Muscle Cs-137 in Fukushima wildlife from rescue and wildlife-control programmes

Hisashi Komatsu , Kiemi Murakami, Kousuke Kanda, Masahiko Kabeya, Kenji Inami

We analysed opportunistic muscle Cs-137 records of wild mammals and birds from rescue and wildlife-control programmes in Fukushima Prefecture, Japan, in 2017–2025. Bayesian hierarchical models, with species random intercepts and left-censored non-detects, related concentrations to a fixed baseline contamination index, the 2012 air dose rate at capture locations. In 182 records of 38 species, a 10-fold higher index was associated with 13.8-fold higher concentrations, adjusted for year and species (95% credible interval [CrI] 8.80–22.1). The ratio was 13.7 (7.55–25.2) with a municipality random intercept and 12.9 within species. Species differed widely (species SD 1.21, 95% CrI 0.80–1.79, a 3.35-fold difference), mostly between classes. With class as a fixed effect, mammals exceeded birds 8.70-fold (95% CrI 4.88–16.6) and the species SD fell to 0.47. Temporal estimates were not robust in direction across specifications and were indistinguishable from physical decay. Because later years had more mammal records (3 of 43 from 2024 were birds) and the year effect was positive between species but negative within species, pooled year estimates partly reflect changing species composition. In two planned guild contrasts with aquatic-associated birds, muscle Cs-137 was 10.3-fold higher in soil/litter-associated mammals (95% CrI 4.80–23.4), whereas other terrestrial birds did not differ clearly (ratio 1.56, 0.716–3.47). The two mammal guilds did not differ clearly (ratio 0.869, 0.291–2.34); a class-only model fitted equally well. Guilds that included class thus largely reproduced the mammal–bird difference, which may also reflect unrecorded time in care or foraging range. Multispecies radiocaesium data should be analysed with species-level effects.