 Good afternoon and welcome to EGU22, the annual meeting of the European Geosciences Union. As many of you already know, this is the union's first hybrid general assembly, where we are bringing back our on-site experience for those joining us in person, while at the same time introducing new concepts from the last couple of years to include our virtual attendees as much as possible. This year we've had more than 12,000 abstracts submitted to EGU's meeting. And during the press conferences, we'd like to highlight some of the most unique studies, which as you'll soon see, have impacts on local communities, industries, ecosystems and the global environment. I'm Julian D'Souza, EGU's media and communications officer, and I'll be hosting this week's press conferences. Each press conference will have time for speakers to make their presentations, followed by a question and answer period at the end. For those of you joining us virtually, I ask that you mute your mics throughout the briefing until I call upon you to speak. If for some reason you experience technical difficulties, you can try to rejoin the session or look for more information on the press conferences section of the media.egu.eu page. A last couple of things to note, please save all of your questions for after the speakers have finished presenting. During the Q&A period, we will take questions from journalists both in the room and online. If you're in the room, please raise your hand so we can pass a microphone over to you. If you're joining us virtually, please use the hand raising function on Zoom so we can come to you for your question. If you prefer to type your question in said, that's perfectly fine, feel free to do so in the chat. All right, so now I'm going to go ahead and introduce all of our panelists to make for faster transitions between them. This afternoon's press conference is titled, From Food Security to Flooding, Technological Solutions to Geoscience Challenges. Our panelists for the day are Elizabeth Safran from the Environmental Studies Programme, Lewis and Clark College, Portland, Oregon, and Elizabeth is joining us virtually. We have Hussain Najafi from the Helmholtz Center for Environmental Research, Computational Hydro Systems, Leipzig, Germany. We have Jessica Murzdorf from the Eustus Liebig University, Institute for Landscape Ecology and Resources Management, Giesen, and Alliance to Feed the Earth and Disasters, Fairbanks, U.S. And we have Keshwan Ginkal from Delte's Deaf, the Netherlands. All right, so we will begin with Elizabeth's presentation. Elizabeth, over to you. All right, great. This collaboration between a geologist, a computer scientist, a psychologist, and a scholar of media studies is focused on what it takes to motivate young adults to prepare for earthquakes. We find that video games can be effective tools to promote learning, confidence building, and action. Sorry. Here we go. We're motivated by the fact that the Pacific Northwestern United States is subject to huge earthquakes and tsunamis, but lacks a culture of preparedness, partly because the last big one, big earthquake, was 322 years ago. And there are few moderate sized earthquakes in between. I've spent the past two months in Japan and felt at least eight earthquakes. In the 20 plus years I've lived in Portland, Oregon, I felt maybe two. Because the regional infrastructure is not resilient to these earthquakes, individual and community preparedness is especially important. We are interested in young adults who are a relatively under-messaged population with significant capabilities and vulnerabilities in times of disaster, like physical strength or language skills on the one hand, and financial insecurity and lack of control over their living environment on the other. We want to know what motivates young adults to prepare for earthquakes. We're using custom-made video games in experiments to address parts of that question. In places where severe hazards occur infrequently, media are essential sources of vicarious experience. But video games compel more active engagement than traditional media and can be used to simulate rare events. They also resonate with the media preferences of many young adults. There are quite a few disaster-related games, and some of them are video games, including some many which have been described in recent reviews. There is even a review of reviews in the Games for Geosciences session tomorrow. Most of those reviews point to opacity of research on the game's effectiveness, and this is a gap that our work seeks to address. Our experiments are guided by social cognitive theory, which emphasizes links among personal cognitive factors, socio-environmental influences, and health-promoting behavior. Our main dependent variables are learning, self-efficacy or confidence in one's ability to carry out a behavior, outcome expectation, which relates to the anticipated consequences of an action, intent to act, and action. We consult on a regular basis with emergency managers to incorporate their priorities into the games and to share our results and games as they emerge. We're just getting ready to make the first game available, so let's get a little taste of it. The first and foremost research tools. The experiment I'm describing today compares the impacts of either playing this video game or searching the web on all these dependent variables listed. Participants in both conditions were primed with a statement about the expected consequences of a magnitude 9 Cascadia subduction zone earthquake. Half the participants then played our video game and half searched the web for earthquake-related solutions. We asked them to perform their task for at least five minutes, but they could do it for up to 45 minutes at will. We gave them surveys beforehand, immediately afterwards, and three months later to assess our dependent variables. The game involved four types of challenges, each of which was solved in three ways over the course of four levels. For example, shown here, capturing and boiling rainwater to get drinkable water. Web searchers were offered three starter links to emergency management websites, but were also told they could search freely. These websites contained or linked to all the content included in the game, and of course, much more. So we found that gameplay had several short-term advantages over web searching. Gameplayers performed their task for significantly longer and enjoyed it more, though they also found it more challenging and more frustrating. They downloaded more handouts and made more inquiries about disaster training immediately after their task. When we asked them to come up with solutions to a problem closely related to one featured in both forms of media, Gameplayers listed more solutions and more kinds of solutions. They perceived learning significantly more new information, though both groups performed similarly in a free recall of lessons learned. We expected the Gameplayers to trust the game less as a source of information, but the two groups actually trusted their respective sources of info equally. Self-efficacy and intent to act are both considered important precursors to taking action. Broadly speaking, both self-efficacy and intent to act increased for eight general categories of action as a result of the experiment. Gameplayers showed significantly greater increases in self-efficacy than web searchers for some of the categories that were really foregrounded in the game, like getting drinkable water or managing bodily waste. After three months, differences in self-efficacy and learning between the web searchers and the Gameplayers kind of evened out. For both participant types, self-efficacy for five out of the eight categories of action remained elevated. For the Gameplayers, that included all the categories of action that were actually in the game. We asked about some that were not. Intent to act was a very different story. It returned essentially to pretest levels for both types of participants. According to their self-reports, both Gameplayers and web searchers took significant steps to prepare for earthquakes, with web searchers perhaps showing a slight edge, although the data are a bit ambiguous. In summary, and to return to the question in this talk's subtitle, games can help prepare young adults for earthquakes and opportunities for immediate action should be integrated into game environments to capitalize on those short-term advantages that Gameplay confers. And I'll close with thanks to our undergraduate research assistants, funding sources, emergency manager collaborators, and I'll note that the findings, opinions, conclusions, and recommendations expressed here do not necessarily reflect the views of the National Science Foundation or any of our other funders. Thank you very much. Thank you so much, Elizabeth, for your presentation. We will now move on to Hussain's presentation. And Hussain is also joined by his colleague today, Heiko Apel, who will also share in with their findings. Over to you, Hussain. Okay, so I'm going to talk about some new insights and lessons learned from the forecasting in Germany, flood forecasting when based on the previous flood that hit R last year in 2021. So basically, if we look at the river and flash floods over the years in Germany, you can see that there are a number of fatalities since 1870s. And the one that was last year is one of the biggest and catastrophic one with around like 180% died, unfortunately. So we look at the statistics of this event. This is done with our colleagues at GFZ. So if we look at the statistics of this event, you may see that the return period of this flood with around 975 cubic meter per second or other estimates also are about like 1100 cubic meter per second. As the return period of this event is almost around 10,000 years. And this is significantly higher than the previous flood, which was seen in this river basin, which was in 2016 with around like 236 cubic meter per second. And then the flood hazard maps basically has been based on the value of 286 cubic meter per second. So when we have this events, we usually use the flood forecasting systems with different components, as you may see here. So the most basic part of this flood forecasting is the observations of meteorological variables. We need this so that we can generate the initial conditions generate the initial conditions for our hydrological models. And once we have the state of the current estate, we can force the hydrological model with the weather forecast. And then this will provide us with some hydrological forecasting, which can have several times of hours to days, or even months. And ideally, the output of this model, which could be the runoff stream flow or water level can be used for some regional andundation modeling, which has been also done by our colleagues at GFZ. And I'm going to present about that in the future slides. So basically, for this particular event, there was some warnings available from the ECMWF Center. This is based on the extreme forecast index. And here you can see the ranges of this index. The value of the index, which once it's more than .8, it shows severe weather. And it could be an exceptional event. And from the figures from left to right, you can see that this index has been provided on July 14. And it has already shown a significant value, which is like around .9 for this particular event. Another forecast was available by the European Flood Awareness System. And this has been updated every 12 hours. And here you may also see that the forecast has shown like the return periods of more than 20 years for this particular event. So back to the components that we show in previous slides. We have two different products in Germany, RADALAN and RIGNI. RADALAN has hourly data time steps and RIGNI is daily. So we use this information for forcing or hydrological model. As you may see here, the RADALAN compared to RIGNI has underestimated the precipitation, but like 20 millimeters less than RIGNI. So we use the spatial, the temporal patterns of RADALAN to disaggregate the RIGNI in hourly. So then we can have quantified the effect of this precipitation and uncertainty associated with this for our hydrological model. The hydrological model we use as MHM, the Meso-scale hydrological model. It has been developed since 11 years at UFZ, Germany. And we have some enhancement for this so that it can read the hourly precipitation data. And that's the same model that is used for the German drought monitor and it's a grid-based model. So in this figure, you see the constructed flood peak with the blue hydrograph and also the red one, which is based on the RIGNI disaggregated and RADALAN with the purple one. And as you may see, the uncertainties associated with this near real-time products is not so different, but it can also help us with simulation of this flood event. One other part of this model, the forecasting chain is like this flood impact forecasting and it's the inundation modeling, which has been done by our colleagues at GFZ. So basically, the figure here shows the water level available for, based on the flood center with the red one, which was also underestimated and the blue one, which was actually the observed value of the flood. And here you will see that this model, which is not computationally expensive, this model has been run for the event of 15-hour time period, just in seven minutes. And this is like the application of how we can use this kind of models in the forecasting chain to help early warning of the forecasts. We also look at the climate change projections for this particular event. And as you may see, based on several climate change scenarios, this event is exceptional and like with around 550 cubic meters per second, you can see that even in the climate change projections, this kind of event can be considered as exceptional. So the lessons we learned is like that the event was predictable, as we have shown in several figures. Or the state of the art, whether in hydrological forecasting tools, can have forecast this particular event. By implement the hydrological, the hydraulic forecast models with low computational resources, we can add these models to the forecasting change for early flood warning. And one last point that I would like to mention is that the data should be available from adjacent countries in case of we are interested in flood forecasting of the transpoundary rivers. This data is available only for the state flood centers and regional flood centers. But if this kind of information is available for the research part, then we can improve our models and have a benchmark so that we can improve our forecasting systems. That was my last slide. Thanks. Thank you, Hussain. Over to you, Jessica, if you're ready. Yeah, hello and welcome. First of all, the sharing of the slides is fine from our side. I want to talk to you about my thesis work, which was conducted as a joint project between the nonprofit organization Alliance to Feed the Earth and Disasters, or short all-fed, and the this is Liebig University Gießen in Germany. It addresses the question, can we feed everyone without our modern infrastructure and industry? If this topic sparks your interest and you attend virtually, you can use the link provided at the bottom of this slide to connect with me or with all-fed throughout the week. Let's have a look at the scenarios which can cause a loss of industry scenario. So first we have high-altitude electromagnetic pulses, which are the result of the detonation of nuclear warheads in the higher atmosphere, radiation from severe solar storms, or globally coordinated cyber attacks. All three of those have the potential to severely damage or inhibit the use of industrial infrastructure. The fourth scenario on the slide is one that we are all way too familiar with at this point. It is a severe pandemic. As we have already seen at the height of the pandemic, people were scared to go to work. If the pandemic were even more deadly, then this could lead to the abandonment of critical infrastructure and it affect the loss of industrial products and services. These losses would have a large effect on our modern agriculture as it is heavily reliant on industrially produced fertilizer, pesticides, and agricultural machinery to do the heavy lifting. To address these impacts, I built a regression model based on current spatially explicit gridded crop data to model the relationship between crop yield and selected influencing factors. Afterwards, this model was used to predict yield in a loss of industry scenario. We selected four staple crops for analysis, the first three are cereals, maize, rice, and wheat, and then also the joom and oil crop, which is soybean. The yield of these crops depends on different growing factors, and according to their relevance to our analysis and the data availability, we chose four factors. The influencing factors we chose are nutrient input in form of manure or industrial produced fertilizer in kilogram per hectare. The pesticide application also in kilogram per hectare if the agricultural production was conducted using agricultural machinery or not, and also we looked at the fraction of irrigated crop land in a specific area. For loss of industry, we're looking at a worst case scenario, but we also split it into two phases. For both phases, we do not have access to any electrically or fossil fuel reliant irrigation. However, on phase one, we assume that there will be some stocks left of fertilizer, pesticides, and fuel due to production surpluses before the catastrophe hit and distributed storage. In phase two, on the other hand, we will not have any of these stocks anymore and then no access to industrially produced inputs, so we can only rely on manure, biological pest control, and draft animal power to aid us in the production of agricultural goods. Now let's have a look at how we would fare in a catastrophe according to my model. Here you can see the overview of the mean crop yield for each crop, so we have the four crops, maize, rice, soybean, and wheat, and in blue you can see the current crop yield from the early 2000s in purple, the projected yield for phase one, and in green the projected yield for phase two. As we can see, we have already a strong reduction in phase one, ranging from 17 to 35 percent yield loss. In phase two, this gets even worse and we look at a reduction of 37 to 46 percent in yield. Overall, that means that for phase two, we're looking at at least one third of yield loss for each crop. Still, this can be misleading, so let's have a look at the spatial distribution of these projected yield losses. Here you can see again the different crops, maize, rice, soybean, and wheat. The gray areas are not particularly affected by loss of industry scenario. For the other ones, the deeper the color red, the more severe the yield loss. We can clearly identify some hotspots on these sites. However, they do not vary too much between the phases, so let's have a look at phase two to more clearly see the different hotspots. Here you can see the losses are much larger and the hotspots we're looking at for overall are North and South America, Europe, China, India, and Indonesia. This coincides with the main growing regions and also largely with the industrialization of the specific agricultural production in these regions. We can also see that the losses in these regions are a lot larger than what we saw before with the mean, so here we're actually looking more at 50 percent and more yield loss in the hotspot regions. What exactly does that mean now for the question we posed ourselves in the beginning? Can we feed everyone? The short answer, according to my model, is we could, but it takes a lot of preparation. Currently we are producing more than double the amount of food that is needed to feed the world population, so the mean that we estimated at 41.5 percent yield loss for the specific crops or for overall crop yield would still be leave enough food for the global population. However, as we've seen in hotspots this can be a lot more and so we would really need to invest in preparing logistics and communication systems for this scenario and also this doesn't address the political or economical dimension of the problem which is a whole separate question that needs to be addressed separately. The work presented here provides the first spatially explicit and crop specific estimate on crop yields and two phases of loss of industry scenarios. However, this is not a final answer and we need further research as I've said into the logistical, political and economical dimensions but also to improve the model that we have right now. One step which is really important is to include more factors that we could not address here due to the lack of sufficient data and also we have a lack of analysis or a lack of data for the whole African continent so we need to do more regional analysis to provide in-depth information. As said before, if you want to connect with me if you're attending virtually then please use this link and thank you so much for your attention. Thank you Jessica for her presentation. We will move to now our last speaker, Kiesh, if you're ready over to you. Thank you and so my name is Kiesh Vaginkel. I work at Deltares Research Institute and pursue a PhD at VU University in Amsterdam and today I will be talking with you without a mask I forgot. I will be talking with you about tipping points. So what is a tipping point? I will do a very small demo so let's say this is our climate driver and the bottle is a system you see that is initially in some sort of stable state but if I push it a little bit then it will return to the original state right push it a bit if I push it more there's some critical point you see it going the tipping point but push harder the system collapses so it moves to a new state which is fundamentally different well I put the cap on it now but if I haven't done it would have large impact so we know that these kind of tipping points they happen in the climate system but also in ecosystems but the objective for my research was to investigate if these also happen in social economic systems to make that a bit more concrete I have three examples so the first is bankruptcy of ski resorts so here our driver would be the snow line that is retreating and the impact would be that the ski resort could go bankrupt the second example is a house price collapse so again the driver here is the climate risk that could be increasing and what suddenly changes is the house price so it could collapse the last example I actually took from the air valley that's the picture you see there and you can imagine if we get really big floods so that's the driver again at some point so much of the road network is destroyed that like the whole country is very poorly accessible and it's also some sort of a tipping point today I will focus on house price collapses in a coastal city based on a study we recently did so this is our city and you can see it's threatened by sea level rise and every year there is a storm and we don't know how high it will be we know a little bit about the range but it's still uncertain what will happen in the future and there's people living in the city and they sell property and buy houses and as a result there is some house price now we all know about house prices that some people take very rational decisions so if you are rational and you you buy a house you would pay less if the if you would expect flood damage right because you need it to repair it however you also know that many people are not aware of the flood risk so just after flood they might panic and they think it's they pay less for their house and after some time they even forgot that they live in a dangerous city and you don't see an effect on the house price we will be looking very much in the future so till 2200 so many things are uncertain not only sea level rise in housing markets but also how we will adapt so how will we heighten the dykes so the model assumes that we will not do nothing but that we will continuously adapt to the changing sea level to account for uncertainty we did 400 or nearly 400 thousand experiments which all represent a possible future it's a bit complicated figure but good thing is if you understand this you have the whole point so let's go so if we have a small sea level rise especially here oh actually you see the results in the grid so if it's like reddish it's a very high tipping point likelihood you can see how it could look like a couple of tipping points over time where the house price collapses and if it's wide there is no tipping point at all so in low-end sea level rise scenarios we hardly see tipping points but if we go up to the most extreme sea level rise scenarios that for example follow from Antarctic ice sheet collapses it's a very serious risk now here horizontally you see the role of adaptation so if you are very reactive so the government waits for something to go wrong or almost go wrong before they do adaptation you see that we end up with a lot of tipping points but if you are very proactive we should be good at these most of the in the most of the scenarios now building wanting to build a dyke is one thing with having it and having it protecting your city is another right so it takes a time to implement the dyke heightening and that is what you see here fast or slow so if the procedure of building and planning for the dyke is really fast you see it's a really big advantage and if you're slow even if you're proactive so we have good ideas but we're just slowing implement implementing it it still fails well also watch these persons i will come back to them in the next slide so what does this all mean well first of all that accelerating sea level rise is really a game changer so we if we have a collapsing west Antarctic ice sheet and there's a lot of presentations here about this topic there could be a serious threat that many coastal cities could have these kind of tipping points in the coming say 100 years or even further in the future so 200 years even the good news is that we can avoid most of them that you can see here so if you're very proactive and we somehow managed to do a very smooth adaptation processes we can avoid most of these tipping points however we know that we cannot take this for granted if we look for example at the Venice most barrier or in Rotterdam the big storm surge barrier it can easily cost like 50 years from the first idea to when it all works fine and protects us now finally the persons you see are they represent the dynamics on the market so these are all likely range IPCC scenarios so more likely than these but you see that here we only see tipping points at in the housing market where people are driven by sentiment so they panic if a flood happens or a near-miss event so they really overestimate the risk at that point and we know from economic research that this really happens whereas if they would be all rational even if there was a flood once and they yeah it would not change the situation that much right so if they are rational you can avoid tipping points whereas people panic for nothing maybe you can have a tipping point so the flood flood risk perception really matters to conclude we see that it is possible that sea level rise causes social or economic tipping points and that especially the accelerating sea level rise is a game changer and if you have very proactive flood risk management you can avoid these tipping points but you must plan very much in advance you must monitor what is coming to you what is the sea level rise that we can expect and what was surprising from this study is also it's not only about building the dive that is one thing that convincing people that they can trust it is another so it's also very important to inform people about the actual risk so they don't underestimate it when nothing happens and they don't overestimate it when something goes nearly wrong if you want to know more about this example or the ski resort or the road network that's jets off to the meeting thank you okay so um thank you very much to all of our speakers um I think we can all agree that there was somebody forward-looking futuristic solutions presented to some very ideal world problems um we now move on to the last part of our press conference today which is the question and answer round so I will now open the floor to questions both from journalists in the room and those online um just to recap if you have a question please raise your hand and I will hand over the mic to you if you're joining us virtually and if you have a question you can type it in the chat or use the hand raising function on zoom and we will call upon you for your question we have a question I'm just going to hand the mic to you it's a question for Jessica so in planning for worst-case scenarios of other kinds I'm thinking like uh by a weapons attack and governments have or shocks the oil production governments have taken a strategic stockpile approach right and just warehouse enormous quantities of the resource that would be needed to weather the initial shock for a while until things could be rebuilt could that approach work for the kinds of shocks you're looking at and what sort of resources would need to be stockpiled for that to work thank you for the question it wouldn't work in all kinds of scenarios there is obviously the uh possibility that some of these shocks could be resolved pretty quickly for example a cyber attack could only be an hour long and then the specific people would have it under control again and everything would be fine however if there is lasting damage to critical infrastructure this can take years to rebuild even on the normal conditions some of these critical infrastructure pieces take even one or two years to be produced and delivered for use so we're looking more at solutions that could work for a couple of years and even longer than that and for these kind of solutions stockpiling does not work because it's just an enormous amount that is not viable and also very very expensive at one point of all that is also to look at the cost effectiveness i have not done this for my research but it has done been done for other kinds of scenarios and uh stock piling is very expensive and not very cost effective so what we are looking at is more what can we do to prepare our society to react to these scenarios or like prepare for them in a way that is cost effective and efficient and in these cases we're looking more at protocols alliances in a political sense or in my scenario at really diffusing the information on how to be able to do agriculture without a modern infrastructure thank you jessica do we have any other questions do we have any questions that are coming in virtually all right just wait another minute you have a question i have a question for elizabeth so very interesting talk about the positive effect of video game for the preparedness of the earthquake so my question is do you have any plan to implement the same game similar game for other for example flat event thank you thank you for the question we don't at present have plans to implement for flood scenarios we have a research agenda for several other experiments also relating to earthquake preparedness for example our next experiment will focus on what kinds of things are important in media in terms of people identifying with the situations depicted there's evidence that that sort of similarities in various dimensions are important for adopting sort of lessons learned from observed models but in the case of disasters what are the important forms of identification so our next experiment will look at identification with living situation we'll have one game that's set in portland oregon and one that's set in a generic cityscape and then we'll also have different living environments because we know that many young people don't own their own houses and maybe don't identify with that situation as it's depicted in in lots of media so that's our our future experiments all deal with earthquakes for now thank you for the question thank you elizabeth okay if we do not have any more questions incoming then we will wrap up the session for today thank you all for joining us for today's briefing i just want to say that if you are struggling to connect to any of our speakers for interviews or comments after this feel free to drop me an email at media at egu.eu we also have three more press conferences lined up tomorrow somebody exciting ones so be sure to visit the media.egu.eu page for more information we will now close this press briefing thank you again to our panelists and for all of you joining us thank you