1. INTRODUCTION

Cyanobacteria are natural components of freshwater and brackish ecosystems, but under nutrient enrichment, warming, altered hydrology, longer residence time, and water-column stability they can form harmful blooms (Paerl and Otten 2013;Huisman et al. 2018;Chung et al. 2019). These blooms can impair drinking-water resources, recreational waters, irrigation supplies, shoreline environments, and aquatic ecosystem services, thereby requiring public-health and management responses (Chorus and Welker 2021;WHO 2021;Health Canada 2022;Villalobos et al. 2025). Their expansion and persistence are further shaped by Microcystis ecology, climate-sensitive bloom dynamics, and nutrient enrichment across freshwater-to-marine systems (Harke et al. 2016;Wurtsbaugh et al. 2019;Paerl and Barnard 2020). Therefore, CyanoHABs should be understood as socio-ecological disturbances driven by interacting climatic, nutrient, hydrological, biological, and institutional factors rather than as isolated analytical anomalies.

Monitoring practice has often centered on how to count cyanobacteria more accurately. That question remains important. Direct microscopy provides taxonomic information and a defensible reference layer; biovolume links abundance to biomass; phycocyanin and chlorophyll-a (Chl-a) fluorescence provide rapid signals; imaging flow cytometry and FlowCam-like approaches increase throughput and archive interpretable images; quantitative PCR (qPCR) and droplet digital PCR (ddPCR) distinguish total from toxigenic populations; and enzyme-linked immunosorbent assay (ELISA), liquid chromatography-tandem mass spectrometry (LC-MS/MS), autonomous toxin sensing, and related approaches connect monitoring to toxin exposure and toxigenic potential (Bustin et al. 2009;Zamyadi et al. 2016;Bertone et al. 2018;Dunker et al. 2018;Menden-Deuer et al. 2020;The dMIQE Group and Huggett 2020;Thomson-Laing et al. 2020;Almuhtaram et al. 2021a, 2021b;Feist and Lance 2021;T-Krasznai et al. 2022;Lim et al. 2023;Eerola et al. 2024;Ussler et al. 2024;Borisova et al. 2025;Janatian et al. 2025;Kim and Park 2025;Kraft et al. 2025;Madany et al. 2025). Yet a field event raises a prior question: what decision must be made, by whom, and by when?

This distinction is operational, not semantic. A beach manager facing visible scum, a drinking-water operator observing an intake anomaly, a public-health officer considering a swimming advisory, an irrigation manager evaluating pump placement, and an analyst calibrating an imaging classifier are not asking the same question. A single cells mL^-1 value cannot determine whether children should avoid contact today, whether a water-treatment plant should modify operations, whether aerosol exposure is plausible during high-wind recreation, or whether spray irrigation should be delayed. International guidance has emphasized that cyanobacterial risk management should combine cyanotoxin values with biomass indicators, visible scum observations, exposure-route assessment, and public communication. World Health Organization (WHO) guidance provides the broad public-health and recreational-water framework (Chorus and Welker 2021;WHO 2021), while U.S. Environmental Protection Agency (U.S. EPA) and Health Canada documents specify advisory logic and indicators for recreational exposure (USEPA 2019;Health Canada 2022). Comparative reviews further show that cyanotoxin risk-assessment and management approaches differ among countries, supporting the need for locally adapted implementation (Ibelings et al. 2014).

Enumeration error is especially consequential for colony-forming taxa such as Microcystis. T-Krasznai et al. (2022) showed that cyanobacterial colonies cannot be reliably approximated as ellipsoids and that even significant colony-volume relationships can generate biased cell-count estimates because colony compactness and hidden cells vary substantially. This finding is methodological, but its implication is managerial: an apparently precise cells mL^-1 value may carry unreported uncertainty large enough to affect risk interpretation, resampling priority, public messaging, or release decisions.

This manuscript is therefore a decision-oriented critical review and operational framework paper. It is not a systematic review, a meta-analysis, a new legal threshold system, or a claim that rapid proxies can replace reference microscopy or toxin measurements. Its contribution is to specify how imperfect evidence should be staged across decision windows so that precaution, confirmation, revision, release, and learning remain logically distinct. The aim is not perfect precision before action, but sufficient credibility to prevent avoidable exposure while later evidence refines the decision.

Rather than attempting to conceptually reorganize or replace established global guidelines (e.g., WHO, U.S. EPA, or Health Canada), this framework is positioned strictly as an operational supplement. It addresses the systemic lag between rapid field dynamics and definitive laboratory reporting, providing clear operational decision pathways tailored for field biologists, local managers, and small-scale utilities who must justify immediate defensive actions before comprehensive results are finalized.

In this review, reference-calibrated means that rapid or proxy evidence is interpreted against documented microscopy, biovolume, image records, toxin assays, molecular assays, or locally validated field protocols, rather than treated as a stand-alone definitive abundance or health-risk estimate. This boundary is essential because the framework is strongest for colony-forming and Microcystis-dominated CyanoHAB settings, and it requires local adaptation for filamentous, picocyanobacterial, marine, or mixed-species bloom regimes.

2. REVIEW APPROACH AND EVIDENCE SYNTHESIS

This is a narrative and scoping review rather than a formal meta-analysis. We integrated three evidence streams. The first evidence stream included method-focused studies published primarily during 2018-2025, covering colony enumeration and multi-instrument plankton assessment (Menden-Deuer et al. 2020;T-Krasznai et al. 2022), image-assisted and computer-vision approaches (Dunker et al. 2018;Eerola et al. 2024;Borisova et al. 2025;Kraft et al. 2025), phycocyanin fluorescence and early-warning tools (Zamyadi et al. 2016;Bertone et al. 2018;Thomson-Laing et al. 2020;Almuhtaram et al. 2021a, 2021b), satellite bloom monitoring (Janatian et al. 2025), molecular assays for total and toxigenic populations (Bustin et al. 2009;The dMIQE Group and Huggett 2020;Feist and Lance 2021;Kim and Park 2025;Madany et al. 2025), and toxin-oriented measurement, including autonomous toxin sensing (Ussler et al. 2024). The second evidence stream included public-health and management guidance from WHO (Chorus and Welker 2021;WHO 2021), the U.S. EPA (USEPA 2019), Health Canada (Health Canada 2022), and comparative international risk-management reviews (Ibelings et al. 2014). The third provides contextual support on bloom drivers, Microcystis ecology, nutrient enrichment, and climate-sensitive bloom dynamics (Harke et al. 2016;Wurtsbaugh et al. 2019;Paerl and Barnard 2020).

Evidence was selected for its role in decision support rather than for technological novelty alone. We prioritized sources that directly addressed at least one of four functions: (i) minimum evidence for precautionary action, (ii) reference-calibrated enumeration or biovolume interpretation, (iii) toxin or toxigenic-potential confirmation, and (iv) exposure-route-specific management or communication. Papers focused solely on general bloom ecology, analytical chemistry, or image classification were excluded unless they contributed directly to decision roles, uncertainty pathways, or reporting requirements. This approach intentionally favors interpretive usefulness over exhaustive coverage.

Papers were not ranked simply by methodological sophistication. Each method was evaluated against five action-relevant criteria: the decision window it can support, field feasibility, exposure-route relevance, reference-calibration need, and ability to communicate uncertainty to non-specialist users. This evaluation separates three roles that are often conflated: a trigger signal that justifies temporary precaution, a reference measurement that anchors taxon-specific interpretation, and a confirmation measurement that evaluates toxin-related health risk.

Because the objective is a practical contribution rather than a technology inventory, the main manuscript is intentionally lean. It uses three figures, three tables, and three boxes. Detailed literature mapping, minimum reporting items, and a one-page field action card are provided in the supplementary material. This structure keeps the main paper focused on a clear conceptual and operational argument while making the supporting evidence auditable and reusable.

The synthesis was organized to carry these four functions through the manuscript. Figure 1, Table 2a, and Table 2b address the first function by defining minimum evidence and international operational precedents for precautionary action within specific decision windows. Figure 2 and Table 3 address the second function by identifying sampling, colony-handling, and reporting controls required for reference-calibrated enumeration and biovolume interpretation. Toxin and genetic assays are treated as confirmatory evidence for health-risk interpretation, particularly when biomass indicators and toxin risk diverge. Figure 3 then addresses the fourth function by mapping evidence needs to major exposure pathways, including recreation, drinking-water protection, irrigation, and aerosol or occupational exposure.

3. RESULTS AND SYNTHESIS

3.1. Decision-ready evidence architecture: method roles and decision windows

Figure 1 summarizes the action-first structure proposed in this review. The framework starts with three practical questions: what decision is needed, how quickly it must be made, and what minimum evidence can support that decision. In a CyanoHAB event, rapid field evidence may be more useful for immediate precaution than a precise laboratory result that arrives too late. Therefore, the key issue is not which method is analytically superior in general, but which evidence is sufficient for a defensible action within the relevant decision window.

This distinction becomes clear when different management contexts are compared. If the decision concerns visible scum at a recreational site, the minimum credible evidence is not a full taxonomic count but a defensible combination of field observation, location, exposure context, and precautionary communication. If the decision concerns a drinking-water intake anomaly, the minimum credible evidence includes phycocyanin, chlorophyll-a, or turbidity behavior together with operator-relevant information on intake depth, source-water condition, and treatment status. If the decision concerns public-health confirmation or escalation, microscopy, toxin analysis, and, where appropriate, molecular information become necessary. Thus, method selection is secondary to decision definition.

This architecture is consistent with international and national guidance. WHO guidance emphasizes the combined use of public-health risk assessment, observation, monitoring, and management action (Chorus and Welker 2021;WHO 2021). U.S. EPA and Health Canada guidance further distinguish cyanotoxin values from biomass-related indicators and recognize that visible scums, exposure pathways, and plausible contact risk can support management responses before complete analytical confirmation is available (USEPA 2019;Health Canada 2022). The same staged logic is supported by comparative method studies: microscopy and multi-instrument assessments provide reference and calibration information (Menden-Deuer et al. 2020;Kraft et al. 2025), fluorescence-based tools support early warning and resampling decisions (Zamyadi et al. 2016;Bertone et al. 2018;Thomson-Laing et al. 2020;Almuhtaram et al. 2021a, 2021b), and satellite approaches support spatial prioritization rather than final taxonomic confirmation (Janatian et al. 2025). The practical implication is not that fast methods are inferior, but that their role should be defined correctly.

The key operational move is to stop asking whether a proxy is inaccurate in the abstract. Phycocyanin fluorescence, for example, is not a substitute for taxon-specific microscopy or toxin quantification. Its value lies in being fast enough to support inspection, resampling, and temporary precaution. Used as a trigger, a proxy can be highly defensible; used as a final abundance estimate without calibration or confirmation, it becomes over-extended. Field fluorometry studies show its value for rapid biovolume prediction and early-warning detection (Thomson-Laing et al. 2020;Almuhtaram et al. 2021a, 2021b), while broader reviews emphasize the need for calibration, interference control, and careful interpretation in real-time cyanobacteria monitoring (Zamyadi et al. 2016;Bertone et al. 2018). The same logic applies across the evidence layers shown in Figure 1: rapid evidence supports early protection, whereas confirmatory evidence supports refinement, continuation, narrowing, or release of that decision.

Table 1 translates this architecture into method roles, showing that the practical value of each tool depends less on abstract analytical superiority than on whether it is used for triggering precaution, confirming abundance, prioritizing space, or interpreting health risk.

Methods are organized by decision role rather than by assumed technical superiority. This framing prevents rapid proxies from being treated as final abundance estimates and prevents slow reference methods from becoming the only basis for urgent precautionary action.

A CyanoHAB event is also a time-structured problem: some decisions require immediate protection, while others require reference confirmation, operational adjustment, release logic, or post-event learning. In the first two hours, the practical question is whether visible or sensor-based evidence is sufficient to prevent avoidable exposure. In the next 24 hours, the question is whether rapid screening supports continued caution, resampling, operational review, or targeted field inspection. Within 24-72 hours, reference microscopy, biovolume, toxin screening, and image records should confirm, narrow, expand, or revise the initial response. Over the following week, repeated measurements support release, continued restrictions, or treatment adjustments. Post-event, the relevant task is learning: identifying sampling bias, false positives, false negatives, communication gaps, and protocol weaknesses.

This staged view is consistent with the U.S. EPA recreational criteria framework, which emphasizes magnitude, duration, and frequency for microcystins and cylindrospermopsin swimming advisories (USEPA 2019). It is also consistent with Health Canada guidance, which distinguishes direct cyanotoxin values from indicators of possible cyanotoxin presence, including total cyanobacteria, biovolume, and chlorophyll-a (Health Canada 2022). The practical point is simple: a field manager should not wait for the best possible enumeration result before taking a low-cost exposure-prevention action when visible scum and plausible exposure are present.

The most common error is to treat rapid signals as poor versions of laboratory tests. That framing is counterproductive. Rapid signals are valuable because they are available early enough to influence behavior. Laboratory tests are valuable because they confirm risk, identify cause, and refine future thresholds. A decision system fails in two opposite ways: it waits for confirmation when it should act, or it maintains action without confirmation after later evidence shows low risk.

Table 2a converts measurement evidence into decision windows, while Table 2b summarizes selected international field precedents illustrating how tiered precautionary actions are successfully deployed before complete laboratory confirmation. These frameworks are not intended to replace legal thresholds, national criteria, or agency-specific advisory systems. Its purpose is to distinguish evidence sufficient for precaution, evidence needed for confirmation, and situations in which delayed action or unconfirmed restriction becomes a management error. Any temporary precaution, confirmation step, advisory release, or operational adjustment should be aligned with the locally adopted legal thresholds, responsible agency, and communication protocol before field use.

3.2. Sampling before counting: procedural uncertainty and action-first protocol

Figure 2 shows that enumeration error does not begin at the counting chamber. It begins earlier, with field context, sample definition, sample handling, and the biological architecture of colony-forming cyanobacteria. The central implication is straightforward: a sophisticated analytical pipeline cannot repair a poorly defined sample.

This point becomes especially important when different sample types are considered. Scum, nearshore accumulation, intake water, mixed water-column samples, finished water, and irrigation-point samples represent different exposure questions and management purposes. If these samples are combined, incompletely documented, or ambiguously labeled, later microscopy, sensor signals, toxin assays, and molecular tests may still be analytically correct, yet operationally irrelevant. In practice, the interpretability of a result depends not only on how well the analysis is performed, but also on whether the sample itself corresponds to the management question being asked.

For colony-forming Microcystis, the uncertainty problem begins even before counting. Colony integrity, mixing intensity, fixation, sonication, sedimentation, chamber loading, and image focus all influence the apparent number of cells. T-Krasznai et al. (2022) showed that simple ellipsoid approximation and regression-based colony-volume estimates cannot reliably resolve hidden cells or variable colony compactness. This does not mean that rapid or simplified estimators are inherently unusable. It means that their outputs must be interpreted in relation to how the sample was processed and how colony-related uncertainty was handled. A reported value should therefore state whether it was derived from direct colony-cell estimation, partial disintegration, image-assisted density correction, or volume approximation.

Furthermore, procedural uncertainty is compounded by fixed-sample artifacts and preservation trade-offs. Cell shrinkage caused by standard fixatives (e.g., Lugol’s solution or formalin) frequently distorts subsequent biovolume calculations and compactness assessments. Critically, chemical preservatives required for reference microscopy are structurally incompatible with downstream molecular and toxin assays; formalin destroys DNA templates for qPCR/ddPCR, while severe chemical fixation alters cell matrix integrity, hindering intracellular toxin extraction for ELISA or LC-MS/MS. This absolute incompatibility underscores why field-level sample splitting (Split) before preservation is an uncompromisable prerequisite for reference-calibrated monitoring.

The practical consequence is that the minimum common sampling and processing protocol should be simpler than a full analytical standard, but stricter than routine opportunistic sampling. At minimum, it should document field photographs, location and time, exposure context, separation of scum from water-column samples, split samples for microscopy, toxin analysis, DNA analysis, and sensor or image validation, together with preservation and preprocessing notes. Without this metadata, disagreements among methods cannot be interpreted as scientific uncertainty alone; they often reflect procedural uncertainty. Table 3 translates this logic into a minimal reporting framework for action-ready CyanoHAB assessment.

This protocol is deliberately minimal so that it can be used by field teams, laboratories, water managers, and public-health users before advanced analytical methods are applied. To ensure clear operational communication, the foundational terms utilized across these stages are explicitly defined in Box 2. Detailed checklist items and supporting references are provided in the supplementary material.

The proposed action-first hybrid cyanobacteria protocol converts these procedural requirements into a staging logic for using imperfect evidence responsibly. It is not a new legal threshold system. Its sequence is deliberately simple: observe, separate, split, trigger, confirm, communicate, and learn. Release or revision is treated as part of the confirmation and communication process rather than as a separate analytical method. Each step converts a common field problem into a documented decision point.

Step 1, observe: record visible scum, discoloration, odor, fish or animal mortality, weather, water use, and plausible exposure. Step 2, separate: treat scum, nearshore water, mixed water, intake water, finished water, and irrigation-point water as different samples unless the management question explicitly requires integration. Step 3, split: prepare separate aliquots for microscopy, toxin, DNA, and sensor/image verification before preservation choices make cross-method interpretation impossible. Step 4, trigger: use visible evidence, rapid fluorometry, turbidity anomalies, and local context to justify temporary precaution when exposure is plausible. Step 5, confirm: use microscopy, biovolume, image archives, toxin assays, and genetic markers to determine whether the initial response should be maintained, narrowed, escalated, or released. Step 6, communicate: state which evidence triggered the action, what remains uncertain, and what evidence will be used for release. Step 7, learn: after the event, review false positives, false negatives, field metadata gaps, and threshold performance.

3.3. Toxin-aware and exposure-route-specific interpretation

The same bloom does not create the same decision in every setting. Drinking-water decisions prioritize raw-water and finished-water toxin concentration, treatment performance, intake depth, and the risk of toxin release during cell lysis. Recreational decisions prioritize visible scum, likely ingestion, skin contact, aerosol-generating activities, and public communication. Irrigation decisions prioritize toxin concentration at the irrigation point, spray versus drip application, crop type, and harvest interval. Shoreline and air-exposure decisions prioritize wind, waves, spray, occupational activity, and bloom proximity.

Because these management questions differ by exposure route, Figure 3 maps the evidence pathways needed for recreation, drinking-water protection, irrigation, and aerosol or occupational exposure. The figure illustrates that a visible bloom near a swimming area, a sensor anomaly near a drinking-water intake, scum accumulation at an irrigation pump, and wind-driven spray along a shoreline are not equivalent decision problems. They require different combinations of minimum rapid evidence, confirmatory evidence, and immediate management action.

The aerosol pathway deserves specific attention. Recent work has quantified aerosolized cyanobacterial toxins in the atmosphere and has highlighted inhalation as a plausible exposure route during cyanobacterial bloom events (Lim et al. 2023;Shi et al. 2023). An exploratory study from the Nakdonggang River reported aerosolized cyanobacteria and cyanotoxin-related signatures, making this issue relevant to river-reservoir systems where dense blooms, wind, spray, and shoreline activity can co-occur (Kim et al. 2023). This example should be used as an illustrative case, not as a claim that Korea is uniquely affected.

Agricultural use also expands the definition of relevant evidence. Microcystin in irrigation water can affect plants and create food-chain concerns, and recent reviews and field-oriented studies have emphasized uptake, accumulation, and health-risk implications for crops irrigated with contaminated water (Haida et al. 2024;Faulkner et al. 2025). For this reason, a low cells mL^-1 value in a mixed water-column sample may be insufficient for an irrigation decision if scum accumulates at the pump intake or if spray irrigation creates occupational exposure.

The broader implication is that CyanoHAB monitoring should not force all exposure routes into a single abundance-based interpretation. Cell counts, biovolume, sensors, imaging, toxin assays, and genetic assays become most useful when they are assigned to the exposure question they can actually answer. Recreation, drinking-water protection, irrigation, and aerosol or occupational exposure therefore require a shared monitoring architecture, but not identical evidence thresholds. This exposure-route framing completes the action-first logic developed in Figures 1 and 2: the decision defines the evidence need, the decision window defines the timing, and the exposure route defines the interpretation.

4. DISCUSSION AND IMPLEMENTATION CONSIDERATIONS

International guidance already provides important threshold values, risk-management principles, and public-health context for cyanobacteria and cyanotoxins (Ibelings et al. 2014;USEPA 2019;Chorus and Welker 2021;WHO 2021;Health Canada 2022). This review is therefore not a replacement guideline. Its contribution is narrower and more operational: it explains how enumeration error, sampling design, rapid proxies, toxin-aware confirmation, and exposure-route interpretation can be staged in time and mapped to action. In this sense, the paper addresses the missing operational middle between technical methods and field decisions.

In the context of the South Korean regulatory environment, this framework does not replace the legally mandated Algal Alert System governed by the Ministry of Environment, which dictates response triggers based on strict cell-count thresholds. Instead, our ‘action-first rapid evidence’ serves as a co-management tool. For example, when field operators detect rapid phycocyanin sensor anomalies or visual shoreline scums, these signals can trigger immediate localized public alerts and preventative water-intake shifts during the 48-to-72-hour window required to process official government cell counts.

Moreover, small-scale utilities, local governments, or educational institutions often lack the specialized infrastructure to perform advanced microscopy, ELISA, or ddPCR internally. To resolve this capacity gap, we propose a tiered hub-and-spoke governance protocol. Resource-limited entities focus entirely on the first three steps–Observe, Separate, and Split–and then immediately route the preserved, split aliquots to regional hubs, such as provincial Institutes of Health and Environment, K-water technical support centers, or university research networks. This operational division allows small programs to deploy advanced technical layers dynamically without maintaining prohibitive in-house laboratory costs.

Three claims distinguish this review. First, enumeration error becomes decision error when it delays action, misclassifies exposure, or produces a number that cannot be interpreted by managers. Second, sampling and preprocessing must be placed before method comparison: a poorly defined sample cannot be redeemed by a superior instrument. Third, rapid proxies should not be rejected because they are imperfect; they should be assigned to the trigger layer, where their imperfection is acceptable if later reference and toxin evidence are required for confirmation.

The action-first logic does not imply that visible scum or rapid sensor anomalies always indicate high toxin risk. It means that plausible exposure can justify temporary, low-cost precaution while reference enumeration and toxin-aware evidence are generated to confirm, narrow, maintain, or release the action. This distinction is important because the framework is designed to avoid both delayed protection and unnecessary prolonged restriction.

The strongest scientific implication concerns colony-forming cyanobacteria. The Microcystis problem cannot be solved by declaring a universal correction factor. Colony morphology, cell density, mucilage, preservation, disintegration, chamber loading, and image focus can all affect reported cell abundance and biovolume (Harke et al. 2016;T-Krasznai et al. 2022). However, the management system does not need to wait for perfect colony-cell enumeration before warning the public about visible scum. What it needs is a transparent statement: the initial action was triggered by visible and rapid evidence; the later cell count is reference-calibrated; the toxin result anchors health-risk interpretation; and the uncertainty is documented.

This argument also clarifies the role of advanced methods. Imaging flow cytometry, FlowCam, Imaging FlowCytobot (IFCB), computer vision, ddPCR, LC-MS/MS, autonomous toxin sensing, environmental DNA (eDNA), and automatic image recognition remain essential for research-quality evidence and for improving future models (Bustin et al. 2009;Zamyadi et al. 2016;Bertone et al. 2018;Dunker et al. 2018;Menden-Deuer et al. 2020;The dMIQE Group and Huggett 2020;Thomson-Laing et al. 2020;Almuhtaram et al. 2021a, 2021b;Feist and Lance 2021;Lim et al. 2023;Eerola et al. 2024;Ussler et al. 2024;Borisova et al. 2025;Janatian et al. 2025;Kim and Park 2025;Kraft et al. 2025;Madany et al. 2025). Yet a protocol that depends on these tools at every site and every event is not field-feasible. A national or regional system should start with common field actions and minimum metadata, then add technical layers where capacity exists. This is especially important for local governments, small water utilities, schools, recreational managers, and field educators who need clear action language rather than a list of measurements they cannot perform.

The framework is most directly relevant to recurrent Microcystis-dominated river-reservoir, lake, and nearshore systems where management decisions are often needed before full analytical confirmation is available. Its strongest transferability is to colony-forming CyanoHAB settings in which visible scums, colony-related counting uncertainty, toxin-biomass divergence, and rapid exposure decisions co-occur. It is also transferable to other CyanoHAB settings only if the trigger, reference, and confirmation layers are adapted to local taxa, water uses, laboratory capacity, monitoring infrastructure, and legal thresholds (Kim et al. 2025).

The framework should not be interpreted as lowering evidentiary standards. It separates evidence standards by purpose. Immediate precaution can be justified by credible field evidence and plausible exposure; confirmation requires reference-calibrated enumeration, toxin analysis, or molecular evidence; and release requires repeated observations, declining risk, and documentation of the decision basis. This separation makes the system more transparent, not less rigorous.

Implementing this architecture requires acknowledging that different precautionary actions carry distinct socio-economic and public-health weights. While low-cost actions like increasing sampling frequency or posting advisory signage require minimal trigger evidence, heavy-handed measures such as full recreational closures or water-supply suspensions demand progressively higher evidentiary certainty. Future regional adaptation must explicit link escalating response tiers to calibrated evidence thresholds to prevent economic over-restriction while preserving public health.

4.1. Agency implementation and governance frameworks

Implementation will differ among agencies and monitoring programs. Large utilities or national programs may be able to combine field observation, sensors, satellite products, microscopy, toxin analysis, and molecular assays. Smaller local programs may begin with field documentation, sample separation, basic microscopy, and clear decision logs, then add technical layers as capacity allows. The purpose of the framework is therefore not to prescribe a single universal method package, but to ensure that each method is assigned to a defensible decision role within the relevant decision window.

4.2. Prospective verification and validation agenda

Future applications should test the framework prospectively across recurrent Microcystis-dominated systems using decision logs, response time, false-positive and false-negative review, confirmation results, and release accuracy. Such evaluation should not ask whether every early precaution was later confirmed as high toxin risk. A more appropriate test is whether early actions were documented, proportionate to plausible exposure, revised when stronger evidence arrived, and released when declining risk and confirmatory data supported release.

A practical validation agenda should explicitly track the following four performance endpoints:

These measurable endpoints will allow environmental protection agencies and water utilities to systematically improve local thresholds and optimize decision windows without altering national statutory criteria.

5. CONCLUSION

Counting blooms is not enough. Cyanobacterial bloom assessment must move from precision-driven enumeration alone to action-oriented evidence use. The goal is not to eliminate uncertainty before every action; it is to identify what uncertainty is acceptable for immediate precaution, what evidence is required for confirmation, and what data are needed to release or revise a decision. The best cyanobacterial monitoring system is not the one with the most advanced instrument, but the one that turns minimum credible evidence into the right public-health action within the available time, then uses reference-calibrated measurements to correct, justify, or release that action.