Effectiveness of Real-Time Near-Field Tsunami Inundation Forecasts for Tsunami Evacuation in Kushiro City, Hokkaido, Japan - Post-Tsunami Hazard: Reconstruction and Restoration - page 154

and considerably similar to the observations. The algorithm required only 3 min to
produce tsunami inundation forecast maps of 15 sites, whereas a tsunami inundation
model, running on a common laptop computer, required approximately 40 h to
produce the maps (Gusman et al. 2014 ).
We previously built a precomputed tsunami database for Kushiro City from
hypothetical earthquake fault models off the Pacifi c coast of Hokkaido (Tanioka
et al. 2014 ). The precomputed tsunami database was used with the NearTIF algo-
rithm to produce a tsunami inundation map in Kushiro City from an assumed future
earthquake event with an earthquake source mechanism that is the same as that
estimated for the seventeenth century great Hokkaido earthquake. In this paper, we
report the use of tsunami inundation forecast maps produced by the NearTIF algo-
rithm in a tsunami evacuation drill in Kushiro City, Hokkaido. The tsunami evacua-
tion drill was set up with the assistance of the NHK Television Network ( 2013 ) and
involved the residents of Kushiro City as participants. In this paper, we show how
the NearTIF algorithm works and report the effectiveness of using a tsunami inun-
dation forecast map in a tsunami evacuation drill.
Section 11.2 describes every component in the NearTIF algorithm. Section 11.3
shows inversion methods that directly or indirectly estimate a tsunami source model from
seismic wave data, GPS time series data, or tsunami waveforms. Section 11.4 describes
the identifi cation of Kushiro City as a tsunami-prone area. Section 11.5 describes the
assumed earthquake for the tsunami evacuation drill in Kushiro. Section 11.6 reports
how the participants in the tsunami evacuation drill learned the effectiveness of
using a tsunami inundation forecast map during the simulated evacuation.

11.2 Method of Near-Field Tsunami Inundation

Forecasting (NearTIF)

If different earthquakes produced the same tsunami waveforms at near-shore loca-
tions, the tsunami inundations in the coastal areas from those tsunamis would be the
same. Based on this reasonable assumption, we developed a methodology for near-
fi eld tsunami inundation forecasting (Gusman et al. 2014 ). The Near-fi eld Tsunami
Inundation Forecasting (NearTIF) algorithm has three main components: a precom-
puted tsunami database, a tsunami numerical model, and a tsunami database search
engine. These three components are explained in the following subsections.

11.2.1 Precomputed Tsunami Database

The fi rst component of the algorithm is a tsunami database that contains pairs of
precomputed tsunami waveforms and tsunami inundations from hypothetical tsu-
nami source models. The hypothetical tsunami source models can be made from
simple earthquake fault model scenarios with variable earthquake source parame-
ters (i.e., strike, dip, rake, slip amount, and depth), or tsunami source models.