Climate Change and Health Seminar: Attribution of Health Impacts to Climate Change

October 21, 2021

Information

Dr. Ana M. Vicedo-Cabrera, Institute of Social and Preventive Medicine & Oeschger Center for Climate Change Research, University of Bern, Switzerland

Dr. Vicedo-Cabrera joined YCCCH for this monthly seminar to discuss her work on climate change and public health, and the Multi-Country Multi-City (MCC) Collaborative Research Network.

October 11, 2021

ID7061

To CiteDCA Citation Guide

00:00<v ->Hi, everyone, as we have long been</v>
00:03waiting for this time,
00:05Dr. Ana Vicedo-Cabrera will talk about
00:09the most advanced sciences in how to attribute
00:13the health impacts to climate change.
00:16Also Vicedo-Cabrera's current research
00:19develops along the intersection of epidemiology
00:22and climate sciences to understand
00:25how different climate factors
00:26and other related environmental stresses
00:29affect health in the context of climate change.
00:32So she has many, many excellent publications
00:37in the climate epidemiology field,
00:39including using the Multi-Country Multi-City Network
00:43data to look at the health impacts also,
00:46and the health impact to heat-related mortality
00:49including the one study
00:50that she will be sharing with us today.
00:54But before we hand it over to Dr. Ana Vicedo-Cabrera,
01:01I want to just mention that housekeeping rules
01:06that if you do have any questions,
01:08especially our online audience,
01:10please feel free to type in your question in the Chat box,
01:14and we will have all the questions answered at last.
01:19So thank you, and without further ado, Ana.
01:23<v ->So just give me a second, sorry.</v>
01:34Do you hear me?
01:36<v Host>Yeah, okay.</v> <v Attendee>Yes.</v>
01:37<v ->Yeah, perfect, so welcome, everybody,</v>
01:40and thank you very much
01:42for being here in this webinar.
01:44And of course, thank you for the invitation
01:46to contribute to this event today.
01:49It's a great pleasure for me being here
01:52to talk about a topic that in a way,
01:54has been a bit of my nightmare, I must say,
01:58or a bit of my priority during the last,
02:01I would say, two, three years.
02:04And I believe that it might be one
02:07of my main research fields in the coming years as well.
02:13So I hope that basically, at the end of my presentation,
02:18you might have already some insights about this topic,
02:23and probably you might get inspired as well
02:28about the specific topic of attribution.
02:31So as you could see from the title, we'll talk about
02:35"Attribution of Health Impacts to Climate Change",
02:37and now, you will see that mostly of my presentation
02:40will be focused on heat and health as an example.
02:44So let's start from there, from the very beginning.
02:48See if I can, okay?
02:49Yeah, so basically, heat is considered nowadays,
02:57an important environmental stressor.
02:59Very recently, it was estimated that
03:01around 1% of, sorry, 1% of all deaths globally
03:10can be attributed to heat.
03:11That translates around seven deaths
03:13per 100,000 population per year.
03:17And as well, it has been estimated
03:19that increase in morbidity, in particular,
03:22for cardiovascular, respiratory, and mental disorders.
03:26And as well, it has been identified several,
03:29let's say, vulnerability populations
03:35such as elderly, pregnant women,
03:37chronic patients, and children,
03:39and finally, it has been estimated an important burden
03:42in terms of economic costs,
03:44and also reduction in labor productivity.
03:46So every signal is taken together is considered
03:49that heat is an important element
03:52to be assessed in our field of climate change,
03:56epidemiologic impact in particular with regards,
03:58let's say, climate change research
04:00and what is gonna happen in the next decade?
04:03So let's say, how heat affects health?
04:10In a way, we can see that first,
04:14the mechanism by which heat likely
04:16impacts human health are complex
04:18and understanding, let's say, how the body
04:20reacts to heat has been the focus of decades of research
04:25in particular, in physiology.
04:28It stresses kind of the overarching term
04:31that is used for, let's say, to describe
04:34the response of human body to the exposure to heat,
04:37and usually happens when the body
04:40is overwhelmed by metabolic heat production.
04:42You can see here from this diagram,
04:44basically, our body, what it has to do
04:46is kind of react to the exposure to heat
04:48in several mechanisms to try to dissipate,
04:51or let's say, emit the overheating
04:52that we have in our body.
04:55And if in a way, this is not efficient,
04:58what we can create or let's say, can cost
05:00or through different mechanisms that eventually
05:03can damage different systems or organs in the body.
05:07In this nice review that was recently published,
05:12they summarized most of the areas,
05:15so the different mechanisms they have seen
05:16that actually, there are several,
05:19let's say, several organs that are directly
05:22affected by heat exposure through different mechanisms
05:25such as ischemia, heat cytotoxicity,
05:29inflammation, et cetera.
05:31This is from, let's say, physiological
05:34mechanism important view.
05:35But if we, let's say, us, epidemiologists
05:38working on climate change research,
05:40we know that we assess heat in a kind of different way.
05:44Basically, what we do is to have our wonderful,
05:47or let's say, so called exposure response functions,
05:51that in a way summarizes the association
05:53between the ambient temperature in this case,
05:57we do focus on heat, and specific health outcome,
06:00that is in this case, mortality.
06:02That basically, you see here
06:04in the y-axis is the relative risk,
06:07and in x-axis is the temperature.
06:09In this case, representing this, let's say,
06:11association that usually is non-linear for the City of Rome,
06:15and how we define heat is basically all temperatures
06:19above a specific threshold that we call
06:21temperature of minimum mortality
06:23that is tightly-defined from the curve
06:26that corresponds to the temperature
06:28for which the risk of dying is minimum.
06:30So every temperature above
06:32this threshold is considered heat.
06:35And we know that risk increases steeply
06:38from this point onwards, as you could see here
06:40in the curve, up to a maximum.
06:43So basically, this is how we assess heat,
06:46the effect of heat on health.
06:49But to make our life a little bit more complicated,
06:52we know that actually the effect of heat
06:55is very different across locations.
06:57I mean, we can assess this expression response function
07:01or curve from a specific population,
07:03but we know that we cannot extrapolate this curve
07:06to other locations because we know
07:08that vulnerability is very specific,
07:11very particular for a specific location.
07:14It's mostly because of the different combination
07:16of let's say, factors, so vulnerability factors
07:19or resiliency factors that make this population
07:22more or less resilient to an ambient temperature,
07:25in this case, for heat.
07:27So in a way, you could see that during the last years,
07:31there has been a lot of, let's say,
07:34developments in the field of climate change epidemiology,
07:37and to clarify how heat affects health.
07:42But if we have to define specific moments in time
07:45in the past but in a way constitute an important,
07:49let's say, kind of turning points,
07:52how we assess the effect of temperature on health,
07:56in what particular in public health
07:58is this event that happened in Europe in 2003
08:02is this massive European 2003 heat wave
08:06that affected very heavily Central Europe
08:11on the Southeast in over the Mediterranean area.
08:15And basically, what happened is that
08:18it was so massive, so unique that made
08:22that everybody turned their, let's say,
08:25focus, in particular, on public health on heat.
08:29Actually, a few years after that,
08:32there was an assessment in which they estimated
08:37that around 70,000 deaths could happen
08:44during this massive event.
08:46So in a way, it gave a kind of very clear idea
08:50about the severity of the event.
08:52And more importantly, what's at that time
08:54is really to say, "Okay, if this is happening now,
08:57"what would happen in the future?"
08:59So we know that probably due to climate change,
09:01these events will be much more frequent.
09:04So while the epidemiology is, let's say,
09:07as we were assessing what were the impacts
09:09of that event at that time, in particular,
09:13try to understand how we can implement,
09:15probably, health measures,
09:16how we can protect population from future events,
09:20climate science community, they were more thinking,
09:25"Okay, and what could have, in a way,
09:28"what was the role of climate change in this event?"
09:33In a way, there's this kind of very,
09:37not very good question,
09:39or let's say, an imposed question
09:41whether this 2003 heat wave was actually caused,
09:44in a very simplistic or deterministic sense,
09:47by a modification of external influences on climate.
09:49Basically, as we said it, due to the increase
09:52in concentration of greenhouse gases in the atmosphere.
09:55Because we know that almost any such weather event
09:59might have occurred by chance in our world
10:01without climate change.
10:04In a way, what we have to do is to think
10:05that in another perspective,
10:07or let's say, to put that question in a different way.
10:10Basically, it's how much did human activities
10:13increase the risk of occurrence
10:16or probability of this event?
10:17Or more specifically, did actually climate change
10:21alter the severity, frequency, or duration of the event?
10:25So this is exactly what attribution
10:28and detection studies does.
10:30It's a field that it has been, in a way,
10:33developed in the last years,
10:37in a way, it's more traditional
10:38from the climate science community,
10:40but not much on the epidemiological side.
10:43In a way, one of the first example
10:45was actually this study led by Peter Stott from the UK,
10:49in which actually they assessed
10:52what happened in this 2003 heat wave,
10:54and what they came up from this study
10:56is that it is very likely that human influence
11:00has at least doubled the risk of a heat wave,
11:03exceeding this threshold magnitude.
11:05So in a way, it's already posting,
11:07putting a certain name into this event,
11:10saying that probably climate change
11:13have altered the, let's say,
11:15probability of the occurrence of this event.
11:18So how this attribution and detection studies work?
11:22So basically, in a very simplistic way,
11:25what happens is that we model, let's say,
11:30we compare our current climate
11:32in presence of climate change
11:35that we can actually, in a way, estimate
11:39or let's say, mimic or get simulations
11:42based on these kind of climate models,
11:45in which they kind of try to mimic current conditions
11:49based on the what we know
11:51in terms of greenhouse gas emissions,
11:54and we compare it with a world without climate change.
11:58So basically, it's what you can see here.
12:00We compare it, here is in this curve
12:03is kind of simulated data just for you
12:05to illustrate this comparison
12:07in which we have our warming climate
12:10that is increasing in red
12:11compared to a kind of climate-free,
12:14climate change-free environment
12:16or what they called a naturalized scenario,
12:20or let's say, without anthropogenic forcing.
12:24So in a way, the difference
12:25between these two scenarios would give us
12:28what is actually the contribution of anthropogenic forcing,
12:34that eventually is what we want to know,
12:35what is the human influence in the climate
12:37that actually might have altered the climate
12:44in current period or historically,
12:45during the last decades.
12:47So as I said, it's a field that has been developing
12:53in the last years in particular
12:54for the climatological statistics
12:56because of course, we know what is happening now.
12:59I mean, we can see whether these simulations
13:03from these climate models really mimic
13:06what we are experiencing today based on observations,
13:10but we don't have data what would have been
13:13the world without climate change.
13:14So in a way, we have to rely on these models
13:18that eventually, when you reduce this,
13:20let's say, forcing, so these inputs in your model,
13:23you're actually mimicking what would have been
13:26the world without climate change.
13:28So in a way, you can see that
13:30there are a lot of uncertainties.
13:31And of course, one thing
13:33that I would like you to put in your,
13:36kind of your front, in your forefront,
13:39is that when we talk about attribution
13:41and detection studies,
13:42we talk about, basically, probability.
13:45This is a term that you will say
13:47that it's very pivotal in this story
13:50because in a way, it's not a matter of,
13:52okay, yes or no climate change
13:54have caused this event.
13:57It's whether climate change
13:58has altered the probability of this event.
14:01So basically, what we, let's say,
14:05people working or researchers working in,
14:07in climate science, mostly on the part
14:10on attribution and detection studies,
14:12what they do is to compare probabilities.
14:15As you could see here in red,
14:17the probability of an event happening
14:19above a specific temperature threshold
14:22compared to the same, let's say, threshold,
14:24what would have been the probability
14:25in a world without climate change
14:27in this counterfactual scenario.
14:31Let's say, in this nice review,
14:35this researcher, Fredi Otto from Oxford,
14:40she very well described in this paper,
14:43in this review, in which basically,
14:45what she said is that out of this exercise,
14:47we would have four different outcomes.
14:50First, could have been made, let's say,
14:53that this event could have been made more likely
14:55because of the anthropogenic climate change, let's say,
14:59or it could have been made, let's say, less likely,
15:03or there is no detectable influence
15:05from anthropogenic climate change.
15:07And the last one, with our current understanding
15:10and available tools, we are unable to analyze
15:12the role of external drivers in this event.
15:15So basically, as you could see,
15:17is to see whether climate change
15:20altered the probability to make this more or less likely.
15:23And then we have resources,
15:25or let's say, our models can help us to really clarify
15:29whether these differences in probability is meaningful.
15:33So to understand the world, this work of attribution,
15:37is we have to talk about,
15:40and usually, you will see this plot
15:43in every study in this field.
15:46For us, epidemiologists, it's basically
15:49something that we, I mean, honestly,
15:51when I saw it, I don't know what the hell is this,
15:53but at some point, I found it very, very interesting.
15:56So I will try to guide you through this plot.
16:00You could see here in the y-axis
16:03is the monthly temperature equivalent.
16:07And basically, in the x-axis,
16:10you have what they call return time.
16:13It's a measure of, let's say,
16:15of probability or severity of an event.
16:17It's basically, you put in here in this point,
16:20what it's saying is an event
16:22has a return kind of time,
16:23or basically, it already happens 1 in 10 years.
16:27If you go in here, it's an event happening 1 in 100 years.
16:32So in a way, you could understand
16:34that the far you are from the origin,
16:38the more extreme an event could be
16:41or the less likely, as you could see here
16:45on the other side of the axis,
16:47less probable an event is.
16:49So basically, what they do, in this case,
16:51it's an example of the 2010 heat wave in Russia.
16:55And it's basically to compare this blue line,
16:59that in a way, is used based on this counterfactual world
17:03in which you have, basically,
17:05this is the probability of a specific,
17:06let's say, event happening at different times.
17:09And in red, the same,
17:11but in our world currently, in our world,
17:13let's say, in our world current conditions,
17:15so we don't have to worry.
17:16So if you go in this line,
17:19this line is basically corresponds to this threshold
17:25that was, let's say, defined during this heat wave.
17:27But basically, this heat wave reached,
17:29I think, it was 24.5 degrees.
17:33So in a way, according to these dimensions,
17:36what they say is that in current times,
17:40this event corresponds to approximately 1-in-50-years event,
17:46while in a world without climate change,
17:50this event would have corresponded to 1 in a 100 years.
17:54So basically, what has happened
17:56is that climate change has made the event more likely,
18:00let's say, from an event in a hypothetical world,
18:031 in 100 years has become 1 in 50 years.
18:07So as you could see again,
18:10is the EDL changes in probability,
18:12making a specific event or a specific temperature threshold,
18:16make it more probable, let's say,
18:18going from 1 in 100 to 1 in 50 years.
18:23So as you could see, it's kind of something
18:28that is not very, for us as epidemiologists,
18:31a little bit difficult because in a way,
18:33it's just talking about probabilities,
18:35but for us, translating this into health impacts,
18:38it requires a bit of work.
18:40But let's say that we're working on it.
18:42The idea is that this work of attribution
18:46has gained kind of a lot of attention
18:49during the last years.
18:51In particular, thinking about
18:52what happened this last summer.
18:54Surely, you might have heard,
18:56or even lived there, or suffered this event,
19:00this massive heat wave that happened in last summer
19:03in West, North in America.
19:05So in a way, what happened,
19:08we know there were few days,
19:11the group came with temperatures above record,
19:15and it was a lot of attention in media, et cetera.
19:18So while all this was happening,
19:21let's say, that there was an initiative
19:23from this World Weather Attribution initiative
19:27that it's a kind of, again,
19:30it's an initiative in which different researchers
19:33working on attribution studies put together
19:35and try to give answers about whether climate change
19:39might have had some role
19:41in a specific extreme weather events.
19:44Not saying that to provide this evidence
19:48a year or two years after,
19:49it's really to provide this evidence in the coming weeks,
19:53because we know that times matter.
19:55If we have suffered a heat wave like this,
19:58it would have much more impact
20:00if this answer comes earlier in time
20:05rather than wait years ahead
20:08that people might completely forget
20:10about the severity of this event.
20:12So the idea is that this group of researchers
20:15and they did this analysis,
20:16and basically, what they came up with this
20:19is that it would be virtually impossible
20:22without human-induced climate change.
20:24At this event, it's estimated to be about
20:271-in-1000-year event in today's climate.
20:30So you can have from this sentence,
20:33that this event was kind of unique,
20:35very extreme, 1 in 1,000 year, it's a lot.
20:39And actually, they provided this plot.
20:40You will see that it's very similar
20:42to what I just shown.
20:43And actually, this event that was in here
20:47around 40, almost 40 degrees,
20:50they saw that it was actually even outside
20:53the probable range of events within a year,
20:57within let's say, in our current climate.
21:00So in a way, it's saying already about
21:02this huge severity of this event that happened.
21:05So this study and the savings they provide at this time,
21:10I said it was a couple of weeks after the event
21:14was very, very powerful because it's giving clear,
21:19let's say putting the finger into the idea
21:22of the role of climate change
21:24and the human influence in this event.
21:26So the message was very, very, very strong.
21:29At the same time, we have to bear in mind
21:31that surely you know that there was this,
21:34the new report of the IPCC,
21:38that part on "The Physical Science Basis"
21:41was already published in August.
21:43And the difference, let's say,
21:47of this report compared to previous reports,
21:49really to put in more weight into the influence
21:54of the human activities on carbon,
21:56let's say, extreme weather events
21:59happening in today's world.
22:02So again, you can see that the idea of attribution
22:07is gaining much more attention lately,
22:10but more importantly, because we know
22:14that evidence from attribution studies
22:18can be used in lawsuits.
22:20Basically, has been used for a specific,
22:23let's say, companies, individuals, et cetera,
22:27to kind of give some complaints,
22:31or let's say, to ask for some compensations
22:34of a specific losses to, let's say, governments
22:37or companies emitting greenhouse gases.
22:40So the idea is that during the last years,
22:43it has been a tremendous increase in,
22:46let's say, different lawsuits
22:48that have been implemented against climate change
22:51using evidence from attribution and detection studies.
22:54In this plot, you can see that it actually
22:56was mostly during the second half
22:59of the previous decade that actually went super up.
23:02And most of these, let's say,
23:04these initiatives happened in the US.
23:08Most importantly, it's like, okay,
23:12we know that there's this tool,
23:13but we need scientific, robust scientific evidence
23:18that could help us to gain or let's say,
23:21to win these different, let's say,
23:23initiatives in the courts.
23:25At the same time, we know that,
23:28let's say, that the idea of these initiatives
23:31is that beyond individual litigant,
23:33but it is seek to advance climate policies,
23:36drive behavioral shifts by key actors,
23:39and or create awareness,
23:41and encourage public debate.
23:43So it goes beyond the idea of compensation.
23:45It's really to gain more weight,
23:48or let's say, to put more emphasis
23:50on the role of climate change on the different,
23:53let's say, events, extreme weather events
23:54that are happening.
23:57At the same time, it's something
23:59that has been in a way highlighted.
24:03That's why that nowadays, it's not an easy task.
24:08There are variants such as accessing to justice,
24:10and difficulties in dealing with scientific evidence,
24:15and the conservatism of many courts
24:19that eventually confronted the different policy issues.
24:22So in a way, the idea is that
24:25there's a lot of now, research,
24:27going on, putting together climate science
24:31and low, try to gain or let's say,
24:35to create some synergies that eventually
24:37would help advance this field on climate litigation.
24:41And one important, let's say,
24:43call that I take from a recent publication
24:45of a colleague of mine, of Rupert Stuart-Smith,
24:50they say that greater appreciation
24:52and exploitation of current methodologies
24:54in attribution science could address obstacles to causation
24:58and improve the prospects of litigation.
25:01So in a way, it's really saying,
25:02"Okay, we know that we can do something.
25:05"Law can be a very good path for doing that,
25:08"but probably, science is we need to, in a way,
25:11"advance knowledge in this field
25:12"and try to provide better, let's say,
25:15"scientific evidence that could help, let's say,
25:17"winning on these different initiatives in courts."
25:21So let's say that so far,
25:23we have been working more on the part on climate events,
25:29more on the weather events,
25:31whether one weather event can be attributed,
25:33attributed or let's say,
25:34how was the role of climate change.
25:37But what about health impacts?
25:39Okay, we know that one event
25:42might have been more frequent or more, let's say,
25:45the probability has increased because of climate change,
25:47but at some point, we would like to know
25:49what this translates into health impacts.
25:52So, as I said, the idea is how much
25:54of the observed health burden during an extreme event
25:57can be attributed to human activities?
26:00Or more broadly, how much of the historical
26:04health burden of a climate-sensitive outcome
26:07can be attributed to climate change?
26:09So it's not an easy task.
26:11I mean, we know that in there,
26:14there might be some kind of different, let's say,
26:20developments in terms of methods, et cetera.
26:22And actually, one example, for example,
26:24you know that I found this article in The New York Times
26:28that was basically, they showed some calculations
26:31based on a recent report of the CDC,
26:35based on that, let's say, what has happened
26:37in these massive heat waves in the Northwest in the US.
26:42And actually, they just did a very simple estimation
26:46on the let's say, estimated the number of deaths
26:49that were kind of excess,
26:51or let's say, more than normal
26:53during that period of time, during the heat wave.
26:56Attributing that, let's say,
26:57that during this heat wave,
26:59more than 600 people died because that in a way,
27:03one could attribute to this heat wave.
27:08But the other question is how much actually
27:10of this burden can be attributed to human activities?
27:14Again, talking about the probabilities,
27:16not to say yes or no, is to how much of this burden
27:19can be kind of attributed to these events.
27:23So one of the first exercise that has been done
27:28in terms of attribution of health impacts
27:29was this study done by Dann Mitchell
27:33in which they assessed what was the impact
27:36of the 2003 heat wave in London and in Paris.
27:39And actually, what they found
27:40is that anthropogenic climate change
27:42increased the risk of heat-related mortality
27:44in Central Paris by 70%, and by 20% in London.
27:49So eventually, what is really in here
27:51is saying that how much human
27:55or anthropogenic climate change
27:57has either the severity of this event
28:00in terms of how much to really put the value,
28:03a number on this contribution in terms of health impacts.
28:06So in a way, you will see that
28:09it's clearly a different message
28:10compared to what I said before.
28:12It's not about the excess debt during that period,
28:15it's really to say how much,
28:16how many beds can be attributed
28:17to anthropogenic climate change.
28:21So let's say that traditionally,
28:22the way how we have assessed this
28:25is more into the future.
28:27Say compare in what has been there,
28:31the health burden attributed to heat in current times
28:34compared to what will be in the future
28:37using climate change scenarios,
28:38assuming that the difference between today
28:41and the future can be attributed
28:42to anthropogenic climate change.
28:44But you will see that from this idea
28:47of attribution studies is not about future,
28:49it's about present, okay?
28:51This is something that you should be reminded,
28:53is really to use historical events
28:55and try to see what has to be the footprint
28:58of human activities in historical events.
29:01So when we talk about the tradition, as I said,
29:04one could focus on, let's say,
29:07on a specific event to say,
29:09what they call event attribution
29:10for individual extreme weather events
29:12as this example of Dann Mitchell,
29:15but another example is more on the trend attribution.
29:19Basically, for long-term changes
29:21in the mean of climatological statistics.
29:23So not really to assess specific events,
29:25it's to see how much the observed trend
29:29can be attributed to human activities.
29:32So basically, using this approach,
29:35not really to focus on extreme events,
29:37but on the trend during a period of time
29:40is we did this analysis that it was, I mean,
29:46I had the pleasure to lead together with my colleagues
29:48of the Multi-Country Multi-City
29:49Collaborative Research Network
29:51was recently published is here.
29:53And this is the reason why I'm talking today
29:54about this topic, because thanks to this opportunity,
29:57I had really the option to dig a bit into this topic.
30:02So in a kind of general terms,
30:05this study like the general framework
30:10was about combining data and methods in epidemiology
30:15with modeling, let's say, climate projections,
30:20climate, let's say, simulations of the past years,
30:25we were able to estimate how much
30:28of the observed heat-related mortality
30:31can be attributed to human-induced climate change.
30:35So I will go step by step.
30:38First, as I said, we used data from the
30:40Multi-Country Multi-City Collaborative Research Network
30:42in 732 locations in 43 countries in the world.
30:48Here, you can see the different location
30:50of the different places.
30:51And the idea is that we combine,
30:54let's say, we took all this data
30:56on observed temperature and mortality,
30:59and we derived this, the vulnerability function
31:03or the exposure response functions of each city.
31:06You've seen the state of the art methods
31:09in climate change epidemiology
31:10is basically to a stage and serious analysis
31:13with distributed lag non-linear models
31:15and multivariate multilevel meta-regression.
31:17Yeah, it sounds super fancy,
31:19but in a ways, it's not as complicated,
31:22and you'll have all the information
31:24on the methods in the paper.
31:26I invite you to have a look,
31:27review if you would like to learn more
31:29about the methodological part.
31:31So basically, what we did, as I said,
31:33is to estimate the vulnerability of each city,
31:37which in a way, was already kind of an advancement
31:41compared to previous assessments.
31:43Because again, here, the idea is that we clearly
31:45or we aim to assess the specific vulnerability
31:50of each population to have a better estimation
31:53of heat-related mortality in each location.
31:56And you see here that it was clearly heterogeneous.
31:59We saw as we found in previous assessment,
32:02that actually, most of higher risks
32:05are usually found in Europe, in the Mediterranean area,
32:09and other locations in Asia.
32:12So as I said, we combined these exposure response curves
32:16with moderate climate data
32:18that we got from our colleagues from there,
32:21the DAMIP Project, this is
32:22the Detection Attribution Model Intercomparison Project
32:24that is based on the last CMIP6 simulations.
32:29And idea is that for each location in this assessment,
32:34we derive a series, let's say, of moderate pairs
32:39of moderate climate on daily temperature
32:43under current conditions
32:45and under our without climate change,
32:48that is our counterfactual scenario.
32:50Here, you have a kind of illustration of the trends.
32:54Basically, in red, you have the observed trend
33:00with a warming trend.
33:01That it mimics current conditions with climate change
33:04while the orange one mimics somewhere
33:07without climate change in the absence of warming.
33:09So basically, we focused in this period here
33:12between 19, yeah, 1990, oh, 1990, oops,
33:19I forgot, 1991 to 2006, 2018.
33:24Actually, sorry about the numbers,
33:25I'm very bad with that.
33:26And basically, what we did is as I said,
33:28for each location, we obtained these pairs,
33:33and we translated these observed,
33:36or let's say, simulated temperature
33:39into hypothetical excess mortality
33:43under these two scenarios.
33:45And this is basically what you can see here
33:49in this panel A.
33:52In solid, you have the anthropogenic,
33:58let's say, the heat-related mortality
34:01under current condition, let's say,
34:03in presence of anthropogenic climate change,
34:05while in light here, these bars,
34:08you have what would have been heat the excess,
34:11or let's say the heat-related mortality
34:14in a world without climate change.
34:16So basically, we estimated on this
34:18for each of the 700 something locations,
34:21and we aggregated by country,
34:23and this is what you see here.
34:24And eventually, we estimated the difference
34:27in terms of excess mortality between these two scenarios.
34:31That is basically what you find here.
34:33So what we saw overall
34:35is that 0.98% of excess mortality,
34:41heat-related excess mortality
34:43in the counterfactual scenario,
34:45and of course, more excess mortality
34:47in the factor is null,
34:49that is with anthropogenic climate change
34:51that is currently slipping to 1.56%.
34:56So the difference between the two
34:58that is basically, this number here is 0.58%.
35:03It represents the all-cause mortality
35:06that can be attributed to heat induced
35:08by anthropogenic climate change.
35:11So the idea is that in a final step,
35:13what we did is to kind of rescale this difference
35:17over the observed, or let's say, the impact
35:21or the excess mortality in anthropogenic climate change.
35:25In a way to estimate what is the proportion of this,
35:30the excess mortality happening today,
35:33that can be attributed to human-induced climate change.
35:37So it's basically, what we try to illustrate here,
35:41and we found that overall, 37% of heat-related deaths
35:46can be attributed to human-induced climate change
35:49in this assignment, these locations that we included.
35:52And in a later step, what we did
35:54is to kind of extrapolate this
35:57and compute what would be the mortality rate
36:00attributed to heat-related or let's say,
36:04heat-induced climate change.
36:06So in here, what we observed that on average,
36:092.2 deaths per 100,000 population per year
36:14can be attributed to heat induced in human influences,
36:18and let's say, of climate change.
36:21So as you could see in this assessment,
36:23it had very powerful message.
36:27It's really we provide evidence on the clear
36:31to tell the impacts of climate change
36:34over health burden that we observed today.
36:38And you can see that, of course,
36:39this evidence can be very, very useful
36:42for let's say, to support policy-making processes.
36:45And more importantly, I think,
36:46it was about the key message about the relevance
36:49of these findings in terms of to put
36:52a little bit more attention to what is happening,
36:55saying that climate change is not something
36:57that will happen in the future,
36:59it's already happening today.
37:02We can talk about the projections,
37:04but we cannot focus on your projections
37:06in terms of impacts of climate change
37:07is really that already we are suffering.
37:11So it's really to say, "Okay, we need to do,
37:14"or put more emphasis in terms of implementing
37:18"a strong mitigation policies to abate
37:20"this warming in the future,
37:22"but more importantly, to implement adaptation strategies
37:27"that would help us to reduce our vulnerability,
37:31"in this case, for heat."
37:33But of course, we had to acknowledge several limitations,
37:37and understood that, although it was one of the biggest,
37:39let's say, assessment on this field
37:41in terms of attribution of health impacts,
37:44we know that for example,
37:45it was cannot be considered a worldwide study
37:48because we focused our assessment on the locations
37:51that were already included in the MCC,
37:53and we know that there are important regions
37:55in the world that were not covered.
37:57This is an important limitation
37:59that we have in our study environment directly
38:01because we are very much aware
38:04that vulnerability is very heterogenous
38:07and changes from one location to the other.
38:09So at some point, we can extrapolate risk
38:12that we observed in Europe
38:14into places like Africa or Asia.
38:17So at some point, we need better data
38:19that would help us to better identify
38:22or let's say, assess what is the vulnerability
38:25of these locations that currently,
38:26are unobserved or unexplored.
38:28On the other side as well,
38:30something that we have to bear in mind,
38:31we have to do a simplification in terms of risk.
38:34We assume that, in a way,
38:38we did a cultural adaptation in the sense
38:40that we assumed a kind of average risk
38:44across the 20 years, 30 years that we assessed.
38:47And the idea is that okay, it's likely,
38:50and we know that as you could see here in this plot
38:53that actually, there might have been
38:57a partial adaptation of the population to heat.
39:01Though at some point, we don't know
39:04which impact this could have had
39:05because probably, the idea is that probably,
39:08at the end of the period, the risk
39:10might have been lower compared with the beginning.
39:12So eventually, as you could see,
39:14we had to do a kind of group simplification
39:17and something as well that we have to bear in mind
39:19is that the risks that we applied to both scenarios
39:22is the observed risk.
39:24That is the one that we estimated
39:26in our world with climate change.
39:28So we don't know what would have been the risk
39:31without climate change.
39:33So again, it's very difficult,
39:35and I expected in the future,
39:37it's something that we will implement in them,
39:40in this field or in climate change epidemiology.
39:43And finally, the lack of epidemiological causal basis.
39:46This is important because this assessment
39:48is purely based on an ecological design
39:51that as most of the climate change
39:53and epidemiological studies.
39:56So at some point, if we want to talk about the attribution,
39:58we have to improve our way,
40:01how we can assess causal links in this field.
40:06So just as a kind of final wrap-up on this subject,
40:10and as I said, I really want you
40:11to make it fun about this kind of a study,
40:14is first, because as we know,
40:18it can be a powerful tool for climate change policy,
40:22and as well, it can help understanding the mechanism
40:25by which climate change effects have.
40:28Can support in finding projections
40:29of future health effects of climate change,
40:32and as well, improve adaptation to climate change impacts
40:35and disaster recovery.
40:36As well, it can increase motivation
40:39for climate mitigation, as I said,
40:41just learning about what is happening today
40:42and the urgency to really do that.
40:46And also, demonstrate causal links
40:48between greenhouse gas emissions
40:49and climate change impacts
40:51that serve as a basis of evidence
40:52underpinning climate-related losses,
40:54as I said, previously.
40:56And finally, what I believe is also very, very attractive.
40:59It's an excellent platform for interdisciplinary research,
41:03really to put together experts from different fields,
41:05from climate science, working more on the modeling side,
41:09climate epidemiologists, working
41:11on the ascertain the health impacts.
41:13And at the later stage, other experts in other fields
41:16like the economy, law, et cetera,
41:18can take part on these investigations.
41:20So definitely, it's an excellent platform
41:22for latching our kind of research area,
41:26grab information, address knowledge from other fields
41:29and reach our risk portfolio,
41:31which I think is also very relevant for young researchers.
41:35And just as our final point,
41:39something that I think it has to do, bear in mind,
41:41and as I said for me, this research field
41:45can be considered kind of very powerful research line
41:48in the future in climate change epidemiology.
41:52Let's say, climate attribution
41:54is something that has been developed
41:56for years in climate science sphere,
41:58but not much in epidemiology.
42:00And if we really want to advance in climate litigation,
42:06really advance on the fight against climate change,
42:09we have to put a value on what is happening
42:12in terms of extreme events,
42:13in terms of X is that burden, economic cost, et cetera.
42:17And all this can help people change your mind,
42:21and as well, help, advancing or let's say,
42:24winning different initiatives in courts, et cetera.
42:27So as important elements that I believe
42:30we should focus in the future,
42:31is first assess causality, use advanced methods
42:35in environmental epidemiology
42:37that help us to clarify causal links.
42:39Second point, to provide innovative frameworks
42:43that probably, I mean, you think about
42:46as the world attribution initiative,
42:48they provided this evidence on the role of climate change.
42:52If we can couple this with health impacts,
42:55that could be even much more powerful.
42:58And finally, we have to think
42:59about how we can address this research question
43:02in a more broader perspective
43:04and provide probably, global estimates
43:06that are closer to the, let's say,
43:08the real, what is happening today.
43:11So yeah, that's all.
43:13Thank you very much for your attention,
43:15and I'm happy to take questions, thank you.
43:19<v ->Thank you, Ana.</v>
43:19Thank you for the wonderful presentation.
43:21I think you gave a superb view
43:24like an introduction from kind of science, how to tackle
43:27and attribute extreme weather events,
43:29and these type of extreme events attribution
43:32to the trend attribution,
43:34and to the landmark study that you have,
43:36the MCC quality you've been working on.
43:38So thank you very much.
43:40And I think there will be a lot
43:43of questions from our audience.
43:45So while our online audience is typing
43:48your questions in the Chat box,
43:51we do have already collect some questions from our students.
43:55So there are several types of questions
43:59that students are particularly interested in.
44:01For example, the first type,
44:03I think for some of the students still wondering,
44:06you have given this great example
44:09of attributing heat-related mortality.
44:12So they're wondering if this type of technique
44:14can be used to attribute other extreme weather events,
44:17like hurricanes or wildfires?
44:21<v ->Yeah, exactly, I mean, as I said,</v>
44:23in this assessment, in this presentation,
44:25I focused on heat on health,
44:27because in a way, I mean, of course,
44:28it's a bit biased because it has been
44:31my research field for already several years,
44:34but we know that within attribution science,
44:38it's not only about heat waves.
44:40Actually, there's also a very new report
44:44published by this Global Weather Attribution initiative
44:49on the floods happening in Central Europe even this summer.
44:54Again, put in, estimated that actually
44:56the role of climate change was very substantial
44:59in let's say, in increasing the probability of this event.
45:01So definitely, this kind of framework
45:05can be extended to other extreme weather events.
45:08Of course, with some caveats and some limitations,
45:12but I believe that if we try to, let's say,
45:17it would be easy to adapt this framework
45:21to other extreme weather events
45:22if data, of course, is available.
45:27<v ->Thank you, Ana.</v>
45:27I think we have a typo from our online audience.
45:31Exactly, the same question
45:34some of the students are also asking.
45:36But Mona is asking,
45:39"Why is the A and B data missing environments
45:42"from most of Africa?"
45:44And also, it's kind of related to the question
45:47student's asking in the Multi-Country Multi-City
45:50is that they only have South Africa,
45:54doesn't have much of Africa.
45:56And also, one of my students is asking,
45:59why there's no data from the South Pacific,
46:03where she have experienced doing this one
46:07and like fuzzy.
46:08So why there's no such coverage?
46:12<v ->Well, maybe I can give you a little bit of story</v>
46:15about how the MCC started.
46:17And basically, it was, I think in 2014
46:21during a conference, with a group of researchers
46:24working on climate change epidemiology,
46:28mostly on the temperature-related health impacts.
46:31They had an informal meeting,
46:34and they were discussing the possibility
46:36of probably putting together some data from their countries.
46:39For example, one have data on temperature mortality
46:42in the UK, other have in Japan,
46:45the other one had in Spain.
46:47So eventually, they realized that,
46:51"Okay, maybe we can start putting all this data together
46:54"instead of assessing our impacts
46:56"or let's say, our estimates in our country,
46:59"it would be nice to compare different locations
47:01"at the same time."
47:02Because as I said, the idea of,
47:06the peculiarity in a way of temperature-related
47:10health impacts is that this,
47:13the effect is very dependent on the location.
47:16So it's nice to compare these estimates across locations
47:19to understand vulnerabilities
47:21and potential vulnerability factors.
47:22So as I said, it started kind of informal way,
47:26and they started opening the door to other collaborators
47:32and colleagues to work in,
47:33and eventually, it grew, grew, grew,
47:35grew until nowadays that we are around, I think,
47:3870 researchers from 43 countries
47:43with all these bunch of locations
47:45with different data sets.
47:46And also, not only focusing
47:48on the idea of temperature mortality,
47:51but also, air pollution, on projections,
47:55on I mean, in a way, it's a project
47:58that greatly grow in an exponential way.
48:04But the idea how this, then the beauty of this project,
48:08how it's developed and how it started
48:11is that it works in a very informal way
48:13in the sense that the idea how you contribute,
48:16you take part of this consortium by providing data
48:20on a specific country that is missing
48:22because you had it because of your research or whatever.
48:27And it's surprising that it is not directly funded.
48:29I mean, it works, let's say,
48:31off each funds of each partner.
48:34The reason why there are some places in the world
48:37that is not, let's say, covered within this spread
48:40is basically, because so far,
48:43we didn't manage to get data from these locations.
48:46And I mean, it's a problem of course,
48:49of places like in Africa,
48:51where good quality on mortality,
48:54daily mortality in specific locations in Africa
48:58is very difficult to find.
48:59Especially because at some point,
49:00whether you need this data is somehow
49:03comparable in terms of quality and temporal scale.
49:07And especially, this idea
49:08that it should be daily mortality, et cetera,
49:12because the part on them, whether we know
49:14that is relatively easy to get it
49:16from the analysis data, et cetera,
49:20but the main limiting factor here is the mortality data.
49:23And that's why in a way,
49:25we didn't manage too far to kind of get this information
49:29here in terms of observed mortality in this assessment.
49:34However, very recently,
49:37as I mentioned in my first slide,
49:39we performed a global assessment
49:42in which basically, based on information
49:44of the observed locations, our colleagues in Monash,
49:48they managed to extrapolate
49:50the risk in an observed location
49:52and eventually, provide kind of comprehensive
49:55assessment on the team,
49:58non-optimal temperature-related mortality across the globe.
50:02I invite you to have a look in,
50:06I think, it was recently published in (indistinct).
50:10<v ->Thanks, Ana, I think,</v>
50:12if you collect it with the history
50:13and also development for MCC,
50:15why it's not covered?
50:16And what's the most recent
50:18that MCC predict in the temperature mortality
50:21association in places where you don't have mortality data.
50:25There are always a lot of questions,
50:27but I do have one kind of question
50:30related to your answer.
50:32This one student is kind of were astonished about
50:36since the heat-related mortality risk
50:38varies across places that you have shown me on slides.
50:42So the question is why do places have,
50:48even we have similar latitude,
50:50maybe even with the same organization level,
50:53why do we have different heat-related mortality risk?
50:58<v ->Well, it's a very good question,</v>
51:00and I must say, difficult to answer
51:03in a very clear way.
51:11In a way, we know that vulnerability to heat
51:14or let's say, non-optimal temperature
51:16depends on a complex network of different factors
51:22that are highly interconnected.
51:24It's not like we know so far
51:26that what makes one city more vulnerable to the other
51:31is not because of one unique factor.
51:33It's because of combination of different factors
51:37that actually are very much dependent between each other.
51:40Thinking that, for example, we published,
51:42I think, it was in 2018, a study was led
51:46by our colleague, Francesco Sera,
51:49in which we tried to assess specifically this,
51:54to try to understand what were the contextual factors
51:58defined at city level that can give us some hints
52:01about which locations are more vulnerable
52:04in terms of higher excess mortality
52:06due to heat compared to others.
52:08And eventually, what we saw in this assessment
52:10is that it's not only one factor,
52:13it was a combination of probably cities
52:15that are more urbanized,
52:17but also more unequal are those
52:20with a higher heat-related burden
52:24compared to others with a lower level in this case.
52:27Well, for cold, the story was much more complicated
52:30with no clear patterns around.
52:32But again, the idea how all, let's say,
52:36the main factors driving this difference
52:38is nowadays, have very important
52:40or very crucial point that we are trying to disentangle,
52:44especially because we know that if we understand
52:47what are the mechanism, let's say,
52:48the reasons why one city is more resilient
52:52compared to other, this can help us
52:56to understand which mechanism in terms of adaptation
52:59we can apply to other places to try to protect
53:01to reduce our vulnerabilities in the future.
53:04So hopefully, if you ask me this question in a few years,
53:08I hope I will answer this question,
53:12but I think right now, it's very difficult to say.
53:15<v ->Yeah, yeah, I think it's excellent answer now.</v>
53:18So it's kind of related to one,
53:20our online audience questions
53:22regarding the difference in the heat-related mortality,
53:25whether it is rural or regional kind of communities.
53:28I think it's more related
53:30to Francesco Sera's paper you mentioned.
53:34<v ->Yeah, in a way, I mean,</v>
53:35it's still that in this assessment,
53:37and I must say that in the MCC,
53:39most of the locations that we have are cities.
53:43So in a way, the risks that we obtained
53:46are mostly represented for urban locations.
53:49This is one of our limitations in this assessment.
53:52And probably, if you don't,
53:53you need to go a kind of national level assessment
53:58in which you can better disentangle the different,
54:02let's say, patterns in terms of vulnerability to heat
54:06and cold in a rural versus urban.
54:09And as I said, it's also it's a story that needs to,
54:13we need to address in the next years.
54:16And I know there are initiatives in terms
54:17of nationwide assessments try to see patterns
54:22between urban and rural locations, et cetera.
54:27<v ->Yeah, I think kind of the final</v>
54:29group of questions students and also online audience
54:33is interested is adaptation.
54:36So I mean, the adaptation matters
54:41students are kind of wondering,
54:44how must immediate needs to deal
54:47with increasing temperature can be balanced
54:50against the long-term goals of emission reduction?
54:54Basically, asking using adaptation methods
54:59to talk to the long-term global warming paths.
55:04And also, if there are some studies like this,
55:07are there any practical suggestions
55:11on how local communities can do
55:14about the adaptation methods?
55:17<v ->Yeah, and I must say it was one of the key messages</v>
55:20of this assessment that yeah, I presented today,
55:23in this attribution study,
55:24because of course, we give a little bit,
55:28it gives them the message about the urgency
55:30in terms of abating or let's say,
55:35reducing the warming in the future.
55:38But more importantly, what it is saying
55:40is that we really need to reduce our vulnerability
55:44because the idea is that with mitigation,
55:46we know that these benefits will come
55:48in the next decades while with adaptation,
55:51these benefits can come earlier.
55:53And probably, this can be even more efficient
55:55compared to just waiting for, let's say,
55:57the mitigation strategies to have some impacts.
56:01And it's true that we have to think
56:02about even in the best of the scenarios today,
56:06in which we set emissions to zero,
56:08we will be any way exposed to warmer climate
56:11in the next decades.
56:13So it's about, really again,
56:14to put emphasis into the idea of adaptation
56:17that it might be the key on this story.
56:20And with regards on how we can counteract
56:24future warming in terms of how much
56:28we can decrease our vulnerability
56:30to counteract this warming.
56:31I know that there have been some initiatives
56:34of some studies published in the past.
56:37For example, there's a study by our colleagues in Romania
56:41that they simulate what this kind of how much
56:45we would need to reduce our vulnerability
56:47in the future to reduce or let's say,
56:51to keep our heat-related deaths in the future constant
56:56despite the global warming.
56:58So in a way, this is a very nice exercise.
57:00That is certainly something that as well, I'm leading,
57:02an initiative within the MCC to try to address this.
57:05Because as well, this can help us about them,
57:08how much we need to adapt to really do something,
57:13to have some impacts in terms of reduction
57:15of heat-related mortality.
57:17Because imagine that if warming continues,
57:21and let's say, the pace at which we adapt
57:25is not quick enough, let's say,
57:29to kind of counteract this warming,
57:31we eventually will have the same heat-related deaths
57:34today but in the future,
57:35which of course, it would be fine.
57:36But ideally, what we would like is that
57:39the heat-related deaths happening today
57:41won't happen in the future anyway.
57:45<v ->Thank you, Ana.</v>
57:46I think, I saw Tobias posts a comment,
57:49"A really fantastic talk."
57:51So I think is there are any final questions?
57:58If there's no final question,
58:01thank you, Ana, very much for this
58:03really, really amazing talk.
58:04And I think both the students
58:06and I'm sure, our online audience
58:08learned a lot from you, but thank you so much.
58:11<v ->Thank you, thanks a lot</v>
58:12for the invitation, my pleasure.