Wednesday, January 02, 2013

In memoriam of Prof. Ambraseys

Professor Nicolas Ambraseys Dip.Eng DIC PhD FICE FREng (Greek: Νικόλαος Αμβράζης, born 1929 in Athens) is a Greek Engineering Seismologist. He was Professor and later Professor Emeritus of Engineering Seismology and Senior Research Fellow at Imperial College London. Professor Ambraseys studied Rural and Surveying Engineering at the National Technical University of Athens (Diploma in 1952) and then Civil Engineering at Imperial College specialising in Soil Mechanics and Engineering Seismology. He worked with Professors Alec Skempton and Alan W. Bishop and obtained his PhD degree in 1958; his thesis title was "The seismic stability of earth dams". He joined the staff in 1958 as a Lecturer and he was appointed a Reader in Engineering Seismology in 1968 and full Professor of Engineering Seismology in 1974. In 1968 he established the Engineering Seismology Section (ESEE) (now part of the Geotechnics Section) in the Department of Civil and Environmental Engineering of Imperial College and served as its first Head from 1971 to 1994, until he retired and was reappointed as Senior Research Investigator. He founded and became the first chairman of the British National Committee of Earthquake Engineering (more on wikipedia).
Professor Ambraseys passed away on the 28th of December, 2012.

I remember Prof. Ambraseys from my MSc studies at Imperial College London back in the mid 90's. No need to say much; anyone who has attended his lectures has an idea of the experience. His lectures were not based on writings on the board but on discussion. I particularly remember an example he gave one day on probability. He showed a slide on the overhead projector, with a train somewhere in Turkey, overturned by a surface fault which happened to be crossing the railway lines during an earthquake. I think it must had been near Erzincan. This time he used the blackboard. He calculated the probability of occurrence given the seismic hazard at the site, the probability of a surface faulting at the site of the incident, the probability of a train passing exactly the time of the faulting from this very site and the probability of the train being overturned. All and all, something of the order of 1e-10. Nearly impossible to happen. Nevertheless, the train had crashed and people had died. Still possible.

Therefore, he concluded, "you should see statistics in a relative and not in an absolute sense". Every time I crash on uncertainty issues and every time I compute statistical measures I remember this quote. Statistics mean nothing on their own, unless you critically compare them with a carefully selected measure.   

No comments: