Hemispheric contrasts in recorded volcanic explosivity are sensitive to temporal adjustment and missing VEI

Abstract This study quantifies the sensitivity of the hemispheric contrast in recorded volcanic explosivity to temporal coverage. We compare Volcanic Explosivity Index (VEI) ≥ 4 fractions in an archived Smithsonian Global Volcanism Program catalogue using fixed-sample standardization and missing-VEI sensitivity analysis. Among 7,721 VEI-scored eruption episodes, 11.64% of Northern Hemisphere (NH) entries and 6.72% of Southern Hemisphere (SH) entries reach VEI ≥ 4, with a crude odds ratio of 1.83 and proportion ratio of 1.73. Logistic models use the same 7,595 episodes and empirical covariate distribution throughout. Adding a post-1500 CE indicator reduces the standardized NH–SH difference from 6.00 to 1.04 percentage points, with a volcano-cluster 95% confidence interval of − 2.08 to 4.16. A post-1826 indicator gives a difference of 2.50 percentage points; its interval includes zero under volcano clustering and excludes zero under regional clustering. Excluding geographically ambiguous assignments gives a difference of 1.50 percentage points. In a separate analysis of all 9,913 recorded episodes, three of 25 specified missing-VEI scenarios reverse the hemispheric contrast. The comparison quantifies how calendar specification, geographical support and missing-outcome assumptions affect this catalogue result. Reporting the episode denominator, temporal composition and dependence structure makes the hemispheric comparison reproducible and its interpretation explicit.