Power, W.L.; Wang, X.; Barberopoulou, A.; Mueller, C. 2014 Validation of a deaggregation-based approach for tsunami evacuation mapping. Lower Hutt, N.Z.: GNS Science. GNS Science report 2014/36 56 p.
Abstract: In the recent report ‘Review of Tsunami Hazard in New Zealand (2013 Update)’ a process for defining tsunami evacuation zones was outlined. That process involved modelling tsunami inundation scenarios based on earthquake events selected from a deaggregation of the probabilistic tsunami hazard at the 2500 year return period and 84% level of confidence. Deaggregation is a procedure for identifying individual scenarios that correspond to a particular level of hazard. It has been proposed that this process could be adopted as a replacement/update to the ‘Level 3’ tsunami evacuation mapping criteria, as recognised by the Ministry of Civil Defence and Emergency Management. Extensive data on tsunami inundation was collected following the 2011 Tôhoku tsunami in Japan. These data provide an opportunity to validate the proposed evacuation zoning method, by comparing zones developed according to the proposed technique against the actual inundation of the Tôhoku tsunami. A limiting factor in this validation is the quality of the topographic data. The data available for developing the tsunami inundation model was of lower quality than recommended under the proposed ‘Level 3’ criteria – being based on interpolated 10m contour data, rather than having 1m or better vertical accuracy as under the proposed criteria. Four study sites were used for the validation exercise: Kesennuma, Ishinomaki, Shiogama, and the Sendai Plains. All of these are on the east coast of the Tôhoku Region, Honshu, Japan. In three of the four sites the proposed ‘Level 3’ approach to evacuation zoning produced suitable evacuation zones encompassing the inundation area of the 2011 tsunami. Areas of the remaining study site that were not on reclaimed land were also identified as evacuation zones by the method. Any exceptions were very small in area and appear to be attributable to the inaccuracy of the digital elevation model (DEM). Spot height measurements reveal to be within a metre of sea level – significantly lower than specified in our DEM model. Some of these areas were not inundated in our tsunami model, but were inundated by the 2011 tsunami. There is good reason to believe that if a DEM of the recommended accuracy for Level 3 had been used these areas would also have been inundated in our tsunami model, and would therefore have been included in the evacuation zone. Within the constraints of the topographic data quality, the results support the use of the proposed ‘Level 3’ criteria. (auth)