Berna Baradan studied architecture in Izmir, northwestern Turkey. Her Masters thesis focused on temporary shelter after disasters, based on research after an earthquake struck her city in August 1999. She is currently completing a doctorate at Izmir Instutite of Tecnology.

More than 14,000 people lost their lives when a severe earthquake devastated northwest Turkey in August 1999, and of those that survived about 600,000 lost their homes.

The disaster highlighted the importance of temporary disaster accommodation and good housing conditions.

Turkish authorities started work immediately to try to meet the need for temporary shelters.

For short-term use, tents were distributed in the two weeks following the earthquake and prefabricated units intended for medium-term use were constructed within 18 months.

In order to meet the urgent need for shelter, 121 tent cities were built. They were used for almost two years in some districts, in spite of the recommended use of maximum 90 days.

The users complained about unhealthy living conditions, shared bathrooms and laundries, and the inadequacy and poor waterproofing of tents during the winter and rainy weather.

There was insufficient stock of prefabricated shelters for a disaster of this magnitude, so an alternative system was chosen hurriedly to deal with the urgent need.

As result, a decision was made without consulting experts. The system chosen turned out not to be reusable, and consequently uneconomic.

Residents moved to the accommodation sites were not informed that their shelters were intended to be a medium-term solution.

During my investigations in the region, I saw that one-roomed units with a floor area of 30 square metres were insufficient for most users and 59 percent them made some kind of addition to the shelter. The shelters were not designed to take account of families of different sizes.

In the same survey, 68 percent of inhabitants said they had to repair their shelters because of poor workmanship and bad weather conditions.

Conventional methods and materials used for repairing the units were quite different from the original prefabricated system, and meant that the units would not be reusable on future occasions, which was the reason that temporary shelters were used in the first place.

From my case study, I concluded that dividing the production of disaster shelters into medium-term temporary and permanent phases caused a considerable waste of time and money.

This was because of the large number of homeless people compared with the stocks of shelters and the impossibility of re-using units at a later date.

It would have been more appropriate to temporarily locate the earthquake victims in tent cities suitable for winter use, and in summer and recreational camps, so that the production of rapid prefabricated systems for permanent use could have begun without wasting time.

The best policy for temporary shelters in countries in disaster-prone regions such as Turkey is to design them based on general requirements, local needs and their specific characteristics and to manufacture them before disasters.

It is best to stock about 1,000 shelters with their installations in probable earthquake sub-regions.

If this is not possible, alternative settlement options would be to use tent cities suitable for winter usage and summer camps, while permanent settlements could be planned and built as soon as possible.