Since the 1980s, radiocarbon (¹⁴C) has gained attention as a valuable tool to quantify the amount of fossil and non-fossil emissions of CO₂ and CH₄ in the atmosphere. Since the 1970s, however, important ¹⁴C emissions in the atmosphere also occur through the operation of nuclear power plants. The limited knowledge about these emissions challenges the use of ¹⁴C as a universal source apportionment tool. Depending on the reactor type, ¹⁴C is emitted in different forms; in particular, pressurized water reactors emit ¹⁴C as a mixture of ¹⁴CH₄ and ¹⁴CO₂. However, few atmospheric ¹⁴C measurements close to nuclear power plants are available, which mostly address ¹⁴CO₂ emissions. Argon-37 (³⁷Ar) can also be produced in nuclear reactors; however, its atmospheric measurement is challenging, resulting in limited available data. In this study, we sampled ambient air during 20–75 min into 18 individual bags around the pressurized water reactor in Gösgen, Switzerland, at the beginning of the annual revision period in 2019, when ¹⁴C and ³⁷Ar emissions can be expected due to the depressurization of the reactor. These samples were analyzed for ¹⁴CH₄, ¹⁴CO₂ and partly for ³⁷Ar. About 1 km downwind of the stack, we found background-corrected activities up to 1900, 370, and 93 mBq m⁻³ respectively. Considering corresponding background activities of 0.3, 48 and 2 mBq m⁻³ for ¹⁴CH₄, ¹⁴CO₂, and ³⁷Ar, this represents an excess of about 6300, 7.4, and 47 times, respectively. Using an atmospheric dispersion model, we satisfactorily simulated the ¹⁴CH₄ and ¹⁴CO₂ activities in the surroundings of the reactor during this event. Our measurements emphasize the importance of nuclear power plants in the interpretation of atmospheric ¹⁴C measurements and show that pressurized water reactors represent a serious limitation in the use of ¹⁴C for source apportionment of CH₄ sources. Our results also provide insights into the approximate magnitude of civilian ³⁷Ar emissions from nuclear facilities specifically during maintenance operations.