Bounding atmospheric dispersion factors for nuclear facility siting without site meteorological data

Abstract

Atmospheric dispersion factors (χ/Q) are typically used in accident dose analysis for nuclear regulatory compliance. Regulatory guidance supplies two routes: an onsite record of a year or more, or qualified offsite data from representative nearby stations. At many proposed advanced reactor sites, the second is unavailable because no representative station exists: remote siting, sparse networks, complex terrain, or coastal exposure. A third route needs no site meteorology: a dispersion factor demonstrated conservative rather than representative, but has never been supplied with a bound or an acceptance basis. This paper provides both. ARCON96 was characterized by full-factorial parameterization and one-factor sensitivity over the admissible design region, in 3,386 code executions using constant-condition meteorology that removes weather as an input. Common intuitive bounds are non-conservative: a low-wind file understates the bound by up to 2.8 times near the source, and a near-field file is non-conservative by up to 40 percent at 1,200 m. The highest values occur at a moderate wind rather than the lightest: χ/Q is non-monotonic in wind speed, and the worst case switches regime with distance. The presented per-distance envelope bounds two real records at every distance and interval evaluated, by factors of 1.07 to 2.68. The results provide dispersion factors for an envelope at the exclusion area boundary and low population zone distances when considered with the building-area stated applicability conditions.