GROUNDWATER MONITORING BASICS

What RCRA groundwater monitoring is, and why it needs statistics

30 September 2026 · about a five minute read

Every municipal solid waste landfill in the United States watches the groundwater underneath it, for as long as it operates and for decades after it closes. This article is the picture behind all of that work, without the statistics: what is monitored, how, and why the answer cannot simply be read off the laboratory's results. The other articles in this series each take up one part of it.

It covers landfills regulated under RCRA Subtitle D, the part of the Resource Conservation and Recovery Act that governs municipal solid waste landfills. The groundwater rules for them are in 40 CFR Part 258, Subpart E. States with approved programs add their own rules, and a facility's permit can add more, so check what applies to your site.

The picture

The whole of the work is one chain:

1. Groundwater flows underneath the landfill, in one general direction.
2. Wells are drilled around the landfill to sample it. Some stand where the groundwater has not yet passed under the landfill. Others stand at the edge the groundwater flows toward.
3. At least twice a year, the wells are sampled.
4. A laboratory measures the concentration of each constituent the facility's monitoring program requires.
5. Statistics decide whether what was measured at the downgradient wells is evidence that the landfill has affected the groundwater.

The rule requires the network of wells to do two things at once: represent the groundwater that has not been affected by the landfill, called background, and represent the groundwater passing the landfill's point of compliance. Everything that follows is a comparison between the two.

The uppermost aquifer is the first water-bearing layer below the ground surface, together with any deeper aquifer connected to it beneath the facility. It is the water a leak would reach first, which is why the rule requires the wells to sample it.

The wrong question

At U1, the laboratory measured arsenic at 3 micrograms per litre. At D1, it measured 4. Is 4 greater than 3? Of course it is. But that is not the question the rule asks, and answering it would get almost every site wrong.

The question is closer to this: is what we measured at D1 different enough from how background groundwater normally behaves that it is evidence of a real change, and not ordinary variation? The rule calls a yes to that question a statistically significant increase.

Why a straight comparison fails

EPA's Unified Guidance, the agency's guidance on the statistics these rules require, puts the reason in one sentence:

Uncertainty is present because measurement data exhibit variability, with limited knowledge of the medium being sampled.

In practice that uncertainty has several sources, and each one is a reason a single comparison cannot be trusted:

• Groundwater varies naturally. Concentrations move with the seasons, with rainfall and recharge, and from one well to the next, with no landfill involved. Wells can even differ from each other permanently, which changes which background each well should be compared with.
• Measurements vary. Two samples of the same water, analyzed by the same laboratory, do not return exactly the same number.
• Many results are below what the laboratory can detect. A result reported as "less than 1" is not zero, and it cannot be compared with a number as if it were one.
• Many comparisons produce false alarms. A site with ten wells and fifteen constituents makes one hundred and fifty comparisons every time it samples. Even when nothing is wrong, some will come out high by chance. The Unified Guidance notes that across all of a site's wells and constituents, the rate of these false positives can be much higher than the rate for any one test, and much of its advice is about keeping it under control.

The Unified Guidance runs to nearly nine hundred pages because each of these has to be handled, along with others: whether the data follow a normal distribution, outliers, trends over time, and whether consecutive samples are truly independent. The rule lists the statistical methods a facility may use, and the facility records in its operating record which one it uses.

What happens with the answer

If the statistics find no significant increase, the facility keeps monitoring, and the new results become part of the history for the next time. If they do find one, the rule sets out steps with deadlines: a notice within 14 days, and within 90 days either a move to a more extensive monitoring program or a demonstration that something other than the landfill caused it. Our article on statistically significant increases covers those steps.

A statistically significant increase is evidence to be explained, not a verdict. The rule itself allows it to be shown to come from another source, from an error in sampling, analysis or the statistics, or from natural variation, and that demonstration has to be certified by a qualified groundwater scientist or approved by the state.

If something here does not match how your facilities work, we would like to hear it. Write to support@koynasystems.com.