There are three limitations to the methods I used to estimate the number of vulnerable people:
1) People tend to have multiple characteristics simultaneously/ characteristics may change over time
If the numbers of vulnerable people are combined, even though the real number of vulnerable people is not so many, the vulnerability would be increased depending on how many characteristics they have. This means that I have counted some people multiple times, depending on their characteristics. For instance, there are the indicators ‘people with disabilities and no car access’, and ‘older adults over 65 years old and having multiple disabilities’ which have overlap.
This can only be prevented in future research by collecting data on a more regional scale, which is now being introduced in Japan (Action Policies for Supporting Evacuation Activities of Persons Needing Assistance During Forced Evacuations (2013)). The municipalities are mandated to collect data on individual people needing evacuation assistance. My study shows the limitations of the currently available data and the necessity of collecting more data on an individual level.
2) Vulnerability is assumed as binary rather than a continuous variable
If a person has any characteristic, they are equally vulnerable to any other person with any other characteristic. For instance, a person in a wheelchair is equally vulnerable to a person without car access. This may seem unfair. It may be possible to apply weights or gradients, but these are subjective and depend on the culture or even on an individual level; therefore I chose to work with unbiased numbers.
3) The characteristic ‘restricted by commitments’ – are people vulnerable?
If people are only restricted by commitments, they themselves do not have any characteristic that makes them intrinsically vulnerable. However, these people choose not to evacuate themselves immediately, because of sense of duty, relationship to a dependent, or to protect valuable assets. As they don't evacuate themselves immediately and remain in the exposed area, they become vulnerable to the disaster risk.
Pictures by Daniel Vrielink (2014).