90th Percentile Lead Result Above the Action Level: Calculation, Verification, and the Corrosion Question

Rank from highest to lowest (the 18th lowest is the 3rd highest):

1. 0.048
2. 0.031
3. 0.026
4. 0.022
5. 0.019
6. 0.018
7. 0.017
8. 0.017
9. 0.016
10. 0.016
11. 0.015
12. 0.014
13. 0.012
14. 0.010
15. 0.009
16. 0.007
17. 0.005
18. 0.004
19. 0.003
20. 0.002

With 20 samples, the 18th value counting from the lowest is the 3rd value counting from the highest: 0.026 mg/L.

Conclusion. The lead action level is 0.015 mg/L, and an exceedance is triggered when the 90th percentile is greater than that value. 0.026 > 0.015, so the action level is exceeded, and it is exceeded by a wide margin, not marginally.

Sanity check on the arithmetic style. Do not average the results and do not use the maximum (0.048). The 90th percentile is a positional statistic: 10 percent of sites are allowed to be above the level. Also note that 11 of the 20 results are at or below 0.015, which shows how easily a mean would have obscured this problem. A mean of these 20 values is about 0.014 mg/L, below the action level. Reporting the mean instead of the 90th percentile would be a reporting error with real consequences.

Verify the data first (days):

• Site eligibility and tiering: are these genuinely Tier 1 sites, and were any sampled out of the correct tier?
• Sampling protocol: first-draw after the required stagnation, correct bottle, no aerator removal or line flushing before the sample, correct lab and method with QA/QC and detection limit well below 0.015 mg/L.
• Site history: have any of these sites been sampled before, and what did they show? Is the 0.048 site new (recent plumbing work, partial service line replacement, new brass fixture) or a long-term high site?
• Field blanks and a duplicate at one high site to rule out contamination.

Investigate the spatial pattern (weeks):

• The two highest sites are on the same street, both with lead service lines. That is not a coincidence in a system with no inhibitor and a finished pH of 7.3. Check whether that street had a partial lead service line replacement recently, which is a well-documented cause of lead spikes from galvanic coupling, and whether there was a source or blend change that altered water quality.
• Characterize water quality at those sites: pH, alkalinity, temperature, and any inhibitor residual, compared with the plant. At pH 7.3 and 45 mg/L alkalinity the water is only marginally passivating, so small changes matter.

What the exceedance triggers (confirm against current federal and state rule text, since the lead and copper rule has been revised and compliance dates have shifted):

• Public education delivered within the required timeframe to all bill-paying customers and to occupants of the sites with elevated results, with prescribed health-effects and exposure-reduction content.
• Continued tap monitoring on the required schedule and water quality parameter monitoring at the plant and in the distribution system.
• Corrosion control treatment evaluation and, if the system is not already at optimal corrosion control, a treatment recommendation and state review.
• Lead service line replacement obligations and planning, with prioritization of the highest-result sites.

Because these requirements have changed with the federal rule revisions, the plan must be checked against the state's current written rule and the system's most recent state correspondence before the public education and replacement schedules are committed.

Focus of the corrosion control evaluation:

• pH first. At a finished pH of 7.3 the water has little capacity to form and hold a lead-scale passivation layer. Raising pH toward 7.8 to 8.0 (with jar/pipe-loop confirmation that coagulation and disinfection are unaffected) is the cheapest first lever, especially since a chloraminated system also benefits from higher pH.
• Alkalinity second. Alkalinity around 45 mg/L is not obviously deficient, but the interaction between pH and alkalinity on scale solubility is what matters, so evaluate them together, not separately.
• Inhibitor decision third. With no inhibitor in place and a demonstrated exceedance, evaluate orthophosphate addition using a bench or pipe-loop study on actual system water and actual pipe, targeting a residual that can be verified at the far end of the system. An inhibitor that never reaches the high sites does nothing.
• Distribution water quality parameters at the high sites, not just at the plant, because the plant may be fine while the ends of the system are not.

In the affected homes, now:

• Offer certified filters (and pitcher/certified filter replacement supplies) to the two highest-result households and any household on that street, with instructions and a reminder that a filter only works if it is certified for lead and replaced on schedule.
• Give practical exposure-reduction advice: flush the tap before drinking or cooking when water has been sitting, use cold water for formula, and do not use hot tap water for consumption.
• Prioritize replacement of the lead service lines on that street and check those two sites specifically for galvanic connections from a prior partial replacement.
• Retest those households after corrosion control changes, and tell them when and why.

Judgment point. The temptation is to reach straight for the inhibitor and declare the problem solved. The pattern here (two adjacent homes, both with lead service lines, no inhibitor, low pH) points to a source-and-scale problem that may be substantially reduced by pH adjustment plus focused replacement, and the corrective action should be chosen on data from these sites rather than on the general principle that inhibitors are good.