Radon and Time

A 48-hour measurement, a 3-month measurement, and a continuous monitor can all produce different values without any "error" — they are measuring different time windows of a constantly varying concentration. Indoor radon at any instant reflects entry rate (advective soil-gas flow driven by pressure differences, plus diffusive entry from soil and materials) balanced against removal rate (air exchange plus radioactive decay, half-life 3.8 days).

WHO states that "high temporal variation of indoor radon makes short-term measurements unreliable for most applications" and "temporal variations of a factor of two or more commonly occur in repeated short-term measurements." UK validation schemes cover only measurements ≥3 months "because radon levels are highly variable on all timescales."

A single measurement tells you the mean concentration during that window under the ventilation, weather, and occupancy conditions of that window — nothing more. Whether it estimates the annual average well depends on window length, season covered, and whether the building responds strongly to weather or unusual conditions.

Minutes to hours: Spot readings can differ from annual mean by >200% (Tsapalov & Kovler 2024). Door/window opening, exhaust fans, HVAC cycling, wind gusts, and sudden barometric drops produce rapid changes. WHO excludes grab samples from recommended methods because they "do not capture fluctuations in radon over time."

Daily (diurnal): Indoor peak at night/early morning, minimum in afternoon, typically within a factor of about 2 in homes. Day-use buildings with HVAC off at night can show night levels 3,000–6,000 Bq/m³ vs daytime <200 Bq/m³. Health Canada notes levels commonly change "by a factor of 2 to 3 over a 1-day period."

Weekly: In public buildings, weekend/non-operational periods are significantly higher. Study of 20 multi-use facilities found WHO guideline exceedance rates "increasing by up to two- to threefold" during weekends. Accumulation was "more strongly influenced by duration of ventilation shutdown than by air exchange rate."

Monthly (synoptic weather): Finnish dwellings showed geometric standard deviations of short-term-to-annual ratios declining slowly with duration: 1.65 (1 day), 1.56 (3 days), 1.48 (7 days), 1.38 (14 days). June values were higher: 1.89–2.09. Weather in the preceding 5 days significantly affects short-term results.

Tsapalov & Kovler provide indicative 95th-percentile deviations of duration-t results from true annual mean under normal living conditions: 2 days ≈160%, 7 days ≈120%, 1 month 105%, 2 months 100%, 3 months 85%, 6 months 45%, 7 months 35%. Under closed-building conditions (12 h before, maintained during test), values are lower but still substantial: 2 days ≈105%, 7 days ≈75%, 3 months ≈38%.

Independent Romanian data (71 houses) agree: CV of temporal correction factors was below 10% for measurements >7 months, but close to 20% for measurements up to 4 months. Field correlations between short- and long-term: Slovenian study found 7–10-day continuous averages agreed with 2-month etched-track averages within ±20% "only in about 1/3 of cases."

US paired dataset (2,245 buildings): short-term tests explained about 75% of variance in subsequent long-term measurements at best, leaving a quarter or more unexplained. Direction of bias: closed-house winter short tests tend to overestimate; summer or well-ventilated short tests underestimate.

UK national survey (2,300 homes): individual-home correction factors were log-normally distributed with GSD of about 1.36, meaning a corrected 3-month result can still be wrong by a third or more. Miles et al. found that spring and autumn measurements gave better estimates of annual mean than the best seasonal correction factors applied to all-season measurements.

Factors must be regionally derived and do not apply to buildings with atypical radon drivers. UK 1-month correction factors range 0.73 (January start) to 1.61 (July start), implying winter/summer ratio of about 2.2. Spain 1996 survey: summer/winter GM ratio 0.40 (ratio 2.5). These are population averages — individual buildings vary widely.

Even 12-month averages are not fixed. Rome 10-year follow-up: overall year-to-year CV of 17%. Minnesota (1,700 year-long measurements over 1983–2000): median variation 26% (range 3–110%). Colorado: mean CV 25%. Other European and US studies ranged 15–62%.

WHO notes that correcting for this random year-to-year variability roughly doubled the excess relative risk in the European pooled analysis, to 16% per 100 Bq/m³. Tsapalov & Kovler suggest this year-to-year variation sets a floor on achievable measurement accuracy, making instrument calibration uncertainty much below ~20% of little value for long-term tests.

UK study (470,689 homes, 1980–2015): homes with double glazing had GM radon 67% higher than non-retrofitted homes. Irish modeling: increases up to 107% after energy retrofits, driven mainly by reduced air permeability. Adding ventilation measures prevented the increase.

French survey (3,233 Brittany houses): thermally retrofitted homes had median 180 Bq/m³ vs 114 Bq/m³ in non-retrofitted. Finnish follow-up of apartment buildings showed outcomes depend on whether ventilation is upgraded alongside sealing. Any foundation penetration, basement conversion, heating system change, or ventilation switch can shift the baseline.

Screening / quick check: minimum 48 h under closed-building conditions (starting 12 h before); preferred 4–7 days, multiples of 24 h, heating season. Temporal uncertainty still >70–100%, so not adequate for compliance decisions.

Decision to mitigate (homes): minimum 3 months avoiding summer (WHO, Health Canada, UKHSA, Spain CSN GS-11.4); preferred 12 months. UK validated results require 3–12 months. Year-long results remove seasonal correction error.

Regulatory compliance (workplaces, Spain/EU): 3 months between 1 October and 31 May (CSN IS-47), or full year in quarterly periods. ISO 11665-1/-8 defines long-term as ≥2 months; France uses ≥2 months in heating season. Add ≥1 week continuous monitoring if passive result exceeds 300 Bq/m³ and occupancy is time-limited.

Post-mitigation verification: short-term test after system has run ≥24 h (immediate check) plus long-term (≥3 months, same rooms as original test) to confirm annual average is below reference level across seasons.