Magnus Tumi Gudmundsson

The considerable effects that wind can have on estimates of mass eruption rates (MERs) in explosive eruptions based on volcanic plume height are well known but difficult to quantify rigorously. Many explicitly wind‐affected plume models have the additional difficulty that they require the use of centerline heights of bent‐over plumes, a parameter not easily obtained directly from observational data. We tested two such models by using the time series of varying plume heights and windspeeds of the 2010 eruption. The mapped fallout and photos taken during this eruption allow us to estimate the plume geometry and to empirically constrain input parameters for the two models tested. Two strategies are presented to correct the difference in maximum plume height and centerline height: (a) based on plume radius, and (b) by using the plume type parameter Π, which quantifies the relative influence of buoyancy and cross‐wind on the plume dynamics, to discriminate weak, intermediate and strong plumes. The results indicate that it may be more appropriate to classify plumes as either wind‐dominated, intermediate or buoyancy‐dominated, where the relative effects of both wind and MER define the type. The analysis of the Eyjafjallajökull data shows that the MER estimates from both models are considerably improved when a plume‐type dependent centerline‐correction is applied. For one model, we varied the wind entrainment coefficient β . For this particular eruption, we find that the best value for β lies between 0.28 and 0.36, unlike previous suggestions that set this parameter to 0.50.