How do you verify a cooling tower biocide program is working?

This method covers open recirculating cooling towers and evaporative condensers under a routine biocide program. It does not replace a jurisdiction's Legionella rules or an outbreak investigation led by public health authorities, and its thresholds are New York City and UK values, not universal limits.

When should you verify a cooling tower biocide program?

A cooling tower biocide program should be verified on a fixed routine, after every change to the treatment, and after any event that could have let bacteria grow. In New York City, a qualified person must evaluate the system before an existing chemical treatment is changed or a new agent is introduced, and that evaluation must describe the optimum chemical level in a form usable as a system performance measure (24 RCNY 8-05, NYC DOHMH, 2026). So the target exists before the first result.

HSG274 Part 1 recommends weekly Legionella sampling at first when a system is commissioned or a new or modified treatment programme is established, with the frequency reviewed once the system is shown to be under control (HSE, 2024). New York City requires additional emergency Legionella sampling after a power failure or a loss of biocide treatment long enough for bacteria to grow, or a failure of conductivity controls to hold cycles of concentration (24 RCNY 8-05(f)(3), NYC DOHMH, 2026).

Four verification points follow from these rules:

1. Routine: a monthly review of the weekly indicators and the residual log.

2. Change: a new biocide, feed point, control setpoint or water source.

3. Upset: loss of feed, power, circulation or blowdown control.

4. Performance drop: a fall in heat transfer that has no scale or air leak explanation.

A heat transfer drop counts because, at one power plant, condenser terminal temperature difference improved soon after the biocide program was corrected (POWER, 2013).

What data do you need to verify cooling tower microbial control?

Verifying cooling tower microbial control needs four kinds of data: the dose and residual record, the microbial results with their sampling details, the operating record that explains changes, and an outcome measure such as heat transfer. A microbial count alone cannot tell you whether the biocide failed or the system changed around it.

The dose and residual record covers biocide feed, oxidant residual and pH. CDC recommends that disinfectant residual be monitored and adjusted by an automated system, and that disinfectant residual and pH be monitored on a regular basis (CDC, Legionella Control Toolkit). The pH matters because HSG274 Part 1 states that it is essential to pH-adjust DPD No 1 halogen results, since the more active hypochlorous and hypobromous acids decline as pH rises (HSE, 2024).

The microbial record needs each result with its sample point, time relative to the last dose, method and laboratory. New York City requires the Legionella result in a final laboratory report with a signed chain of custody and the laboratory's ELAP certification number (24 RCNY 8-05(f)(3), NYC DOHMH, 2026).

The operating record holds make-up and blowdown volumes, conductivity and cycles of concentration; the US Department of Energy advises keeping that log and watching trends to spot deterioration (DOE FEMP, Best Management Practice 10). Add shutdowns, low flow and makeup changes.

The test method belongs in the record too. At one plant, manganese in the makeup water interfered with the DPD chlorine test, so staff switched to an amperometric method (POWER, 2013).

Which indicators show cooling tower microbial control: oxidant residual, HPC, ATP, biofilm or Legionella?

Each indicator of cooling tower microbial control answers a different question: oxidant residual shows the dose is present, HPC or dip slides and ATP show bulk water bacteria, a biofilm check shows surfaces, and Legionella culture or qPCR shows the hazard itself. No single indicator replaces the others, because each one misses something.

The table below sets the criterion for each indicator and what to do when it is not met.

The key row is Legionella culture: a positive result is not a score in the biocide verdict but a trigger for the corrective actions written into the site's water management plan, and in New York City those actions run by Table 8-1 (24 RCNY Table 8-1, NYC DOHMH, 2026).

Oxidant residual and HPC are the indicators most sites already collect, and together they can carry a weekly verdict. ATP, biofilm checks and qPCR add speed or surface coverage, but each needs a site baseline or a culture result beside it before it can carry a verdict alone. On a program that uses a non-oxidizing biocide, the residual row is weaker, because HSG274 Part 1 notes that non-oxidizing biocide concentrations cannot easily be measured (HSE, 2024); the dose record and the HPC row then carry more of the verdict.

Where and when should you sample to verify a cooling tower biocide program?

Samples to verify a cooling tower biocide program should come from circulating water downstream of the heat source, away from the biocide feed point, at the point in the dosing cycle where the residual is lowest. HSG274 Part 1 recommends dip slides downstream of the heat source, usually from the return line to the tower, with the sample point flushed first so the sample is representative (HSE, 2024).

Timing decides whether a result means anything. HSG274 Part 1 recommends a sampling point remote from the biocide dosing point and, for shot-dosed biocides, sampling when the residual is at its lowest, ideally at the same time each week (HSE, 2024). New York City allows indicators at any time on constant chemical treatment, but requires samples from intermittent biocide systems to be taken before biocide application (24 RCNY 8-05(f)(2), NYC DOHMH, 2026). A sample taken just after a shot dose measures the dose, not the control.

Hold the method constant as well. HSG274 Part 1 recommends incubating dip slides for at least 48 hours, usually at 30 °C, with the same time and temperature each time (HSE, 2024). For Legionella, samples come from circulating water near and downstream of the heat source, the biocide is neutralised where possible, and culture follows BS EN ISO 11731:2017 (HSE, 2024).

Write the sample point, time, method and incubation into the monitoring plan before the first result, so a later change shows up as a method change, not a program change.

What HPC and ATP levels signal a cooling tower biocide program problem?

An HPC or dip slide result of 10,000 CFU/mL or more under New York City rules, or more than 10,000 CFU/mL under UK guidance, signals a biocide program problem, while ATP has no published threshold and signals a problem only against the site's own baseline. In New York City, such a result requires disinfectant residual checks three times per day until targets hold for at least 24 hours; if targets are not reached within three days, a Legionella sample is collected (24 RCNY Table 8-2, NYC DOHMH, 2026). The team then retests bacteriological indicators 48 hours after reaching the target residual (24 RCNY Table 8-2, NYC DOHMH, 2026).

HSG274 Part 1 splits the range. Results above 10,000 and up to 100,000 CFU/mL call for a programme review, biocide adjustment if appropriate and a resample after 24 hours; above 100,000 CFU/mL, a shot dose or higher continuous dosage and a resample after 24 hours (HSE, 2024). It warns against relying on frequent shot dosing without finding the underlying problem (HSE, 2024).

Trends matter below the limit. HSG274 Part 1 recommends watching week-to-week changes and reviewing the system after any increase (HSE, 2024).

ATP measures cellular ATP as an indicator of total metabolically active microbial contamination, and ATP data generally covary with culture data, but different factors affect each (ASTM D4012-23a). HSG274 Part 1 recommends running such alternative tests in parallel with culture methods for a period so action levels can be set (HSE, 2024).

Why can HPC look fine while a cooling tower tests positive for Legionella?

HPC can look fine while a cooling tower tests positive for Legionella because routine culture media do not grow Legionella and because Legionella can persist in biofilm that bulk water counts do not see. HSG274 Part 1 states that dip slides are coated with nutrient agar on which many microorganisms grow, "but not legionella" (HSE, 2024). A low count therefore shows general bacterial control, not the absence of Legionella.

Published data show the gap. In an industrial cooling tower in Bologna, untreated water held HPC at 36 °C of 2.54 log CFU/mL while Legionella pneumophila serogroup 8 was 5.06 log CFU/L (Iervolino et al., 2017). Converted, that is roughly 350 CFU/mL by HPC, far below the 10,000 CFU/mL line, against about 115 CFU/mL of Legionella, which would sit in level 3 of New York City's Table 8-1 (our conversion of the log values; 24 RCNY Table 8-1, NYC DOHMH, 2026).

The same study shows why biocide choice needs verification. Hydrogen peroxide and silver treatment raised Legionella to 6.14 log CFU/L, while a sodium hypochlorite shock followed by continuous dosing at 2 to 3 mg/L brought it to 1.77 log CFU/L by the last sampling (Iervolino et al., 2017). Between those points, a fall in residual chlorine came with a rise in Legionella (Iervolino et al., 2017).

HSG274 Part 1 adds that sporadic positives are not uncommon even with low counts, but repeated positives, or positives with poor biocide control or poor counts, should be investigated (HSE, 2024).

Worked example: a monthly cooling tower biocide program scorecard (illustrative)

A monthly cooling tower biocide program scorecard puts each indicator, its result and its fixed criterion on one page, so the verdict follows from the criteria rather than from the reviewer's impression. The example below is illustrative: the tower, the results and the residual target are invented to show the format, while the criteria come from the cited sources.

The scorecard below decides whether the month's control is accepted, needs action, or goes to a separate procedure.

The key row in this illustrative month is the biofilm check: the dip slides met the limit, yet slime on the fill and a rising ATP trend point to surface growth that bulk water counts miss. The verdict would be "control partly verified, action required on residual gaps and fill cleaning", with the two low-residual days explained before the next month. In a real scorecard, each illustrative value is replaced by the logged result, its sample record and the name of the person who took it.

Two rules keep the scorecard honest. First, criteria are fixed before the month starts; a criterion changed after the result is a new review, not an adjustment. Second, the Legionella row records the result and the date it went to the water management plan; it never averages into the verdict.

How do you report cooling tower biocide program performance?

Cooling tower biocide program performance is best reported as a verdict per indicator against criteria fixed in advance, with the sample point, timing, method and laboratory attached to every result, and with life-safety signals reported in a separate section. Each verdict should trace to the record behind it.

A useful report has five parts:

1. The criteria and their sources, dated before the period began.

2. The results table, with sample point, time relative to dose and method for each value.

3. Trends: residual compliance, weekly counts and ATP against baseline.

4. Operating context that could explain a change: cycles of concentration, makeup changes, shutdowns, feed failures.

5. A separate section for Legionella results and the corrective actions taken, with dates.

When a customer says the biocide stopped working, the report has to show which results reflect the product and which reflect feed, flow or water changes; the same separation is the core of showing whether a product complaint is really the product.

State what the data cannot show. Bulk water counts do not prove surfaces are clean, qPCR genomic units are not directly comparable with culture limits (HSE, 2024), and a month without a Legionella positive does not mean the absence of risk (HSE, 2024).

The Performance evaluation Task in Lubinpla ties each sample, method and result into one evidence record for the verdict, and hands life-safety signals such as a Legionella positive to a separate procedure, in a cited document for your review.

Where cooling tower biocide verification does not apply

Cooling tower biocide verification as described here does not apply to outbreak response or to jurisdictions whose rules differ from New York City and the UK, and applies only partly to non-oxidizing programs without a measurable residual. When an outbreak or associated illness is suspected, the public health authority having jurisdiction decides that, and CDC recommends testing in conjunction with public health before and after remediation (CDC, Legionella Control Toolkit).

Non-oxidizing biocides are harder to verify by residual. HSG274 Part 1 notes that their concentration cannot easily be measured, unlike oxidizing biocides with a simple field test (HSE, 2024), so the dose record and microbial results carry more weight, and HSG274 Part 1 recommends taking Legionella samples immediately before a non-oxidizing shot dose (HSE, 2024).

The thresholds here come from New York City's Chapter 8 and the UK's HSG274 Part 1. Other jurisdictions set their own Legionella bands and sampling rules, so the criteria table must be rebuilt from the local rule before use. qPCR results also sit outside the culture bands, because the assay detects DNA from live and dead cells (HSE, 2024).

Finally, this method verifies microbial control, not corrosion or scale control. A program can pass every microbial criterion while oxidant overfeed corrodes metal, which is why HSG274 Part 1 recommends good mixing that avoids localized high biocide concentrations (HSE, 2024).

Key takeaways

• Fix the criteria, sample points and timing before the first result; a target set afterwards is not verification.

• Sample downstream of the heat source, away from the feed point, before intermittent doses.

• HPC or dip slides at about 10,000 CFU/mL are the shared trigger for a program review: 10,000 or more in New York City, more than 10,000 in UK guidance (24 RCNY Table 8-2, 2026; HSE, 2024).

• Low HPC does not rule out Legionella; route any positive to the water management plan instead of scoring it.

• Report method, timing and laboratory with every number, and say what the data cannot show.