Understanding Radon Measurements

Indoor radon concentrations are approximately lognormally distributed, meaning log(concentration) follows a normal distribution. This arises because radon is the multiplicative product of many independent factors (radium content, emanation, permeability, entry rate, ventilation), each with its own distribution.

Consequence: the distribution has a long right tail with occasional very high values. The arithmetic mean is pulled upward by these extremes and exceeds the median. For example, global indoor radon median is about 40 Bq/m³, but arithmetic mean is 50–60 Bq/m³. The geometric mean, being the exponential of the mean of log-concentrations, is close to the median and more representative of typical exposure.

When summarizing radon data, WHO and national authorities typically report geometric mean (GM) and geometric standard deviation (GSD). For example, UK housing stock: GM = 20 Bq/m³, GSD = 2.6. This means about 68% of homes fall between 20/2.6 ≈ 8 Bq/m³ and 20×2.6 ≈ 52 Bq/m³.

Detector calibration uncertainty: alpha-track detectors typically ±7–15%, electrets ±10–25%, charcoal ±15–30%, continuous monitors ±10–20%. This represents the precision with which the device can convert its signal (track density, voltage drop, alpha counts) to radon concentration.

Temporal variability: the dominant uncertainty source for tests shorter than about 6 months. A 2-day test has 95th-percentile temporal uncertainty >100% relative to annual mean; 7-day test ~70%; 3-month test ~85%. This is not detector error — it is genuine concentration variation over time.

Spatial variability: within a single room, replicate simultaneous measurements show coefficient of variation (CV) typically 10–30%. Between rooms on the same floor, CV can reach 30–50%. Basement-to-ground-floor ratios are typically 1.3–2.5×. Protocol specifies lowest-level frequently occupied room to capture highest exposure.

Year-to-year variability: even 12-month averages vary 15–30% CV from year to year due to weather and building changes. This sets a practical floor on long-term measurement precision.

For a 3-month alpha-track test result of 250 Bq/m³: calibration uncertainty ±10% contributes ±25 Bq/m³, temporal uncertainty ~85% (95th percentile) means the true annual average could plausibly be 135–470 Bq/m³ (using rough factor-of-2 range), spatial variation adds another ±20%. The result is compatible with both being well below and well above a 300 Bq/m³ reference level.

This is why results close to reference levels warrant follow-up measurement. UK guidance: if a 3-month result is within a factor of about 1.5 of the action level, conduct a second 3-month test in a different season or a 12-month test. Spain requires 3 months outside summer or 12 months for workplace compliance.

For a 2-day closed-building test at 180 Bq/m³: temporal uncertainty alone makes this compatible with annual averages from ~60 to >400 Bq/m³. Short tests are suitable for rapid screening (clearly very high vs clearly low) or post-mitigation verification, not for borderline compliance decisions.

Seasonal correction factors adjust sub-annual measurements to estimate annual average. UK factors for 1-month tests range from 0.73 (January start, divide measured value by 0.73 to estimate annual) to 1.61 (July start, divide by 1.61). For 3-month tests, range is 0.74 (December start) to 1.52 (June start).

Factors are population averages and carry their own uncertainty. UK national survey found individual-home correction factors log-normally distributed with GSD ~1.36, meaning a corrected 3-month result can still be off by a third or more. Factors do not transfer between climates: UK/northern European factors do not apply to Mediterranean Spain.

Miles et al. found spring and autumn measurements gave better annual-mean estimates than summer or winter measurements corrected with seasonal factors. The preferred approach is to measure during heating season (October–May in Spain per IS-47) or measure 12 months to eliminate seasonal bias.

Every detector has a lower limit of detection (LLD) below which a measurement cannot reliably distinguish radon from background. For alpha-track detectors, LLD is typically 10–30 Bq/m³ for 3-month exposures, lower for 12-month exposures (longer exposure accumulates more tracks).

For short-term detectors (charcoal, electret), LLD is higher: 20–40 Bq/m³ for 2-day exposures. Continuous monitors have LLDs of 4–20 Bq/m³ depending on averaging time. Results below LLD are reported as "< LLD" or not detected.

At very low concentrations (e.g., upper floors of high-rise buildings, outdoor air), measurement precision becomes limiting. A result of 15 ± 10 Bq/m³ means the true value could plausibly be 5–25 Bq/m³ — still useful for confirming low exposure but not precise.

Certification programs validate detector performance: NRPP (North America), PHE/UKHSA (UK), national programs in EU countries. Devices are exposed to known radon concentrations in reference chambers, tested across humidity and temperature ranges, and must meet accuracy and precision requirements.

Duplicate measurements: deploying two detectors side-by-side checks for device failure and provides precision estimate. If duplicates differ by more than 30–50%, investigation is warranted. Blank detectors (unexposed controls) check for contamination or storage exposure.

Chain of custody: detectors should be tracked from deployment to analysis. Sealed or tamper-evident packaging confirms the detector measured only the intended location and period.

A complete radon report includes: measured concentration (Bq/m³ or pCi/L), detector type, deployment start and end dates, measurement duration, room location, building conditions (closed/normal), calibration laboratory and certificate number, measurement uncertainty, and comparison to applicable reference level.

Results should be expressed as "measured concentration ± uncertainty" when uncertainty is quantified, or as a range when temporal variability dominates. For example, "250 ± 25 Bq/m³ (detector calibration uncertainty) over 3 months October–December" is more informative than "250 Bq/m³."

When comparing to reference levels: a result slightly above the reference level from a short test does not necessarily require mitigation — it warrants further measurement. Conversely, a result slightly below from a summer short test may underestimate winter and annual exposure.