Projects per year
Comparisons of the observed global-scale cooling following recent volcanic eruptions to that simulated by climate models from the Coupled Model Intercomparison Project 5 (CMIP5) indicate that the models overestimate the magnitude of the global temperature response to volcanic eruptions. Here we show that this overestimation can be explained as a sampling issue, arising because all large eruptions since 1951 coincided with El Niño events, which cause global-scale warming that partially counteracts the volcanically induced cooling. By subsampling the CMIP5 models according to the observed El Niño–Southern Oscillation (ENSO) phase during each eruption, we find that the simulated global temperature response to volcanic forcing is consistent with observations. Volcanic eruptions pose a particular challenge for the detection and attribution methodology, as their surface impacts are short-lived and hence can be confounded by ENSO. Our results imply that detection and attribution studies must carefully consider sampling biases due to internal climate variability.