Based on the analysis of the household survey on the previous “potential tsunami” earthquake experience, especially in September 2007, it was found that there was only a small proportion of people evacuating after receiving the warning (35%). The main reasons given by the respondents who did not evacuate were “did not trust the warning” and “wanted to wait and be sure about the tsunami occurrence”.
Moreover, 42% of those who evacuated did not immediately proceed to evacuate but gathered the household members first, collected important items at home and/or secured the house before evacuating, etc. The analysis also found out that 98% of the respondents performed evacuation in groups; most of them (75%) evacuated with the family members. A simple correlation analysis was
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Received through TV & radio Received only f rom other media 89,80%
67,60%
10,20%
32,40%
"Could you understand the warning easily?"
Yes No
88
performed between the variable “did you evacuate” and some independent variables, resulting in significant correlations as shown in Table 5‐7 below.
Table 5-7 Correlation analysis between evacuation in the past and some selected variables
Kramer's V Kendall´s tau b
Distance to the coast (further from the coast, less people evacuated)
‐0.133
Household income (higher income, more people evacuated) 0.082
Gender (more females evacuated) 0.128
Understanding the received warning message (when better understood, more people evacuated)
0.122 Source: Own analysis based on the UNU‐EHS Household Survey 2008
These results also show that spatial settings, socio‐economic characteristics, and the format of the warning message have a direct influence on the decision to evacuate. Additionally, the households that have elderly family members (with no children) showed a tendency of lower evacuation rate, although this has not shown a significant correlation due to the small sample size for this particular group.
It is also worth considering how people perceived the evacuation urgency conveyed by the tsunami warning. The official message given by the early warning centre was “earthquake with potential tsunami”, followed by instructions from various sources at the local level to evacuate or stay in alert.
Many people also received information indirectly from word of mouth, so that various interpretations were possible. Even for the official “potential tsunami”, there have been different interpretations. As seen in Figure 5‐14, more than half of the respondents still perceived a “potential tsunami” as an uncertain tsunami threat. They tended to stay in alert and wait for further instructions rather than immediately evacuate. It was also emphasized in the interviews with the households and actors at the community level that clear information and evacuation guidance from the authorities or warning centre is really needed in emergency situations so that people do not panic and are able to follow instructions adequately.
89
Figure 5-14 Various interpretations of tsunami warning message (“Do the following statements come into your mind if you receive a “potential tsunami” warning?”)
Source: Own analysis based on the UNU‐EHS Household Survey 2008
Although the evacuation rate was very low in the past experience in 2007, 75% of the respondents intended to conduct evacuation in case of future tsunami early warnings. However, only 37.7%
intended to evacuate immediately alone or with family members while 37.2% would conduct other actions before proceeding to evacuate, i.e. by seeking further information from the media or by observing the coast. An attempt to visualize the impact of the evacuation delay on the overall evacuation time was done in cooperation with TU Berlin / Last‐Mile. The evacuation behaviour was derived from the intention of the respondents to evacuate and classified as immediate evacuation, evacuation with delay, and no evacuation. For each evacuation behaviour class, an assumption for the delay time in starting evacuation was arbitrarily made. The following classification of the evacuation response based on distance from the coast was used:
Table 5-8 Classification of evacuation behaviour for evacuation delay modelling
Intended response Living > 3 km
from the coast
Living within 3 km from the coast
Total
Not evacuating (assumed delay 25 minutes) 35.7% 21.4% 25.1%
Evacuation after finding further information (assumed delay 15 minutes)
23.9% 41.7% 37.2%
Evacuation immediately alone or with family members (no delay assumed at all)
40.3% 36.8% 37.7%
Source: Own analysis based on the UNU‐EHS Household Survey 2008, assumption of delay time based on discussions with TU Berlin/Last‐Mile
The evacuation modelling produced is shown in the Figure 5‐15 below.
0%
20%
40%
60%
80%
100%
not sure whether tsunami will
occur
in alert and wait for
further instruction
immediately evacuate wherever
they are
I don't know Strongly agree Agree
Disagree
Strongly disagree
90
Figure 5-15 Estimated evacuation time with and without evacuation delay (morning scenario)
Source: TU Berlin / Last‐Mile, 2009
The model on the left side is evacuation time without consideration of evacuation behaviour, and on the right side with consideration of evacuation behaviour. Generally, the evacuation time needed was much longer due to delays – which had been expected and assumed. However, the modelling shows the magnitude of total evacuation delay and potential losses of lives caused by evacuation delay (from about 38% of people who needed longer than 30 minutes it increased to 78%). Such a number, although it cannot be presumed to be an absolute number, serves as eye opener of how important risk communication and education of the population in parallel to physical measures really is. The issues of perception and soft measures for effective risk communication and community education are crucial factors that determine the effectiveness of physical measures.