Derivation of effect benchmark values for each taxonomic group, which has been difficult due to lack of experimental effects data, is required for more adequate protection of the environment from ionising radiation. Estimation of effects doses from nuclear DNA mass and subsequent species sensitivity distribution (SSD) analysis were proposed as a method for such a derivation in acute irradiation situations for assumed nuclear accident scenarios. As a case study, 5% hazardous doses (HD 5s), at which only 5% of species are acutely affected at 50% or higher lethality, were estimated on a global scale. After nuclear DNA mass data were obtained from a database, 50% lethal doses (LD 50s) for 4.8 and 36% of the global Anura and Caudata species, respectively, were estimated by correlative equations between nuclear DNA mass and LD 50s. Differences between estimated and experimental LD 50s were within a factor of three. The HD 5s obtained by the SSD analysis of these estimated LD 50s data were 5.0 and 3.1 Gy for Anura and Caudata, respectively. This approach was also applied to the derivation of regional HD 5s. The respective HD 5s were 6.5 and 3.2 Gy for Anura and Caudata inhabiting Japan. This HD 5 value for the Japanese Anura was significantly higher than the global value, while Caudata had no significant difference in global and Japanese HD 5s. These results suggest that this approach is also useful for derivation of regional benchmark values, some of which are likely different from the global values.