stirred and heated to 60 °C, and then the odor number and odor profile is determined. The odor profile uses the T&O wheel, first described by Mallevialle & Suffet (1987), and updated by Lin et al. (2019). Categories include Earthy/Musty/Moldy (E/M/M), Grassy/Hay/Straw/Woody (G/H/S/W), Fishy/Rancid (F/R), Marshy/Septic/Swampy/Sulfurous (M/S/S/S), Medicinal/Phenolic (M/P), and Chemical/Hydrocarbon (C/H). This allows technicians to know which GC-MS method to run to determine T&O compound concentration. Sensory methods are reliable, essential when used correctly, and can serve as an early warning before concentrations are high enough for analytical confirmation.

Orion meters are used to monitor pH, free ammonia, and conductivity in the laboratory. KEM auto-titrators are also used to monitor alkalinity and hardness. These parameters help build a historical baseline of water quality so that irregularities can be quickly identified and investigated (Buerkens et al. 2020b).

The traditional method of algae/cyanobacteria identification and enumeration by light microscopy is time-consuming and not always a good option for small laboratories. To reduce analysis time, CEL implemented FIM and purchased a Fluid Imaging Technologies FlowCam (www.fluidimaging.com). It is used to image algae and cyanobacteria, and then sort them into groups, i.e., T&O-producers, cyanotoxin-producers, filter-cloggers, etc. Knowing which taxa are present enables CEL to determine whether a T&O or cyanotoxin event would be likely. Chemical analysis can indicate if there is a current problem, i.e., T&O or cyanotoxin compounds already present in the water column, while identification of organisms indicates that heightened monitoring is prudent even if the compounds have not been detected.

The presence of a cyanobacteria bloom does not mean that the organisms are toxic or producing T&O compounds. However, the absence of toxins or T&O compounds does not mean that a problem is not emergent (Buerkens et al. 2020a). Intracellular compounds can be present, but not readily detected until they are released during normal growth or senescence. Their undetected presence does not absolve the water utility of a potential future problem. When a T&O event begins, it is important for quality data to be systematically collected to give guidance for effective mitigation. For this reason, more utilities are moving toward FIM and successfully incorporating it into their monitoring programs (McKay et al. 2019; Barrowman et al. 2023, 2024). This technology requires a large capital investment, but its ease of use and time savings prove its justification for implementation by water systems, and is useful in other applications such as particle analysis for filter studies and distribution system investigations (Adams et al. 2023a).

Actinomycetes, fungal-like bacteria, are also monitored because they can produce many different T&O compounds, including geosmin, 2-methylisoborneol (MIB), pyrazines, and halogenated anisoles. They are near-ubiquitous in the environment, occurring in soils, are present in surface water, and can colonize filter media in water treatment plants and water distribution systems (Adams et al. 2021a). Analysis is performed by single-layer agar culturing using SM 9250B. Culturing can take up to 7 days, but provides an indicator whether they were the source of a T&O event.

T&O compounds are detected and quantified by selected ion monitoring (SIM) using a Thermo Scientific (Waltham, MA, www.thermofisher.com) Trace 1310 ISQ LT GC-MS with a TraceGOLD-5MS column, TraceGOLD-1MS column, or TraceGOLD-WaxMS column, with sample injection by a Supelco SPME fiber or by headspace (HS). A Thermo Scientific TriPlus RSH autosampler is used with a three-position sample tool changer. This injection is performed by RSH sample tool 1–3 to minimize instrument reconfiguration by the analyst. High purity standards and SPME fibers are sourced from Sigma Aldrich (St. Louis, MO, www.sigmaaldrich.com). Method parameters are based on SM 6040D and are discussed in detail by Adams et al. (2020). A study of 18 T&O compounds from several T&O categories was performed by Pochiraju et al. (2021), and provides new information on preservation and hold times. It should be noted that single quadrupole GC-MS is limited to a low (∼5 ng/L) detection range for many T&O compounds, but some have OTCs lower than analytical detection ranges. For this reason, absence of T&O compound detection during an event does not necessarily mean that the compound is not present, only that it is not present at the detection threshold. If the OTC is lower, then sensory analyses must be relied on. Triple quadrupole, while less common in utility laboratories, can achieve detection limits several orders of magnitude lower, as demonstrated by Pochiraju et al. 2021.

SIM is a mode in which a limited mass-to-charge ratio range is transmitted and detected, resulting in a higher selectivity and sensitivity. Sodium chloride is added to samples and dissolved with heat to increase ionization potential and drive the volatile organic compounds and semi-volatile organic compounds (VOCs/SVOCs) from the aqueous phase to the gaseous phase. SPME is a sample injection method in which target compounds adsorb onto a SPME fiber in the headspace of a closed sample vial, which is then immediately inserted into the heated injection port of a split/splitless (SSL) injector. Compounds then desorb from the SPME fiber onto the GC column. After separation in the GC column, target compounds are detected in the MS by monitoring the largest of multiple target ion masses for confirmation, and quantified.

The most widespread and problematic compound category for water systems is the E/M/M category. It includes two terpenoids, geosmin and MIB. Also included are 2-isopropyl 3-methoxypyrazine (IPMP), and halogenated anisoles, i.e., 2,4,6-trichloranisole (2,4,6-TCA). Compounds in this category are produced by a wide variety of taxa, including algae, cyanobacteria, fungi, bryophytes, protozoans, and some plants. This is currently the only category included in Standard Methods (SM 2022). The G/H/S/W category includes semiochemicals like esters, alcohols, and fragrant apocarotenoids. The Fishy/Rancid category consists of aldehydes and amines that are produced by taxa with high cellular polyunsaturated fatty acids.

The M/S/S/S category is rarely produced by cyanobacteria during normal conditions, but they are major constituents produced in anaerobic decomposition of harmful algal blooms (HABs) and other organic material (Watson & Jüttner 2017). Surface waters heavily impacted by agriculture and wastewater input realize these compounds in abundance. The M/P category is not one that is a common source of concern for surface water systems. The odors produced are commonly described as pesticides, herbicides, and disinfectants. The Fragrant/Vegetable/Fruity/Flower (F/V/F/F) category is comprised of alkylphenols and aldehydes. These compounds are not typically nuisance compounds but can be disinfection byproducts (DBPs), e.g., produced by ozonation.

The C/H category also includes compounds not normally found in surface waters, but they can be present due to spills or the introduction of sewage. This category includes compounds like methyl tert-butyl ether (MTBE), benzene, toluene, ethylbenzene, xylenes, styrene (BTEX), and others. They are easily monitored by labs utilizing EPA 524.3, which outlines the detection of purgeable organics by SIM using Purge & Trap (P&T) and single quadrupole GC-MS with a drinking water VOC column. Like the previous category, the Chlorinous/Bleachy (C/B) category includes compounds not normally found in surface waters. This class consists of constituents used as disinfectants in the treatment of surface water such as free chlorine, monochloramine, dichloramine, and ozone, so it is already understood that they will be present to some degree in tap water. They are added/formed in the disinfection process, and therefore labs should already have the capability to analyze this class of moiety by colorimetric testing or amperometric titration.

Since many of the region's most problematic cyanobacteria are known to produce cyanotoxins, qPCR was incorporated using Phytoxigene CyanoDTec (www.phytoxigene.com) molecular-based assays to determine presence and abundance of cyanobacteria and cyanotoxin-producing genes. The Phytoxigene test can quantify both the amount of overall cyanobacteria present in a water sample as well as the number of genes that are responsible for the production of the cyanotoxins. There are several classes of cyanotoxins associated with cyanobacteria, but most are either hepatotoxins (causing damage to the liver) or neurotoxins (causing neurological damage). The hepatotoxins include microcystin, nodularin and cylindrospermopsin, while saxitoxin is the primary neurotoxin produced by cyanobacteria. The assay's target genes are the 16S rRNA gene for total cyanobacteria presence, mcyE/ndaF for microcystin/nodularin, cyrA for cylindrospermopsin, and sxtA for saxitoxin. An assay for anatoxin was added in 2023.

Presence of genes does not always correlate to toxin production, but does indicate that the genetic sequence is present for production. For this reason, toxin confirmation is necessary for any positive samples. Positive qPCR detections are followed up by analytical confirmation with a third-party lab by liquid chromatography (LC-MS/MS) using EPA 544, 545, 546, and the newly released SM 10110, and have shown direct correlations with low-level detection qPCR results (Adams et al. 2021b). qPCR can be a relatively inexpensive option for water systems to use to monitor source water for the potential of toxin-producing genes.

Due to the expense of an LC-MS/MS system, CEL opted to implement a fluorometric planar wavelength assay in 2022 to perform confirmatory testing in-house. A LightDeck Mini (www.lightdeckdx.com) was purchased, that performs an assay in cartridges in which a laser-excited dye binds to antibodies. Fluorescence signals are captured only if the antibody has reacted with the target cyanotoxin (Bickman et al. 2018). Currently cartridges are only available for microcystins/nodularin and cylinderospermopsin. Saxitoxin and anatoxin-a cartridges are anticipated to be released in 2024. The system is an inexpensive option for water systems for rapid proactive monitoring.

Water systems must remove algae and their metabolites from the source water while meeting all state and federal treatment regulations. This can be done early in the process by aerating the water near reservoir intakes, alternating intakes, adjusting the pH to deter pH-sensitive organisms, and by the use of algaecides such as copper sulfate (CuSO₄). Most systems incorporate a multi-barrier approach, where reservoir management strategies previously listed are used in conjunction with physical pretreatment, physical removal, conventional treatment, biological treatment, oxidation, and/or adsorption. CWF has two sources for conventional treatment, so sources can be switched to pull from the lake not experiencing a T&O event. However, water systems should take care to understand how source switching and mixing may affect and even exacerbate T&O events (Zhu et al. 2023).

Chemical oxidation is a process used to remove algae and their metabolites. Potassium permanganate (KMnO₄) is used early in treatment processes to maximize contact time, and is widely used as a pre-oxidant. KMnO₄, unlike many other oxidants, produces little to no DBPs. One advantage of KMnO₄ is that it helps to demobilize algal cells, which helps them settle out in the treatment process prior to contact with a disinfectant and prevents cellular lysis and the release of intracellular compounds into the water. However, it is not as effective as others at removing geosmin and MIB. Chlorine dioxide (ClO₂) is a pre-oxidant that is more effective at removing geosmin and MIB than KMnO₄, but it is unstable and can be hazardous and is usually generated on-site. Care must also be taken to dose appropriately so that chlorite, one of the products of ClO₂ degradation, does not exceed the EPA MCL of 1.0 mg/L in the tap water. Pochiraju et al. (2023) evaluated the treatability of 18 T&O compounds using oxidants (Cl₂, ClO₂, and ozone), which can serve as a starting point for utilities once the T&O compound is known.

CWF utilizes an adsorption process, which is a highly effective treatment method that using powdered activated carbon (PAC). PAC is a fine carbon powder that is fed early in the treatment process. Organic compounds are adsorbed onto the carbon particles and settle out as they become heavier during coagulation. Overfeeding PAC can turn clarifier effluent black and can potentially clog conventional filters. PAC is also commonly used with KMnO₄. Several types of PAC are marketed, so they should be bench tested to determine which is the most effective for the water system's source water. PAC dosage can be expensive, but an advantage is that it can be increased or decreased as needed. Pochiraju et al. (2022) evaluated the treatability of 18 T&O compounds using PAC, which can serve as a starting point for utilities once the T&O concentration is known.

In the event that CWF has maximized all treatment capability and T&O breakthrough into the distribution system is still experienced, a template press release was developed to alert the news media. This is triggered by Public Utilities Administration and is a proactive measure to educate the public on the T&O event (including extent and anticipated duration), and what measures are being taken by the CWF to mitigate the event. The goal is to maintain consumer confidence by proactive communication.

The majority of T&O events detected by CEL have been biogenic, i.e., produced by living organisms. The primary biogenic producer has been cyanobacteria, but actinomycetes have also produced events for CWF, along with three unknown events and a non-biogenic source. Zhu et al. (2022) provides a comprehensive review of the types of biogenic events that have been reported and is an excellent source of information for utilities building a program.

The greatest concentration of geosmin detected in source was 666 ng/L, while the greatest concentration in tap was 164 ng/L. This speaks to the effectiveness of early detection, followed by mitigation measures including source switching, oxidation, and adsorption in the treatment plants. Average concentrations were <20 ng/L for both source and tap, indicating that even when geosmin is detected, it is usually slightly over the OTC of 10 ng/L. It is also important to understand that OTCs are determined to be the point at which half of a sample population can detect the compound by sensory analysis, and half cannot. This means that even when the OTC of 10 ng/L is hit, half the population should still not be able to detect the compound.

Table 1 shows each T&O event detected by CEL, by year and month(s) of duration. The T&O compound is listed if identified by CEL, along with the detection method(s) for the T&O event, maximum source concentration reached, T&O source, organism count (if applicable), mitigation strategies undertaken by CWF treatment plants, maximum tap concentration reached, and the number of customer complaints received. Organism count (#/mL) is used by CEL as cells if found individually, colonies, and/or chains/filaments. These may be interpreted in various ways depending on type of information needed. There were three events produced by an unknown source (all in 2022), one event produced by elevated salinity (TDS) one event produced by actinomycetes in filter media, one event produced by the chlorine dioxide generation process, and the remainder produced by algae and cyanobacteria. The following subsections discuss each type of T&O event.

Most T&O events detected by CEL have been HAB-related, i.e., produced by algae and/or cyanobacteria. The largest events occurred in 2019 and 2020, which were both warm dry years, and consisted of geosmin produced primarily by Dolichospermum (formerly Anabaena), with Aphanizomenon and Microcystis also in occurrence. Detection methods included field monitoring, sensory analysis, FIM, and GC-MS. In each instance, CWF switched sources to mitigate the event, along with CuSO₄ treatment at the lake and PAC at the treatment plants (for two events). In two instances, customer complaints were received, but treatment was maxed out, and the raw water transmission main was broken in once instance, preventing the use of the reservoir not experiencing a bloom. Both events were described in detail by Adams et al. (2021b, 2022a).

In February 2021, plants were cycled for winter maintenance, and a localized event occurred at CWTF. The T&O compound was unknown, but was detected by sensory analysis two weeks prior to confirmation by GC-MS due to its low OTC. The compound was confirmed as 2,4,6-TCA once the analytical detection threshold (∼5 ng/L) was exceeded. It was determined that actinomycetes had colonized filter media in one plant, and were converting 2,4,6-trichlorophenol to 2,4,6-TCA in the media. Filters were taken offline and oxidized with KMnO₄, which eliminated the event. A full description can be found in Adams et al. (2021a). This event provided a strong example that sensory analysis cannot be removed from a robust T&O program, because it can allow the detection of an event before it can be detected analytically.

In February 2022, an event described as ‘fishy’ was detected by sensory analysis only. It resulted in a single complaint, and the source of the event is still unknown. This is another example of sensory analysis detecting T&O before analytical detection.

CWTF has one microfiltration/reverse osmosis (MFRO) plant, which blends permeate with conventionally treated water. In 2022, the RO elements were nearing their end of life, and CEL staff detected an increase in TDS and conductivity in the overall blend leaving CWTF. Under normal conditions, the salinity produced by CWTF is less than that produced by JWTF, but the RO membrane rejection had degraded to a point that CWTF was producing water with greater salinity that JWTF. Normal concentrations were around 250 mg/L, but had increased to a maximum of 442 mg/L at CWTF. This resulted in plans for the MFRO plant to be taken offline while the new membranes were in transit. This small example supported the findings of McGuire & Pearthree (2020), who reported fast changes in TDS could be detected by consumers and result in complaints.

In January 2022 and 2023, a musty odor was noted in the tap water at both WTFs. The T&O compound was unknown, but was detected by sensory analysis. It was similar to the moldy odor from the 2021 filter event (refer Section 3.3), so other haloanisoles were investigated. The compound was confirmed as 2,4,6-tribromoanisole. It was determined that this compound was being produced in raw water lines post-ClO₂ injection, due to the reaction of excess free chlorine with naturally occurring bromide and humic anisoles in the raw water. Jar tests were conducted to confirm this method of production, which was described in detail by Adams et al. (2023c).

Customer complaints can be viewed negatively, or as opportunities for water providers to get quick feedback on water quality and take proactive measures in the future to prevent similar complaints. It is important for water providers to establish water quality goals for aesthetics, such as T&O (Burlingame 2012). To accomplish this, it is important for providers and customers to understand the basics of tap water taste and how to communicate with providers when issues arise (Burlingame et al. 2007; Carniero et al. 2021). Providers must systematically collect complaint information, maintain records, and follow up with consumers to ensure water quality has returned to normal and they are pleased with the provider.

The CEL workflow has become more robust and more effective over the years as its monitoring program has been refined. New data allow staff to adjust triggers and more is understood about the water quality of CWF's reservoirs. These data show that the program has effectively detected T&O events early, allowing for proactive treatment measures to eliminate the event at the source and/or mitigate the event at the treatment plant. Sensory analyses must be used in tandem with analytical chemistry techniques, because some T&O compounds have OTCs lower than detection thresholds. In these cases, non-detect by analytical techniques only indicates that the T&O compound is not present above the detection threshold, and cannot conclusively indicate the complete absence of the T&O compound.

Additional technologies, such as qPCR and fluorometry have allowed CEL to analyze cyanotoxin-producing genes and cyanotoxins in-house, for faster turnaround times and decision-making. Future needs include methods for the analysis of multiple T&O compounds, beyond geosmin and MIB, on single quadrupole GC-MS because most utilities do not have the funds or expertise needed for triple quadrupole analysis. This research group is currently working to meet this need and has formed a Joint Task Group to expand SM 6040D and evaluate a suite of known T&O compounds that can be problematic for water systems, using a single method for 20 T&O compounds by GC-MS. A deeper understanding of biogenic sources will help utilities develop stronger monitoring plans can also be used to detect non-biogenic sources, as demonstrated by CEL.

H.A., M.S., and D.N. conceptualized the study, wrote, reviewed, and edited the article. S.R. and E.A. wrote, reviewed, and edited the article.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.