Environmental distribution and long-term dispersion of reactor 14CO2 around two German nuclear power plants

Carbon-14 data on atmospheric CO2 as well as on plant material (tree leaves and wheat) from the vicinity of two German boiling water reactors (Philippsburg and Isar/Ohu) are reported. Atmospheric CO2 samples taken routinely with an integration time of one or two weeks 1.75 km downwind of the Philippsburg reactor (900 MW electrical power) show a maximum 14C excess concentration of delta 14C (excess) = 300 +/- 7%, corresponding to 12.7 mBq m-3 (STP air). The long-term average excess amounts to delta 14C (excess) = 47 +/- 3%, corresponding to 2.0 mBq m-3 (STP air). The concentrations observed with plant material at the same sampling site range between delta 14C (excess) = 0% and 120%, corresponding to 0 and 27 mBq (g carbon)-1. With the meteorological dispersion parameters actually measured at the nuclear power plants, the dispersion factors for the various sampling sites and for the individual periods of sampling were calculated on the basis of a one-dimensional Gaussian plume model. With the observed 14C excess concentrations and the dispersion factor, a "theoretical" (i.e. calculated) reactor 14C source strength is then determined. For the Philippsburg reactor, which is situated in the flat Rhine valley, the "theoretical" and the observed yearly mean 14C emissions compare rather well (within a factor of 2). A significant systematical deviation from the model was found in the concentration decrease with source distance: the decrease predicted between the 1.75-km and 3.25-km distances is steeper than actually observed. The 14C excess concentrations found in tree leaves around the Isar/Ohu reactor (907 MW electrical power) at 1-2 km distance fall into the same range as observed at Philippsburg. In the hilly terrain at this reactor site, the model calculations agree well with the observed values up to a distance of 1 km if the relative elevation of the sampling site is taken into account by introducing a "reduced stack height" in the model calculations. This method fails in predicting the concentrations at distances greater than 1 km from the source.