Risky conversations with: Professor Anne Kiremidjian

Anne Kiremidjian is a full professor at the Civil and Environmental Engineering Department at the Stanford University. One of the most well-known earthquake engineers in the world, Professor Kiremidjian has been involved in pioneering studies in seismic hazard and risk in regions such as California, Costa Rica, Guatemala, Honduras and Nepal. With more than 100 publications in peer-reviewed journals, active participation in several advisory committees and 10 distinguished awards, Professor Kiremidjian has contributed strongly to the advancement of science in the various fields of structural modeling, earthquake loss assessment and seismic hazard estimation. On the 8th of March 2016, Professor Kiremidjian took the time to talk to us about her career.

Professor Kiremidjian (in the centre) supported the presentation of GEM models, datasets and results at the Stanford University.

GEM: Throughout your career you have explored the various fields of earthquake engineering and seismology, from the development of seismic hazard models to the estimation of losses, and even the assessment of earthquake resilience. What was the most interesting topic you had the chance to investigate, and what was the most challenging?

Professor Kiremidjian: What I found to be particularly exciting was developing models for the earthquake occurrence in different regions in the world, developing the actual model and trying to understand what are the physics behind it and what are the right analytical formulation to represent this phenomena. That was certainly challenging, but also very exciting. We always recognized that our results were not the ultimate solution, but we felt that every time we tried to solve a portion of the problem, we made progress, and that I found quite rewarding. The most challenging part was definitely finding data. Nowadays we have so much data and we can do so much more with these models. Now we can talk about proof-of-concept and verification of models, but several years ago we did not have this luxury and it took a lot of effort from my part, and also from my students, to collect the right kind of data, and process it to make it usable.

GEM: Some of these models were probably a big surprise to the scientific community, were they applicable right away?

Professor Kiremidjian: Some of them were not immediately applicable. We developed the very first probabilistic time-dependent model. It was challenging enough to do the time-dependent model, and then we dared to even venture to do time and space dependent models, which at the time were new ideas to the seismic hazard and risk community, but talking to the seismologists and the geophysicists who looked at the model fundamentally from a physics perspective, for them it was the right way to approach the problem. I think it has only been in the last several years that people have began to merge the two approaches. I am hopeful that in 10 or 15 years from now the models will be extremely sophisticated and we will marry the geophysics and the stochastic components. Of course data is always a challenge, but we have so many more sources of information now. From spatial imagery, deep penetration radar, electromagnetic wave propagation, various kinds of data that we have not even tapped in yet in order to understand how we can utilize these data in the development of new models.

Seismic hazard map for California (peak ground acceleration for a return period of 500 years) developed by Professor Kiremidjian in 1975.

GEM: That actually leads me to my next question, related with the future of earthquake hazard and risk. Now we have the ability to employ physical models, stochastic simulations, develop numerical models for almost every phenomena, extract information from satellite imagery, so what is your opinion about these new tools that are changing the way we approach seismic hazard and risk.

Professor Kiremidjian: I see a new generation of models coming now. As you already said it, and I will add to your list the very sophisticated methods to analyze structures with nonlinear models. These models are flourishing and have reached a great level of maturity and acceptance. Now we have so many experts that can handle nonlinear time history analyses and develop more accurate fragility functions than we could previously. We have satellite imagery that can helps us identifying heights of buildings and extracting other types of information, which was very difficult to do before. Compiling inventories of structures can be done much easier now. Satellite imagery I see it as a first step, and the second step is to extract additional structural information and develop correlation models between the structural types and what is observed in these images. All of this today is possible because of the computational and analytical power that is available to us. I remember twenty years ago when we were facing computational challenges, we discussed utilizing parallel and distributed processing, but those methods were still new to us. It took many years and several experts to achieve a higher efficiency. Now something that used to run for 20 or 30 hours came down to 20 or 30 minutes. Today using a super computer has become quite normal, and most of our students are using it already. Now we have the computational power and also the memory to handle the enormous amounts of data required for earthquake analyses. I think, in some ways I wish I were born a little later so I could take advantage of all of these new technologies.

Fragility model developed by on of Professor Kiremidjian’s student, establishing the relation between probability of collapse and ground shaking conditional on the level of corrosion. Such models allow considering the age of structures in risk analyses.

GEM: As you know GEM has activities at the global level. It has several projects and partnerships in various parts of the world, and you too have worked in many regions such as South America, Central America, Central Asia, and of course, the United States. So what was your favorite region and what was the main difficulty in working there?

Professor Kiremidjian: Well I started my career as part of a team that worked on hazard mapping of Nicaragua, and then I did California by myself. That was one of the first probabilistic maps of seismic hazard for California. Of course at the time the main challenge was the data. I actually had a fairly large database, but organizing that data and developing a source model was complicated. A lot of the things that are taken for granted today, we were facing for the first time. In terms of regional application, I have to say that I fully enjoyed working in the Central American countries. I fell in love with every single country I visited in Central America. There was something so genuine and unique about each region, this culture really affected me. I found the local people extremely kind and warm, I worked a lot with the civil engineering societies of each country. I am not sure we were able to convince all of them to consider earthquake hazard, but in some countries they ended up endorsing seismic codes. I found working in Central America really fascinating, and amazingly enough there were many seismologists and geologists that had studied the region extensively over the past century. We were able to collaborate with the local experts, and working collaboratively at the time was not so common. Today we all collaborate but at the time it was not usual for engineers to work directly with geologists, and I found that quite rewarding.

First seismic hazard maps developed for Guatemala and Costa Rica in the seventies (peak ground acceleration for a return period of 500 years).

GEM: I am glad you mentioned the value of collaboration. GEM is developing models, tools, datasets and risk results in collaboration with several experts and institutions globally. You are obviously quite aware of the GEM mission, so what would be your recommendation in order to ensure that GEM is successful?

Professor Kiremidjian: I think GEM has done a tremendous job developing all of this software. I think the next challenge for GEM will be to disseminate the tools and make it available to regions in the world through educational programs and by reaching out to the appropriate communities. Not just researchers, but also practitioners, designers and policy makers. It is important to reach all of the different groups that are affected by earthquakes, and show them what are the answers that these tools can provide. I think that would be extremely effective. In my opinion, one way to convince people that earthquake hazard and risk must be taken seriously is to show them what are the potential consequences of an earthquake in their region and in their community. Demonstrate that not just the buildings will be damaged (which is the first thing to do, of course), but show how the infrastructure will be affected. For example, how will hospitals be affected? How will the fire stations be affected? What is the number of homeless that might result from the earthquake? And most importantly, propose solutions! It is one thing to say you have a high risk and this is how many people will be affected and how many buildings will be damaged, but it is certainly more efficient to provide alternative solutions. That was our approach when we were working in Central America. Providing a hazard map was half the work, but it was just one of the steps. Supporting different countries implementing a design code or other similar measures; that is the real challenge in many regions in the world. GEM can play a very important role in this part, using the available tools and results to demonstrate the impact of mitigating risk. You can demonstrate what happens if you design according to the current construction practices, and this is what happens if you construct with an improved code. This would be a great contribution from GEM.

Professor Kiremidjian and her students have worked extensively in time-dependent earthquake risk. This figure illustrates the number of buildings expected to sustain extensive damage or collapse in Nepal, should a seismic event similar to the 1934 Bilhar occur at different times in the future.

GEM: Thank you, very inspiring. My last question is not scientific. So the field of earthquake engineering and seismology is still quite dominated by male experts. There are of course very brilliant female experts such Helen Crowley, Tiziana Rossetto, Dina D’Ayala, Emily So, but still nowadays every time you go to conferences or scientific meetings there is still a gender unbalance. What is your recommendation for young female engineers who want to be as successful as you?

Professor Kiremidjian: I think the opportunities are out there to be successful. My approach has always been the following: do your work and demonstrate what you are capable of doing. If you demonstrate what is expected of you, then everyone will accept you. And I think today women are being accepted much more easily then they were 40 years ago. When I first started out I was the only one in my classes and the only one in all the conferences I went to. But I was also very fortunate to have strong supporters. The difference today is that women do not need supporters anymore, now they can stand on their own and they just need to show that they are capable, just like everybody else. Can we change the mentally of man? I think that that has already started, we do see more women in engineering classes, and I see in my classes that woman and man form working groups together, and work cohesively. I do believe that in the next generation we will simply see an evenly distributed engineering force. Women, men, and underrepresented groups. It is nice to see that this is starting from grass roots and it is growing in the various stages. I think women may still face challenges from some of the older generations, but the newer generation is much more open about accepting women in their ranks. Again, my recommendation is don’t even think about these challenges, just demonstrate what you are capable of, do your work, and nothing will stop you.