Justifying Use of On-Site Meteorological Observations in Modeling Annual Integrated Impacts on Population within 50 Miles of a Coastal Nuclear Power Plant

As part of renewing a permit for an existing nuclear power plant, the U.S. Nuclear Regulatory Commission (NRC) requires air modeling exercises to calculate potential impacts of releases of radiological pollutants. This paper concerns the Severe Accident Mitigation Analysis (SAMA), which uses of the MELCORE Accident Consequence Code System version 2 (MACCS2) comprehensive modeling system. MACCS2 includes the ATMOS segmented Gaussian plume transport and dispersion model. The endpoint of the MACCS2 model system is the determination of health and economic effects integrated over a year and integrated over the population and economic entities within 50 miles of the plant. The standard procedure is to use the data from the onsite meteorological tower and from a nearby airport as input to the model. However, in the case of the Pilgrim Nuclear Power Station, located on the coast between Boston and Cape Cod, some groups objected to the use of data from the single 67 m tall meteorological tower and the nearby small airport, on the basis that there are known to be complex wind fields and precipitation patterns in the area caused by land and sea breezes, rolling hills (with approximate 50 to 100 m heights), and the contorted shape of the coastline.

We justified use of the single meteorological tower by pointing out that the endpoint did not concern a single one hour maximum impact at a single location during the year (a typical endpoint for EPA regulatory modeling of non-nuclear industrial plants), but integrations in time over a year and in space over a 50 mile-radius area. We obtained meteorological data (the main interest was winds and precipitation) for several years from about 20 meteorological sites in the 50 mile-radius geographic domain around the Pilgrim Station, including ocean buoys. We compared annual wind roses and annual averaged wind speeds for all these sites and showed that the onsite tower data were within the range of the other sites. Further, we applied the CALMET mass consistent diagnostic wind model using wind observations from the full set of over 20 meteorological sites, as well as vertical profile data from the nearest (Chatham, MA) and the next nearest (Gray, ME) radiosonde sites, to calculate trajectories from the plant for each hour. The resulting annual distribution of trajectories was similar to that calculated using the onsite tower.

In addition, because precipitation is one of the main delivery mechanisms for radiological exposure, an analysis of the precipitation data from 2001 from the Plymouth Municipal Airport (used in the SAMA analysis) was made. A comparison to precipitation data measured at eight other stations in the region was conducted and it was shown that the Plymouth precipitation data were a satisfactory representation of the annual precipitation conditions over the SAMA geographic domain. Furthermore, multiple years, before and after 2001, of precipitation data from Plymouth were analyzed to show that the 2001 data are representative of other years at the same location.

An obvious caveat is that the representativeness of a single meteorological observing site will decrease, even for annual average impacts over a broad area, as the surrounding complex terrain has larger variation (such as a site in a deep valley or along a steep coastal mountain range). Precipitation can also strongly vary with location in a mountainous area.