Performance of the program ASHFALL for forecasting ashfall during the 1995 and 1996 eruptions of Ruapehu volcano

The program ASHFALL was developed to predict the ashfall that would result from a volcanic eruption, primarily for civil defence purposes, such as giving public warnings and planning mitigation measures. It can also be used for volcanic hazard assessments and for studies of past eruptions. During the 1995 and 1996 eruptions of Ruapehu, a volcano in the North Island of New Zealand, the program was used to provide forecasts of where ash would land if a significant eruption occurred, based on forecast wind patterns. This paper compares the program predictions with the actual ash distribution of three major ash-producing events from Ruapehu in 1995 and 1996, and looks at the main sources of error. The accuracy of the forecast wind direction is the main factor affecting the ash distribution. If the forecast wind direction is correct, ASHFALL gives a good indication of where ash will fall. The wind strength has less effect on the ash distribution. The quantity of ash falling downwind depends on the volume of the eruption, and it is difficult to give more than an order of magnitude estimate of this at the time of the eruption. However, analysis of the three events for which the total eruption volumes could be derived from measured ash thicknesses, showed that the ash thickness at any point was generally within a factor of two of that forecast by ASHFALL.