An initial assessment of short-term eruption forecasting options in New Zealand

There are currently no quantitative short-term eruption forecasts based on peer-reviewed and validated models that are operational for New Zealand's volcanoes. Specific forecasts produced for work-risk assessments are not generally publicised. During a volcanic crisis, eruption forecasts are demanded under high stress and time-restricted conditions. Many forecasting options exist but none are proven as universally viable, with testing and calibration limited to the hindcasting of specific events. Here, we compare the requirements of six methods with currently available data and monitoring capabilities at each of New Zealand's volcanoes to determine which methods are currently feasible, as well as those options that may be implemented with additional effort or equipment. In New Zealand, the major limiting factor in method selection is the low number of past instrumentally monitored eruptions. This data gap may be filled by carefully selected analogue data from a global volcano set and expert knowledge. Event trees and the failure forecasting method may be set up at most volcanoes with minimal effort, but the latter can only forecast eruption onset time. Expert interpretation is the only method available in New Zealand for any forecast output type.