Transportation Infrastructure Risk Assessment on the context of Climate Change

0:01 I can see that people are still joining. I think we can we can make a start. Um hi, good morning everyone. Good afternoon. Uh thank you for joining us this morning um for this uh webinar by the UNESCO chair in disaster risk reduction and resilience engineering. Um in this webinar we explore the idea of uh building climate resilient transport by evaluating networks as interconnected systems um not as isolated parts. And in this webinar we have uh we are very pleased to have professor Si Yang with us uh executive director of integrated research on disaster risk redu disaster risk uh IRDB um to talk to you about her work on this area spanning over several years. So the

0:46 title of the top uh the talk today is transportation infrastructure risk assessment in the context of climate change. um uh with professor Yi Sini Yang with us um um uh she is a professor at Beijing Normal University and also co-chair of another UNESCO chair the UNUSCO chair in mountain disaster risk reduction and resilience uh specializing in infrastructure risk analysis and climate change adaptation and she has led over 20 research projects and PR as principal investigator and has authored more than 100 publications including in nature and she contributes her expertise as a member on the expert committee of China's national disaster reduction commission and UNDR Asia-Pacific Science and Technology Advisory Group and also

1:32 uh acting as uh editorial board of several international journals. Um I I extend a very warm welcome to you professor Siam uh despite the time differences and making this webinar happen joining us today. Um um before we proceed to the talk, I just wanted to let the audience know that uh we are taking a short break from the series of webinars for Christmas and we'll be back with uh a series of more webinars coming in January 2026. So please uh keep tuned on our website for information on the upcoming webinars. Um I uh leave the platform to you Professor Sini. Um please proceed. Thank you. Thank you very much Ashna. So uh do you

2:18 mind if I take some time to share my screen? >> Yes please. >> Thank you. So please let me know if you can see my screen because the the document is quite big so it may take some time. >> Okay. Can you see that? >> It's all good. >> All right. Thank you very much. So uh also I'd like to thank um our organizer inviting me to this UNESCO chair talk um and as the co-chair for UNESCO chair mountain disaster risk reduction and resilience I'm really happy to uh collaborate with our uh dear professor Dina my longtime friend um so I changed my uh the title a little bit but no

3:05 basically saying Bra resilient transportation infrastructure and the climate adaptation. The reason I I I want to make a smaller a very minor change is because I think only risk assessment is not enough. Uh the more important thing is for to improve the residence of transportation infrastructure and to to really work on this adaptation. So again um I'm a professor in Benjamin University and right now I'm also the exact director for integrative research on the disaster risk IDR which is the international science program co-sponsored by international science council as well as UNDR. Um well the the reason I want to pick this

3:54 topic is is because I was trained as a civil engineer before. um but also more about um this global efforts in DR really is a compiling force for me to to carry out this research in the past 20 years. Um because you probably know that in in 1990s um natural hazard and their impact become a global concern. And uh more importantly in the past 30 years the focus has shifted from disaster reduction to disaster risk reduction. And the the difference is that uh we are shifting from this post disaster rescue

4:41 reaction recovery to really better integrate all the information in risk assessment and be more proactive. Of course our best ideal dream is that there is no disaster only hazard right. So when we look at this road map we can see how the how this route was evolving in the past 35 years from the international decades for natural disaster reduction at that time is still reduced the disaster then of course you know very early we had the UN UNFCCCC uh is in 1994

5:29 uh we had the first award conference on natural disaster. And then we have the Yukohama strategy and uh uh of course in 2000 the international strategy for disaster reduction called ISDR was established. Um then uh in 2005 we had a second world conference on natural disaster. Remember that time it's all called a natural disaster. And of course in that year we had a skill goal framework for action stand from 2005 to 2015 and IDL was established 2010 and then uh in 2015 we had a third world conference on natural disaster again and in that conference we had the same framework for

6:16 disaster reduction um disaster reduction and that time it's it has been changed because previous in previous strategy is for disaster reduction but from send it's really for disaster risk reduction and of course in that year we have Paris agreement the SGS um and right now we're in 20 25 and in soon after another five years all this sandi framework on Dr. SGS will come to to an end. So actually I think right now it is a time for us to review what we have done in the past years and also to think about what is

7:05 the new theme post 2030. So that is how I look at it. Okay. Uh well, in case you have any question, please feel free to type that in the chat box or Q&A box. Um and I'll more than be more than happy to take all the question after the the talk after presentation. Okay. Um something I want to point out is that um we are facing a very different global risk landscape. Um first is that the earth's average temperature reached 1.5 degree and

7:50 actually uh well I think the UN or say the WMO has already declared that in 2025 many area has reached the two degree right but anyway in 2024 for the first time earth's average temperature climbed to more than 1.5 uh above the pre-industry level so this has already become deep evidence and this is much more than temperature. This has you know systematic impact. Another another factor we're facing is that we actually has already entered a new period of mega earthquake activity began in 204.

8:36 Um well this is the the global um we always know also notice that the global seismic activity uh is positively uh correlated with global volcanic eruption and it probably also know know that uh volcan volcanic eruption uh has a positive feedback uh to our climate. So this this facts just make things very complicated and uh so as a kind of as a result we see the changing hazards. So from the increased temperature it is

9:23 accompanied by the increased water moisture supply. So in most area we're having more extreme rainfall. We have melting ice and snow resulting large runoff. Flooding becomes a more severe challenge in most area. And also this increased the mass material supply. Uh the degradation of the cyrosophere reduced the stability of slope soil. In the in recent year you probably noticed that we are the globally we're facing more disaster channel or say compound disaster that usually is extreme rainfall uh then induce flash

10:09 floods then uh induce landslide debris flow um and of course know with more mass material supply we actually will increase energy the potential kinetic and thermal energy. Now from the ocean they provides more energy for mutological disasters such as hurricanes and uh and also the the masses increase this uh kinetic energy and overall we have this increased water heat coupling effects. the probability of concurrent occurrence of rainfall in a high temperature increases.

10:55 This has a lot of in influence especially for infrastructure. So, so overall we say uh climate change creates more favorable condition for hazard to occur. Um at the same time we are we are having this global change. Global change is much more than climate change. It's also also the the human activities. Human activity especially is reshaped risk features. So increase the human activity that means more housing, more mining, more deforestation, industrialization,

11:41 urbanization when they couple with changing climate. Um then we see of course there are more extreme climate events and with the rapid urbanization we see new vulnerability with urbanization. we notice that exposure is quickly uh increasing. Um and also when we look at this um further impact is that from the natural hazard sides usually it has longer duration they have larger and more intensive events and higher frequency. Take an example with infrastructure

12:27 system. Um you probably know this is a picture that from Dhaka city which is in 2012 or 2013. This one is not induced by landslide is mainly a technical hazard. And here this the the picture on the bottom this is New Orle you you probably know uh hurricane Katrina. Hurricane Katrina definitely this is the natural hazard but more importantly the disaster is caused by the failure of the dyke the dyke system. So yes, technical hazard is bring bringing more disaster because

13:16 urbanization just to bring more population to cities and on the other side extreme hurricanes, extreme natural hazards, they're just more fearsful, more damaging that they can cause a failure of infrastructure. And I want to share with you another example which is uh the 208 windstorm in China. This is also very interesting case that is a typical example of uh cascading effect in 208 uh in January in southern China because southern China used to be quite warm. It's it's the the the winter temperature

14:02 usually is above 25 degree but in 208 in January in in main the main uh parts of of southern China more than 400 400 uh sorry 4 million uh square kilometers area. We suffered strong uh cold wave in the four rounds of cold waves. And this four rounds of cold waves together with uh the precipitation of course that become the snow frozen rain and the snow the the accumulated huge ice on top of our electricity facilities.

14:51 And and actually the the electric wires are quite strong. But when they accumulate four runs of rainboard and ice, the load is huge. So to pull down this uh many major uh transferring wires. Um then this creates this blockage the the shortage in electricity and the in in about eight provinces in southern China. No this is bad enough right because in a southern people lost the supply of electricity but it's much more than that because in January end of January is the

15:37 spring festival season in holiday season in China. And uh during the spring festival uh season in China usually we will have about uh two billion um persons travel activities. You can imagine it's huge number of people accumulated in transportation hub waiting to be uh transported back to their hometown for the greatest holiday of the year. But because we lost the electricity then basically the communication system is dumb and also the transportation system is done. If without electricity then the

16:24 ice on the rail the ice on the road cannot be cleaned out. They cannot heat it heated it up. Right? So the ice are frozen on the rail on the road. So actually nobody can travel. So this cascaded effect in of course this has huge impact in economy and also this creates a lot of issue with social safety because so many people accumulated in transportation hubs and they cannot go they can go nowhere. So the loss is huge this loss is almost the same level as ventron earthquake you probably know that 8.0 magnetic earthquake in that year happened in May. So because of the year

17:10 you know in 20 I still remember in 2019 UNDR proposed that build to last uh which is to strengthen um the resilience of infrastructure and if we look at the global strategies uh like SDGs like Sai framework for disaster risk reduction they have very clear indicators for infrastructure resilience uh like SDG9, SDG11 and also even inserting they have something relevant and if you look at the standai framework the the seventh goal is reduce the damage to critical infrastructure and

17:56 disruption of basic services. I I like the I like the wording very much. I like this uh the goals targets very much is because it's is point out the importance of service. So it's no longer damage of critical infrastructure but reduce the d disruption of basic service provided by infrastructure that is typically also illustrated in the in the case in 208 in China. So we we can say that infrastructure resilience is the common focus of multiple global strategies. For those who are working on transportation infrastructure resilience, you probably know this paper uh published in 2019 in nature

18:41 communication by Cox and their group. Um the important message released there is that transportation infrastructure is facing systemic climate change risk and in numbers about 74 glo 74% of global transportation effects assets will be affected by one in 100 year floods temperature precipitation no this will shorten the life cycle of highway dramatically increase of maintenance repair cost. More importantly, transportation interruption cost is even higher than physical damage. So I think this is also very important conclusion that yes when we

19:26 are looking at the climate change climate change risk of transportation infrastructure. We need to look at this physical damage. We need to look at it associated maintenance repair cost. Also we really need to look at this interruption cost which is the loss from the function. Um and also you know from a risk governance perspective we know that we are we're facing the challenges for a more connected worlds like everything are connected with each other right um I guess you all familiar with this figures this are from this global risk report by the world economic forum

20:13 we notice that like many of these risks are are connected. So this bring us to a very interesting topic that yes we know that infrastructure connected with each other we know this risk connected each other then what is the best way for us to understand it. So with this I also want to talk a little bit about this the definition of resident infrastructure. I know there are like so many different definition of resilience especially for resilience resilient infrastructure but I think overall we can kind of summarize this resilience into four perspective the capacity to resist the shock the capacity that to mitigate impact and the

20:58 capacity to recover and restore and more importantly is if the infrastructure system can have the cap We have a problem. Uh, okay. >> She's connecting back. >> Yes. Si, you are muted at the moment. >> No, I Well, I I noticed that I I kind like I got disconnected. Can you hear me right now? >> Yeah. >> Yes. Yes. Yes. Yes. >> Can you see my Can you see my screen? Not yet. >> Can you share it again? Yeah. >> Yes. It seems that probably it's a

21:43 connection issue. I'm I apologize you know. Can you see my screen right now? This is the resent infrastructure part. >> Yes. >> Okay. Thank you. >> Uh well that's connect. >> We are seeing the first we are seeing the first slide and not in presentation mode. So >> Oh really? >> Yeah. >> Okay. So how about this one? >> Yes. >> Can you see that's thank you very much. >> This is Yeah. >> Yeah. Yeah. Sorry for the technical issue. But anyway, what I want to say that here I listed the four perspective that I think is very much relevant to how can we define the what is resilient infrastructure. It is much more than the physical resilience. is also about this capacity in mitigation in recovery and

22:31 also how can we better adapt to uh for better future adaptation for better future. So so correspondingly I think there are several key questions uh that we really need to focus when we are talking about climate resil infrastructure. So, so how can we better access the impact the risk of natural hazards in transportation infrastructure system? How can we uh better access the impact or risk in a interdependent infrastructure system? Then how to better restore and recover from a damage a damaged system and what are the potential adaptation measures and how to

23:17 improve the investment for transportation infrastructure system. Um never underestimate the last question because overall it's about money. if is it's all about like if investor is interesting in invest in adaptation how do they look at the cost benefit ratio so I know there are many project global activities but no say national activities in UK in EU in uh in in US in Japan they all have like infrastructure or in infrastructure safety uh projects um and I especially especially because since you know you are based in UK so I

24:02 want to really refer to I think quite an interesting report uh of called UK climate change risk assessment um and I I like this framework very much because I read it in 2018 but this report was released in 2017 and I see when I see this framework I I think it really uh it's a very nice uh summary because if we look at this climate uh risk assessment for infrastructure it has five parts. The first part starts from like the climate variables really to understand better understand the change of temperature, wind,

24:48 precipitation, sea level and from there we can better understand the climate extremes like heat waves, storm, flooding, surge, storm surge and then we can look at its impact or infrastructure as effects like individual components in the system. So for example, what is a damage to a port? What is the toppling of poly uh pylons? What is the the damage to a bridge? These are like individual infrastructures assets. And from here then we can look at the risk to infrastructure network because no transportation infrastructure can work by its own. It is it can provide a function because it's integrated in a

25:34 transportation network. So from here we can look at is reduced transmission capacity. Um we can in electricity system we can look at this reduced generation capacity loss of power and the road closure different when we need to look really look at not as individual component but as a infrastructure network. Then from here we can look at this infrastructure interdependency risk because as the as the example showed in 208 in China no infrastructure system right now can exist by its own. Even electric electricity system cannot because if any electricity system was

26:19 damaged you need to take the transportation network to go there and to fix it. So how can we better access this interdependent infrastructure systems risk? If all this are clear then we can go ahead for systemic risk to better understand its impact to the disruption of supply chains to understand its microeconomic impacts. So I think this framework totally makes sense and uh today I want to introduce some of our work starting from 2011 because back in 2011 in China we started 93 973 project which this is like the national key research project and I I have a sub project of it which is the the

27:06 resilience transportation the climate change risk assessment for transportation system and its resilience assessment and we almost take a very similar approach as in the 2017 uh UK assessment. Then we started from the the hazard side. Then we look at the physical damage. Then we look at the system damage which is like transportation system or coupled system. Then we look at the systemic impact which how it analyze how it influence the economics how it influence livelihood. then we go look at this DR strategies. So again um so starting from 2011 I start to work on different projects but mainly through the whole

27:53 through the whole framework starting from header analysis, exposure analysis, mability analysis, impact analysis, risk analysis. Then of course we did some uh recovery and restoration analysis, cost benefits uh benefit analysis for different adaptation strategies. We try to understand the system behavior in this compact system. Then from there uh we design the principle of DR and also design many very detailed projects for DR for hazard analysis. I think I can make this very quick you know this very straightforward right because for those who are working on climate change um

28:40 then usually is you need to have this projection data we need to analyze the extreme value we need to understand it its trend uh we need to analyze its uh v uh variation these are very different um perspective of of future climate variable, the average, the trend, uh the variance, um and extreme value for temperature, for precipitation. But more than that, of course, you can couple this um climate variable and really look at look at this combined coupled hazard, right? So this of course a lot of things can be done.

29:25 Then when you have this when we have all this um climate variable anal an analytical results we can start with exposure analysis for example like rainstorm uh the extreme heat wave and uh flooding the important thing for for rainstorms impact exposure analysis on transportation network it's not simply you overlay delay infrastructure with uh climate variables. It's more about what is the right variable that we need to pick. For example, when we are study this rainstorm um exposure

30:14 um um transportation network, we didn't look at all the intensity of rainfall. We actually look at two types of rainfall. One is this minor ramp for uh which will influence uh which will make generate this water membrane on the surface and the change of friction um coefficient and creates more probably more traffic uh traffic jam or traffic accidents. And another type of raintorm we're looking at is the very heavy one that really will uh reduce the visibility and create issue for transportation. And then when we are looking at extreme heat uh exposure analysis for extreme heat, we didn't use the degree like 35

31:00 36 because that doesn't make sense, right? And what what we really look at is a change of threshold because for real surface it depends on the materials of the pavement. Know the different materials will have different melting point right. So it can be 45 can be 50. So what we are looking at is really how this uh surface temperature will the days how many days will it will be to exceed this like a threshold because these are the numbers that really indicate the damage of the infrastructure. Similar thing is for flooding. Um if flooding of course we

31:45 can use many if we have the precipitation if we have this like hydraulic model we can you can have the flooding but more important question is will this flooding influence transportation infrastructure because for any infrastructure would definitely have protection right will have this uh usually the the road surface is three to five meters higher than the around. So, so this exposure analysis need to more focus on the actual need of transportation system. Um then the the third types of group of work we're doing is ink analysis

32:33 because for exposure analysis basically what I can get is you know how many mileage or how many bridges or tunnels uh will be exposed into a storm will be exposed into once in a 100 year flooding. But but we really what we we are interested in is what is impact. So what when is exposed to that what is the impact for example like like flooding? The more important question is actually yes if we have the rainfall and we simulate the flooding and how it will influence the transportation network. how many nodes or say road segment will

33:20 be inundated and then lost its function. So uh you know this is u this is a work we did almost like 10 years ago that we look at inundation on highway network. Uh there will see then actually there were generated three types of failure. One is a direct failure which is which means that the road was inundated by by the flood. Right? Then the other type is indirect failure. Say uh my this is a very long road. The one segment or one transaction was inundated and actually all the other linked uh linked um segment will you you cannot trespass

34:05 anymore. It lost its function of connectivity. Right? So all the other linked nodes they're called indirect failures and of course you add indirect failure and indirect failure failures you have this total failure. So in this way we will be able to as if we have this rainfall we can simulate the flooding then run this type of uh network analysis. So here uh we we did analysis for for Chinese highway system and the US highway system and the way we did it in this uh we compare the performance of the road network the highway network under three type of uh attacks. They are

34:53 random attacks um floods and the localized attacks. We pick the three types of text for a reason because random attack is is spatial distribution is totally random like ring war. No. So or more like traffic accidents. It can happen anywhere. Floods is very different because floods happens along the river and the river has a trajectory. So these are like a trajectory. And the third type localized attack it has a epic center and it has attenuation and it has impacted area. So earthquake is more like localized attack. So we looked at we observe of course we run uh

35:40 we we run like thousands of simulation um on the two highway networks in China and the US and we observe the the network performance and what we can see that is the random attacks uh will bring the highest the network function decrease and then is flooding and uh and actually for localized attacks is it's network system influence is the smallest and also if we compare in the Chinese highway system and the USA highway system US highway system we can see that the the impact for the Chinese uh

36:26 highway system is much more severe and of course it's a very interesting observation so we try to understand if in red that means the loss of function the loss ratio of function is higher. If it's in uh if it's in blue that means it can maintain its functional uh function and the system function and the reason is because the road network in China and in US has huge difference because in China our our highway system is highly concentrated in the the southeast coast. So it's a very much uh hetrogenic system. So very dense network on the

37:13 east coast and a very sparse network on the west. But in the US if we look at this network is more evenly distributed. So so that's probably the reason that uh why Chinese system is more vulnerable to random attacks. So this gives us a lot of hint to the planning of this ro network. Um so we also from here we notice that there are many like a tipping and a cascading effect. This is more like the the system behavior in a complex network. Um and uh and of course for this work we also formulate um well we we formulate a new model to predict when

37:59 the system will have this abrupt change. We call it tipping points. And the good uh the the quite a nice part is that we have simulation and we have u mathematical model just use theoretical uh get theoretical value and the simulation result nicely match with our theoretical value. If you're interested of course you can always check uh the papers um and then that is more like impact. So what is the the risk analysis for function from the physical damage to function? So to to study that we we took Hainan province as example. The reason I took

38:45 Hainan province is because it's relatively a closed system because it has a very clear boundary. It's a island, right? And this area has has like 10 to 20 different typhoon which is tropical cyclones every year. Uh and we have about like a 60 years uh cyclone trajectories. Um and also they have very clear uh highway network. So what we did is we collect the information and then we build up a very simple but effective framework that first we collect the the natural hazards which is tropical cyclone. Then we from there we start to uh develop the the the multicolor

39:30 simulation that use this multicolor simulation we can simulate 10 thousands of tropical cyclone uh that will influence highland province. The second step is we we define a functional uh fragility analysis which means for each typhoon we will uh for each typhoon will we will simulate it its progression and we uh uh we generates it windfield for the whole process and we run this multiolor simulation and you probably know that like usually for for highend freeway or highways they have road closure when the wind speed is surpassing certain

40:17 threshold. Of course in China uh that is is more about like 15 meters per second then if we can define this we will be able to uh define we can do some network functionality measurement. So how do we measure the function of this road network? It can be connectivity or it can be uh the shortest trouble time. It can be the delay of the network. It can be the transmission capacity of network. You can you can define your network functionality based on the need. Then with this we can go on with this

41:02 multiolor simulation. And if you can run this multiolor simulation for 10,000 times then definitely we can come up with exceedence curve. But here when we talk about the executed curve it's no longer the physical damage of the of the roadway. it will become the functional damage exceeds curve of the entire rope network. So, so that's what we get. So for from uh so here you can see that for different severity of functional damage we have associate with different probability after 10,000s of multicol 10,000

41:51 tropical cyclone simulation and from there we can see uh the exceedence probability and then we can calculate the return periods. Um, frankly speaking, when I look at the return period, it scared me because I I actually got this data from one of my friends. He he was working department in province. So I asked him, hey, what's going on? When I look at this research outcome, I noticed that almost every two three years uh the entire road network will lost to half of its function. basically is half of the the the highway network in Hanaman province were were kind of disabled. Is that the case? Then my friend told me yes that is a situation every every two three four at

42:38 most years then they will they will come encounter a very severe typhoon and at least the half of the road roadway will be closed and lost lost function. This actually has a lots of implication. So if we can have this this functional exceedence curve actually this is very helpful for us to design uh insurance program for transportation companies and it's it's very convenient uh to transer to transfer uh transform that into the investment what what is the size of losses uh functionally is associated with typhoon every here. So this gave a

43:24 lots of uh scientific evidence for their provinces uh DRR investment and of course I have to say that this type of framework includes very straightforward we can easily adapt it to other infrastructure system no matter it's electricity or communication we can use the similar framework uh to understand its functional um Then actually they asked me one question say that um yes we know that we're vulnerable to to tropical cyclone in in Hanaman province then can you tell me that you know but we have very limited capacity and very limited funding so can you tell me that if I have some money um where should I invest

44:14 now in reality investment or is always very limited so so actually we extend our priv service research to help them to decide that which intersections or road segments has the best value in the sense of reduce functional risk caused by tropical cyclone. And this actually is quite a typical research framework that for inability engineering, right? You can remove you can remove the where's the the previous stimulation system actually we can just play with the parameters you know if we can improve certain sections or intersections or roads basically that

45:01 means uh this road will have less damage less um interruption time then you can run the simulation again to see okay what if we invest in this couple intersections or rows and the entire risk how this entire risk function of risk can change. So this is also very interesting. But of course I have to say that if we look at different parameters connected connectivity based damage based um indicator or distance efficiency based indicator or travel delay based indicator they will have slight difference. So this all depends on this decision makaker's perspective that what kind of function do they

45:46 treasure most? Okay. Do they Yeah. Um there another question is about recovery right if you're talking about resilient infrastructure basically you need to make sure that when it's suffered from certain attack then it has to recover fast. So what is the best recover strategy? Usually we don't think that much but in reality it's also that we have very limited capacity in this uh like construction facilities and human resources. So there are many different ways to recover a damaged area and it also depends on what kind of damage is that it's a random random attack. It's a trajectory

46:34 you know like a typhoon or like flooding with trajectory attack then or it's localized attack like like the earthquake with the different types of external stressors our recovery strategy can be different. So here we we analyze the different ways you know local damage uh malicious attack damage which is we always attack the most important nodes of in the network or random or random attack. We we propose different recover strategy and later on we noticed that there's one strategy is the best with we call it um parapharical recover with with load weight

47:21 um this basically is that I'll recover an area from outside to the inside know and also I always recover from the most important note then go to the next important note in In this way, the whole systems um entire function level will increase the the the quickest and the largest. So again, we did that for highan province for typhoon and also for for earthquake for typhoon. We can see that within the the greedy search random repair greedy greedy repair uh the parapharic repair and the parapharac

48:07 repair with no weights we can see this performance is very very different. So from here we start I start to understand that uh it is necessary or if if this region wants to have more resilient increase resilient it's necessary for them to better understand the local attack type is more like the trajectory or more like more like earthquake the different repair needs different well planned recovery strategy and of course later on we start to synthesize the different assessment tool together to make it a system. Basically that means from micros microsposcopic

48:53 um simulation like car following behavior so that we can better understand the traffic in a city then to this microscopic uh traffic which is distribution of uh of transportation um freight transportation and passenger transportation in the nation. Um this work gave us a lot of things. Uh one important lessons we learned from this research is that we we notice there's something called emergence which means the behavior we observe from the the micro level say at individual level has a huge difference uh with this microscopic.

49:39 So this give us a lot of like systemic uh risk insights that uh if we are looking at very complicated system like transportation system, electricity system or coupled system we cannot just draw conclusion for from this microscopic. We have to conduct the different level of research microscopic, massoscopic and microscopic because the system behavior at different level varies a lot and this gives very different implication for uh the policies uh for the regulations

50:25 um in this design. And of course this system can also be used in adaptation accessory. For example, we talk about uh that the easiest way uh to protect uh transportation trans transportation infrastructure is you can raise the base the the road base right. Um so we we run we use a system the simulation system run very simple analysis then to understand what if if we can raise uh the road base for 2 meters or three meters what is the influence what is the impact and what is potential saving.

51:10 Um so this type of analysis has a lot of uh insights for regional or even national um climate change adaptation uh analytical work or strategy work. Um then from there we start to look at some other impact is like yes we know that infrastructure are are coupled with the tone. So let's look at this this performance of the system when this infrastructure are interdependent. So what we did you know basically is we want to understand the cascading effect uh in a inter in interdependent networks. So what we did is also very

51:56 simple coupled the highway system in China which is the the red lines and also then the railway system in in China because this two are heavily uh dependent on each other. They're not in independent. They're dependent on each other. If you really want to take the take the rail, you have to go to the rail station, right? Then you can travel from you first need to go take the rail highway roadway or say from your home to the station then take the railway then go to the place you want there are a lot of details here I probably don't have enough time to really

52:44 explain everything but the key message we learn here is that when infrastructure system are coupled with each other. This performance of this coupled system will be very much different from an independent system and also in a coupled system say for example a dense network coupled with some sparse network the response in this subsystem um are also has huge variation. So one thing also one short lessons we learned from this case study is we notice that when a dance network is coupled with sparse network and then if they are quite interdependent

53:31 then usually this sparse network will suffer more because a dense network they have more redundancy. though. So they can they they have higher tolerance. But if a sparse network is coupled with a dense network then the sparse network has very low redundancy and those cascaded shock to this sparse network sometimes can create abrupt change or say abrupt failure or say total failure in this sparse network. So this give give us a lot of interesting implication even more than just the infrastructure system is also very meaningful for social network and

54:17 also econ economic network. Um >> I think >> again there has been Yes. A failure maybe. >> Yes. Say >> oh sorry I lost again. >> Yes >> I lost again. But I I the the the thing is I'm kind of uh yeah finished the first part. Right. Okay. So, >> can you can you see my screen now? >> Uh uh yes, we up we could see it now. We cannot. >> I'm very sorry for this unstable

55:06 service. >> Yes, I'm back >> now. It's fine. Okay, that's >> okay. The I I just want to spend a very short time on talk about this adaptation. Um The previous work are all about you know the the risk assessment, physical assessment, functional assessment. Then later on we are really care about the different aspect of how to improve the resilience. But when we talked about breastly speaking we're always talking about the money we're always talking about the cost benefit analysis. So uh I I want to show you one example why this is important. So from here I

55:52 want to really emphasize the importance of economic modeling. Um this is also a real story that one day I was having co have a coffee with one of my friend and work with the ministry of transport. Then in suddenly suddenly he asked me questions says that know every year the the ministry of transport they purchase early warning service from the uh China administration of materology. Every year they they spend a lot of money like hundreds of millions hundreds of millions R&B every year for the service because early warning they believe that this is very important adaptation u measure. Then one day the

56:41 min the minister asked him a question say is it worth the money? Every year we we spend like millions of hundreds of millions in this early warning. then does this worth our money. So just from starting from there then they start to ask me question hey can you help me um better understand is this a wise decision you know if we don't invest money in this early warning system then definitely I can invest this money in many retrofitting projects or in some you know purchasing more facilities. So they gave me some data you know in the past and almost in the past 10 years from 2014 to 20 is 2023 uh the the highway the highway monitoring data. So

57:28 from there I will be able to see that what are the most important hazards uh ordered by total delay time by blockage uh mileage. You probably notice that all most of them are landslide, haze, rainfall, you know, geological hazards, snow form, these other things. So all this after highly relevant to early warning to meteorological hazards. So uh what we did what we did is we build up we set up a loss estimation framework. We use that data to better understanding what is the direct loss uh which is the

58:16 cost they paid for the service the loss of infrastructure the loss for the maintenance and then we also look at this make economic uh losses when you have the road closure you actually reduce the transmission of the freight and human this will have influence direct influence on livelihood and production. So if we can introduce all this information to a uh to a maker economic model such as IO mod output input model or CGE model then we will better estimate what is the the indirect loss. So after we finished everything,

59:02 they were very happy because they noticed that the maker level indirect economic loss from the road mill function was much more than the direct loss caused by the physical part. Um it's 10 times 20 times more than the direct damage. And uh when we look at this early warning and of course they give us a lot of data on like how this early warning system was working then we notice that the weather early warning service really can reduced losses and it can reduce more losses with benefit rate uh this exceed cost by ratio bonding 40 to 50. I was shocked by this numbers because some sounds sound too good, right? Um

59:48 then I checked some um world bank report and they report a similar number. This cost benefit ratio for early warning system range about like 20 to 40 to 50. Um so I I pointed this out is because I was trained as an engineer. Um so I was more on this fragility part or physical damage part but right now I notice that it's really necessary for us to really uh have this transdisciplinary uh research directly look at this function and look at this potential impact of functional loss in economic and this give us a more comprehensive or

1:00:35 systemic view of what is the real risk of transportation infrastructure system. Well, I I still have many other slides uh but I want to go quickly um for adaptation of course there are still many other ways know for hazards. I know that we can in the city um we cannot reduce the rainfall but we can use the green roof or change of land land or the land use planning to reduce the the ground um runoff. Um and for vulnerability we can protect the critical link for exposure we can use uh

1:01:24 early warning or flexible working time to reduce the demand. So all this are is a combination of tool in tool box. You can use this hard adaptation as well as like a soft adaptation. The more important thing is how can we better synthesize this different u adaptation measures to generate the the maxim to maximize the the benefit and uh minimize the cost. So we recently just I I I even forgot to put the link there. We recently just published this one in German. Um which would take Beijing as example. Um we we

1:02:10 use the the the hazard simulation tool and then we use the the sumo which is transportation model model and we introduced the green roof adaptation measure the link hardening measure and also work from home flexible exposure adjustment measures and around this cost benefit analysis and u also it's I just skip this details Um I just want to say that it's it's it's really quite promising that a nice combination um of the sty measures can greatly reduce the damage can can generate a huge economic

1:02:56 um incentives. Um the and actually I I the reason I do this research we in we did this research about five years ago is also because of the real demand when the city are facing more challenge with extreme rainfall and more inundation they're looking for what is the best way for them to proceed. So this type of we give them more insights. So overall I just want to say that you know I think the challenge for engineers changed um we we need a more integrated research that's also part of reason that I I became the CD of IRGR because we really

1:03:43 need we need to shift from avoid to adapt know we need to shift this delayed damage to transform the way of uh operation or the way of governance. So, so I really looking forward to more collaboration on this type of research to promote more transform. Um I do think from this perspective uh the collaboration of our UNESCO chairs and also the engagement of our young talents uh will bring huge potential in the future. Thank you very much.

1:04:30 So with this I just finish my presentation almost in time and I just to take the time for questions. Thank you. >> Thank you very much, Sini. Um, an incredibly stimulating u presentation. I have at least I've written down at least five or six questions that I would like to ask you, but I let the audience go first. And our first um speaker, our first uh person in the audience is Ali Osman Enel. Um, we will be allowed to talk if you can open your mic please and introduce yourself. Ali,

1:05:20 can you hear us? >> Also, >> yes. >> Yes. Yes, I can hear you. >> No, this is me. But we cannot hear we cannot hear Ali. Yes. So I can read >> I can I can I can read. Yeah, I just got my glasses. I just got my glasses. So now I can >> So what does the resilient city looks like in 2030? >> This is very broad question. >> Yes. um resilient city I think for for me is even more complicated than resilient infrastructure because infrastructure itself is just it's difficult enough right like how can

1:06:05 we have better can reduce the shock you know or basically with the same intensity of shock we have no impact or if it's over the shred then we can reduce the negative impact then of course how can we better recover restore the function then how can we better form this uh capacity to adapt to the new challenge what I'm thinking right now is like for for society because when we are talking about city it's it's much more than infrastructure it's also about the environment the institution um the technology um more importantly the people so But my

1:06:51 my understanding is that these dimensions are pretty similar which is a four dimensions. How can we resist the shock? How can we mitigate the impact? How we better or faster recover restore the function and how can we build up this adaptation capacity in long. I think this are similar but uh I just want to say that I think for resident city is it's much more complicated. I know that many cities are are walking towards that. I know that uh you know you probably know about this next resin 2030 projects globally there are more than 2,000 cities have their own

1:07:37 strategy right but I think the four dimensions I mentioned they are in common all countries or cities are looking forward to some progress in those five dimensions just the detailed uh scheme actions toward the four dimensions are different like I see very common you know goals in this four aspects. Yes. >> And Ali is also asking what are the barriers to increasing international cooperation for disaster risk reductions and what role can universities play in testing and refining resilience models. I think this can go together. >> Yeah. I think I ask a very

1:08:25 big questions. Yes. >> Um yes. Um my understanding is that disaster risk correction itself it's it's international uh though we see some local disaster but something like climate change there are just a globe right and of course if you look at like Europe or Asia you see many transpoundary issues. So I think naturally Dr. are uh welcome this international collaboration. I think that the current issue there are many ways you know international cooperation can can proceed the sharing of data. It can be tricky but you know it make a lot of sense because extreme

1:09:12 events is still rare. No it's just unwise to let one city to suffer from all kinds of extreme events. is better that we can learn from the extreme events in other city and then we can get the lessons we can get experience and get prepared. So for me this is already enough incentive for us to to promote international collaboration on VR and of course uh from there we can look at the collaboration on modeling on on methodology right and also we can look at uh the sharing of techn products one example can be targeted the AI in recent years um almost all of the man

1:09:59 has to embrace AI my group is also working on that today I didn't have time to introduce them but when we are we try to set up like AI model for Dr. Then when we're doing that, we notice that uh we just cannot do it by ourself because um because the resources too huge. Um it's just need to synthesize the different data resources, publication resources, methodology resources together. then we can utilize the computing uh resources to to learn from this no computing infrastructure to learn from this resources. This also need collaboration and of course when I

1:10:46 say like learning from each other this sometimes also means like learning from each other's executional progress. I think this is also a very cool uh perspective for international. Um I think another thing I can think of or encourage international collaboration on Dr. R is capacity building because since last year I became the ED of IRD. I have visited to many different places. uh so now I I have a very clear understanding that the capacity in different region or country varies a lot. So >> I think international collaboration can start from there. You will bring the capacity in developing country or develop develop

1:11:33 countries to developing countries to those areas who are which are familiar with one hazard to those areas are not familiar to to those hazards for example like heat waves. our our tropical countries actually they have lots of experience in coping with heat waves but uh in northern countries side they have no experience right so many of this our what we have learned in the past can be shared with different region so I I see huge potential there and I remember this the second question uh is regarding well admit dismiss Um

1:12:21 about it's about the univer about the role of the >> university university. >> No okay >> in developing point models. >> Yes. Yes. I think for universities you know the cool part of university right now I have a is our students come from different places. they actually bringing different local knowledge to a research institute. University is also a racial institute and uh um in this way and and also university is more neutral agency you know because it's will it will be much easier for universities to

1:13:08 collaborate with universities. it will be easier for university you know students to collaborate with different universities. This is a very unique advantage of university but also in some sense it's also a disadvantage of universities just because our students come after three four five years they graduated but but if you really want to work on this modeling stuff really work on this um no tools tool development it needs to be continuous right we we always keep progressing. So sometimes it's an issue for me right now is like uh the former students uh codes cannot

1:13:54 be easily understand by the new students. So so this an issue but the good part is the students they from different areas they really brought new information even for me I think globally that's also the case. So I think I think it's very important to help the university students to collaborate from different universities from regions and from there probably we can have some interesting models and interesting tools. So this how I look at this question. >> Thank you very much. And there is a question from Lakshman Shriant. Uh Lakshman if you would like I will open your microphone if you can uh introduce

1:14:41 yourself please and um Laxmanor from >> Yes I can hear you I can hear you >> thank you so much >> can you tell us a bit about yourself and then your question >> yes yes I'm Lakshman I I work with Delta as part of the safe and resilient infrastructure department and work a lot with Elco Cox. >> Okay. >> And uh with regards to and mainly we focus on transportation infrastructure resilience and I was really curious about your presentation and especially the criticality importance impact analysis that you did phenomenal work first. Thank you so much for sharing. But I was wondering would the impact or

1:15:27 the criticality that you assess would it change between a passenger transport and a freight analysis? Uh yes uh yes uh definitely it has uh difference if I I'm using free transport or passenger transport because if I'm using the free transport uh I may uh more I may look at this um this delayed time and turnover because uh you know transportation of of goods basically is money right it's business so people really care a lot about uh because if there's delayed they're going to be penalty. So when I define this function, I really can define the function keep

1:16:15 function variable indicator as the delayed time loss. No, the the increase delay time or the time loss and uh uh for uh for passenger uh transport. Um this of course is one way but probably you need to care more about the safety itself the physical damage itself because the sim the physical damage sometimes indicate uh this personal person damage right because this is something that people really cares more and also uh for passenger uh for passenger uh transportation system sometime we really care about accessibility. So can can this network

1:17:00 provide the basic service to the people living there. So my point is that know when you are uh we are analyzing this type of uh criticality or impact analysis uh you can always you need to have a very clear goal. So what is your research object? it's a it's a highway itself or you're looking at the freight transportation on the highway or you're looking at the passenger transportation on the highway then in this way we need to have a better understanding of uh what is the real demand there to care about the passenger safety most or are we looking at the the function as a as a as a government I may

1:17:48 look at the function if if my my transportation system can provide accessibility to the people, right? But if you are uh you are doing this consulting work for a for transportation company, then they they care about the loss the delayed time at the loss of uh turnover because that's uh directly related to their uh benefit or or profit. So I think as if we have a clear picture of the object of the demand then we can better pick uh the the analytical variable that we want to analyze. >> I'm not sure if that's clear. >> Yes. Thank you so much. >> And now we have Victoria Posito again.

1:18:37 Victoria, if you can open your microphone and introduce yourself. Victoria. Hello. Well, there is there are two part questions here. Um the first one is um that uh from uh an an issue on data. So um if you can expand on how you deal with data gaps because very often you have um maybe enough data on one of the parameters that you use to determine the impact but not on others. Um and

1:19:23 then well there is a question on fragility curves from uh literature. Um I will I will write I will um read that later but also whether you are using open data or you are using uh customary data saying it >> okay so um uh actually um in the past one hour I introduced the many projects uh you probably can can tell that they are from this this research topics come from different um come from different person or different agencies. Yeah, >> sometimes yeah if I work with the ministry of transport they usually for

1:20:09 example the cost benefit analysis that my my friend was directly from the minister and they are facing a real problem from the minister right so in that case they will be able to provide me uh the real data and uh it's it's very complete data set and of course even when I did that research for them they give me this uh the roadway monitoring data. They gave me uh the because they have very very very nice data inventory. They have each event how much they spend on it and how long did they process it? What is the blockage time? What is the blockage mileage? They have very clear information and also I need to match that with like a mutological data or even hazard data. In

1:20:56 this way I have to coordinate with many of my friends in other industries to to match the data. This is a lot of work. Um though sometimes we just you know show it in one slide but actually this this type of work probably cost me most of the time. This is definitely end issue. Uh because it's very difficult for us you know for me as you know doing a research is more like a make a dinner. So if you want to have a nice dinner you're going to have nice ingredients. only if you have nice ingredients you know then you can say with with this ingredients I have then make I maybe can make this dish that dish right and then they're really writing the paper is more about the plating

1:21:44 so so sometimes yes they will say okay I have a question for you sign I have a question for you then I will listen to their question say ah they want to they want a snack or they want fast food or they have a nice dinner or they have a bandwidth. So from there they will I will always ask them hey what kind of data do you have? Then of course I will with with the data I will try to match with other data resources. Then I will check my the data inventory to see okay these are the the inventory I have. These are the ingredients I have. from here based on what they want maybe a pizza or pasta or they want a fine dine they say okay I make this dish or that dish so that's how I look at it okay

1:22:31 [laughter] >> I like I like the good analogy but uh it is also true that uh it's not just like going to the market and shop around it is true that sometimes getting the right data takes quite a long time and a lot of uh uh um negotiations even with your client even though they might be the people that you know are actually paying for your research they might not be ready to disclose. >> Exactly. >> Yes. >> Exactly. So uh the I I I make the analogy and say that you know we we we cannot say that I really want to do this research so I have to get data in real in reality is sometimes you don't have

1:23:17 that data. So it is very important to look at what data can you have and this collection of data the cleaning of data the the preliminary analysis of data take very long time a lot of energy. So yeah we all know that right? Yes. Yes. There is and there is an issue on there is quite of a a tension let's say between the fact that uh as a researcher we are increasingly being asked to solve these kind of uh systemic problems. But then the stakeholders at the other end they are not prepared to open their data sources for us to actually properly develop robust models. So there is a

1:24:05 tension there that probably you know needs to be resolved maybe by uh international organizations as well as by the uh ethics of research. The second question that Victoriao uh um asked was more specific to the adaptations and it's quite an interesting question because he's stating you know very often you are we are using fragility functions when we when we consider the system that are used that are taken from literature because again in the economy of the project you might not be able to develop

1:24:51 every single fragility functions. So how do you modify that and how do you robustly assess the effect of the adaptations on the fragility functions that you have adopted? Um I think uh yes number one um yes um in most of my research I just adopt the fragility curve from literature. Uh this uh this probably is the case for most of our research because fragility curve itself is very expensive. Right? gradulator carbs usually comes from either experiments or numerical

1:25:38 simulations. Uh and of course sometimes it can come from uh statistics. So what what we do usually is when we adopt some fragility curve, we'll use the fragility curve to to do some estimation. Then we find one real real world case a big test this estimation where is the real losses if you know there's no big problem I I can feel comfortable that yes we can use so this is how we proceed and for adaptation when for example we can know what if I will to raise the the one is more about a scenario it's a scenario what if question is a scenario say I can

1:26:26 say that okay I adopt this green roof strategy um then this green roof strategy I can see that if I adopt it for one square meter what is the cost so in in a in a region or in a city maybe know I can have I can have maybe like not six million square meters I can talk to the strategy then what is my cost and with the scenario I can anal analyze that what is a reduced flooding what is a reduced impact and is There's a TR you know cost beneficiary activity adaptation measure. This is how we look at um so my understanding is that for adaptation analysis most of time we're we're making this

1:27:11 scenario comparison not the exact time not the exact number. So basically is for adaptation you can compare the with without you can look at a different combination to see which kind of each combination or if with without it's a scenario what we are working on with with this analytical framework basically is is running this scenario comparison and pick the best scenario and as that

…