 Good morning 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 here 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 ones which, as you'll soon see, have impacts on local communities, industries, ecosystems and the global environment. I'm Jillian 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 all 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 conference 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 are joining us virtually, please use the hand-raising function of the Zoom platform so we can come to you for your question. If you prefer to type your question instead, that's fine. Feel free to do so in the chat. Okay, so I'm now going to go ahead and introduce our panelists to make for faster transitions between them. So our speakers today are Eleni Dragotsi and the special account division ELKE, National Observatory of Athens Institute for Environmental Research and Development. We have Maria P. Kohler from the Institute of Mountain Risk Engineering, University of Natural Resources and Life Sciences Vienna, and Sylvia Nunes from the Instituto Dom Luiz, Faculty of Sciences, University of Lisbon, Portugal. So from three of our speakers today, two of them will be joining us virtually, that is Eleni and Sylvia, while we have Maria present here in person. Okay, so I will now hand over things to them to share the exciting findings with us. Thank you for joining us today. Good morning to everybody. My name is Eleni Dragotsi and I'm going to present you our study entitled Operational Estimation of Daily Dead Fill Moisture Condits, the case of the risk. Sorry. The research presented today was conducted in the frame of the Project Limax National Network. Over the past years, the Mediterranean areas are experiencing more frequent and more severe wildfires. In this context, the estimation of dead fuel moisture content has become an integral part of wildfire management, since it provides valuable information for the plummability steps of the vegetation. At this point, it should be clarified that the dead fuel moisture content refers to the moisture remaining in dead plants, including annual grass, etc. From now on, we will use the term TSMC to refer to the dead fuel moisture content. Most of the large-scale wildfire in Greece usually coincides with periods of high dryness in forest and woodland ecosystems. During drought, the risk of fire significantly increases, mainly due to the reduction of fuel humidity. Therefore, the estimation of fuel moisture content is becoming an increasingly important component in Greek early warning systems. This need motivated us to develop a daily mapping service. To do so, we investigate the effectiveness of a recently developed dead fuel moisture content model in the light of operational use. To fulfill the objectives of this study, we tested and compared two existing approaches in estimating daily DSMC. In the first approach, we calculated daily DSMC using remote-sensing data, using the no-lamps DSMC model at national and regional level. In the second approach, we produced counter-level DSMC maps from weather station data using the same model as well. For the purposes of the analysis, we used more these remote-sensing data, as well as meteorological data obtained by the dense network of automated weather stations operated by the meteor unit at the National Observatory of Athens in Greece. Here, we can see two examples of satellite-derived modest DSMC maps. On the left figure, you can see the satellite-derived DSMC map at regional level, while on the right figure, you can see the counter-level produced modest DSMC map. Here, we can see an example of a DSMC map derived from weather station data. The date of production is the date of the most catastrophic wildfire events in Greece, which were held in July 2018. The areas with the lowest death fuel moisture content values and subsequently the areas with the highest flammability levels are depicted with red colors. Due to a lack of DSMC field measurement, the validation of the produced maps was possible only in the case of the satellite-derived maps. The results from the implementation of the two previously mentioned approaches indicated that at present only the weather station-based approach meets the requirements for operational DSMC mapping. To this end, the daily DSMC mapping service that operates today by the meteor unit at the National Observatory of Athens uses weather station data to produce free access daily DSMC maps. In Figure 8, here in the last figure, you can see an example from the operational estimation of daily death fuel moisture content. Thank you very much for your attention. National Natural Resources and Life Sciences, we say actually BOKU in Austria, so this is in Vienna, but still we're going to stay in Greece, so it was nice to see the previous presentation. Me and my colleagues from the BOKU went from the University of Athens. We are working, since I don't remember when, with vulnerability. Vulnerability to make it very simple is just all these characteristics that make something susceptible to natural hazards, let's say it like that. So we are looking at houses and buildings, infrastructure, and we have done some work on tsunamis and on dynamic flooding in mountain areas in the past. But there was this, exactly this event that you see here in the photo. It was the event that also the previous speaker talked about. This is the fire in July 2019 in Mati. This is 30 kilometers away from the Greek capital from Athens. What was special about this fire was that first of all it created loads of damage and loads of costs and of course environmental damages as well. But the most important thing which was traumatic for the whole country was that we lost 102 people. So that's an amount of casualties in a European capital, more or less. It is dramatic. So we started thinking, okay, hold on a minute, let's have a look at the place. This is where Mati is. This is Greece, so not all of it, but you know, like Peloponnes, you can recognize, oops, I just have to go back, excuse me. Okay, this is the area. It's a typical wildland urban interface area, we say, where forests and buildings meet. And the fire started five kilometers away on the mountain. Somebody started burning leaves on a very dry and hot day. And this happened around noon and in the evening we had 100 dead people. So what is going on was that dry weather, high temperature and winds helped the fire to propagate towards the settlement. People didn't have any plan. The settlement was very wildly made. It wasn't planned at all. And they got trapped in there so they couldn't escape. Some of them did though. And I will tell you about it in a minute. First of all, our colleagues from Athens were lucky and unlucky enough to be there the day after to collect data from the buildings. So what happened was we see some buildings from the area exactly the next morning. You see that the roofs are burned, so in the top three photos, the roofs are burned and all these features of the building like shutters or fences or all these things that are not made from concrete, for example, let's say, were burned. Whereas other buildings weren't really damaged at all, although the vegetation around them was burned. So three questions were generated in our heads. Are all the buildings equally affected by a wildfire? The second question is, can buildings act as shelters? So in Australia they have what they mean, what they name shelter in place. So people instead of evacuating, they stay at home. But of course, this home has to fulfill some standards. And they don't have to evacuate where also it's not safe because evacuation is also not a very easy thing to do sometimes. And the third question is what can we do to reduce this physical vulnerability? So if we look at the buildings very closely, maybe we can find out that there are some very, very easy and cheap things to do to change in the building. And then we reduce costs. And of course, the most important casualities. So let's go to the next one. This is what we did more or less. I don't want to bother you with very, very complicated figures. But what we did was we went to the literature. We had to look to see what kind of characteristics are they considered as to have some sort of relation to the vulnerability of buildings. So we saw that there are many, many different things like the shutters, as I told you, the material of the building, the shape and the material of the roof, of course, the surrounding vegetation, the type of the vegetation, the distance from the building and all these things. Then we collected all this information. Our colleagues from Athens collected all this information for each building in the area. And these were around 420 buildings. And then we used statistical methods to see which of these characteristics is relevant, because not all of them were, and which one is more relevant than the other. So in this way, we could weight them and put them all together in an index. So each building could have its own wildfire vulnerability index at the end. What we did was, this is again a complicated, maybe a figure, but I will tell you what it means. It means that not all the indicators, all these characteristics that we collected for every building seem to be relevant. But some of them were. And that was that were roof material, structural type, the slope of the terrain where the building was located, the vegetation around it, the possibility for leaf accumulation on the roof, because this is also something that burns. And if the roof has such a shape or such a material that doesn't really leave the leaves go, they accumulate there and they can, they are combustible material. The material of the shutters can be, in Greece it's very often the shutters are from wood, are made of wood. Then the main ground covering and the roof type. So these were the indicators that we can more or less say that they are relevant and they have to be included in an index. Of course, we don't stop there. We would, we have some next steps planned. We would like to validate our index in another area or collect even more events, more information from new events. You see here, these are photos that I made from my car last summer in Greece. We had huge fires in the whole country, but especially on the island of Evia. And if we could get hold on this data, for example, we could provide something more reliable, maybe, or reconsider some indicators. So that would be a wish. Then it would be nice to visualize our results on a map. And finally, to include more in the intensity of the fire itself in the index. And the most important is to transfer this index in other countries. As we have seen also in Austria last October, we had a huge fire not very far away from Vienna. It didn't threaten any buildings, but it was something new for us here. It burned a large area from 100 hectares. And of course, it has become now a very hot topic also for the authorities and for the ministries and for the policymakers. So the whole thing is very relevant, I would say. Thank you very much. Thank you, Maria. We will now move to our last speaker's presentation. If we can have Sylvia on screen, please. I'm so sorry. I think you are seeing everything, right? Hi. Yes, Sylvia, you're good to go. Sorry. Good morning. So my name is Sylvia Nunch. And I'm going to talk a little bit about what we have been doing in Lisbon University with the University of Rio de Janeiro. And we wanted to create an early warning system for fire danger for the Brazilian Pantanal. And so for starters, one thing that we need to understand is that in order to have a wildfire, we need to have three main ingredients. So we need to have an ignition, we need to have vegetation, and we need to have the neurological conditions, the atmosphere in order to ignition to become a wildfire. So in this trend of the thing that is most interesting to us is the neurological conditions, because it's the one thing that is completely predictable. But how can we predict the neurological conditions? So one thing that is widely used is a thing called fire weather index. And this index is basically created and calibrated for the Canadian forests. And it's a combination on several factors as well. Rain, temperature, wind intensity and relative humidity. So they compile everything in one index and this index gives us a number. However, and 40, for example, is a number that it can be considered as high. However, it's just a number. So what is the real meaning of this number? So for example, I'm going to give you three examples. As I said, this was calibrated for the Canadian forests. So in our world, we don't have only Canadian forests, we have different types of biodiversity. And for example, if we imagine that we have this number for these three different areas, if we have for the on your left side, the first image, when you have a desert area that we don't have anything to burn, of course, even though it's a high number of fire weather index is hydrological danger, it's nothing to burn. So the classes of fire danger need to be low. In the meanwhile, on the middle, you have a pasture. So if the ignition starts, it's going to burn. However, even though it's going to burn, it's going to be a burn that probably firefighters could control it quickly. And it's not going to be as intense as for example, your right image when you have a forest, when you have more biomass, it will be a more intense and more dangerous fire. So in here, we have three different levels of fire danger with the same value for the meteorological fire index. So how can we calibrate this number in order to have classes of fire danger? So what we use is satellite information. And we have a lot of good data that's come from satellite, but in this case, we use the intensity of wildfires, the fire radiative power. So we can know how much energy a fire is releasing and how intense they are. And if we have a more intense fire, so we are going to have a more difficult fire to firefights and it's going to be more severe and it's going to have larger bunderias. So for our study, we use the region of Fentanyl divided by nine hydrological areas that you have on your right side. And we've done a model to every single one of those areas. And the model, one of the results is this cumulative density function that is a fancy name just to say that is a probability that the intensity will take a value that is less or equal than a certain threshold. Just to a better understanding, I'm going to explain a little bit closer to the plot. So in the curves, what you have is the cumulative density function for different levels of our meteorological danger. So in blue, you have the function for her lowest values of FWI. And in red, you have highest values of FWI. On your x-axis, you have from the left to your right an increase of the fire intensity. So let's imagine that we have a threshold of 100 megawatts. So if we look to the values of due to the distribution of the FWI, the meteorological danger test, the lowest values, if you look to the graphic, we see that's probably will have 0.8. So 80% of the probability of having a wildfire that is lower than 100 megawatts. So we'll have 20% probability of a larger wildfire than 100 megawatts. If you look to the other hands for the highest values of FWI, we have the complete contrary. So we have 40% of having a wildfire that is below 100 megawatts. So 60% of chances to have a fire that is going to be above that 100 megawatts threshold. So we pass from a probability of having 20% to 60% using this FWI, this meteorological danger, in our models. So with this, we created our classes of fire dangers. So we chose a threshold 100 megawatts and then we decided which probability, knowing the FWI, we would use for each class. So we have low danger, moderate, high, very high and extreme. So for the results, this was the results for the calibration of the model. The most important part is on your left side, you have the classes of fire intensity and above you have the classes of fire danger. And the most interesting part is the fact that for classes of fire intensity, the highest fire intensity, you have a shift for the very high, very high and extreme classes. So it's a very good indicator. For a better understanding, we've used this in a graphical mode. So this was for 2020, one of the most severe wildfire seasons in the region, Pantanal. And in here what you have in red, sorry, yellow, you have high classes, orange, very high and red extreme. And the dots, the black dots are the highest levels of FRP and the gray are a little bit lesser and so on. And what you can see in here is that the fire, the more intensive fire seems to follow the more extreme classes. For the future and what we want to do now is to create a prototype. So is to create a page, a web page, where this information can be presented for the next five days. And we've done this already for Portugal. And this is one example on your right side. For 2017, one of the most severe wildfire seasons for Portugal. And just for you to understand and see the redness, the most darkest areas are the ones with extreme fire danger. And the flames are the fires that were active at that time, so by satellites, so the most intensive fires. And they seem to follow very well the most extreme classes. So this is useful for, for example, civil protection or policemakers or firefighters in order to understand where in the next few days we can allocate our means in order to protect our areas. So for example, if you have a northern part that has more danger than the south, maybe you need to allocate more means like the airplanes, like drugs, like firefighters in that area, or even to prevent the ignition. So we need to go and put more police make police and so on protecting and 17 is okay. So this was my presentation. Thank you so much. And if you have any question, please. Thank you to all of our speakers. I think we can all agree that those were some interesting collections of research. And as you've probably already know, this was all under the press conference theme this morning of communities vulnerable to wildfires find ways to prepare. So now we can move on to the last part of this morning's press conference where we will have the Q&A round. And during this section, we will take questions, both from journalists in the room and those online. So I would like to officially open the floor now for questions. Just to recap, if you are in the room and if you have a question, please raise your hand. And if you are joining us virtually, please either use the hand raising option on zoom, or you can simply type in your question on the chat. Okay, so I think we have a question that's come in from online. I will ask our press assistant, Wong to please ask the question. Yeah, so from the online audience, we have one question from Nicholas. And the question is for Maria, what are the main similarity and differences in the situation between Greece and Austria? Can you give an example on how you are trying to transfer the research from Greece to applied region? Thank you. Very good question. Actually, the two countries have many, many differences. First of all, I mean, the whole the environment, how it looks like, the climate is different. But then again, it's not that much different anymore. So Austria experiences also longer periods of dry days and high temperatures. So in this respect, the two countries have some similarities or they will have in the future. The buildings look very much different. This is also a problem. But on the other hand, Austria has a very large amount of buildings in the wildfire urban interface. So in this area where our buildings meet the vegetation. How we could transfer it? First of all, we could transfer it like that and see if it works and seeing the future if it really made sense. This would be the most simplistic thing to do. The second thing is that in Greece, we made the index based on damage data that we already have. We don't have this in Austria. But what we have is we have experts and we have people that know many things and we could put them all together and try to make based on expert judgment an index. Or to make sure that we document the event so good that we will have this information that we need for an index in the future. Thank you so much, Maria. Do we have any other questions? Okay. I'm just going to pass the mic. Thank you. Those really interesting presentations. So my question is for Sylvia. So you mentioned in your talk that these sort of thresholds of danger perform quite well. You showed where all the fires are breaking out. Have you done any assessment of how much better your danger thresholds perform compared to the fire weather index that's traditionally used? We have some experience with national authorities in here that use that FWI just as a threshold. And for some regions, and Portugal is a small country, however, we have a lot of different places with different biodiversity. So we have some areas that the model performed a lot better than others, just because it has that information, the historical information. So for some areas, it's almost the same because the FWI, it works well if you have the fuel, if you have the ignitions. However, for the areas that we don't have as much fuel, as much as much ignitions, it works better for our understanding. Thank you. Do we have other questions that are from the room or online? Okay, we have a question. Just going to hand you the mic. I have a question for Eleni. So in your approach, you use satellite data, right? So I wonder if Z-data is freely available, so in the way that are the researcher or we can apply for another reason in the country? Yes, there is a modest land safety data that we use in the model. It's free access, and so we can produce the daily DSMS maps every day. Is that your question? Yeah, yeah, thank you. Thank you. Any more questions? Okay, what do you? I have another question for Maria. So are you planning to apply the same assessment for the let's say social vulnerability, for example, the population under five years old or like over a certain threshold, where they have like difficulty to move during the event? Well, I'm afraid then I would end somebody else's field and I wouldn't like to do that. So there are many people that work on social vulnerability for different types of natural hazards, and I guess there will be also for wildfire. What I see, what is for me very interesting is not the social vulnerability, but the institutional vulnerability, to see how people are organized, what kind of legislations they have, how much corruption is it involved. The land use planning, or for example, the place I was talking about, was an illegal settlement in the 60s. In the meantime, it became legal, but still there is no planning there. So there were people that they had occupied the coast, which normally in Greece is not allowed to do, and they had put the fence there so that they have their own beach. What happened was the people wanted to evacuate, and they couldn't, it was the evening also came in the meantime, so it was very dark and very smoky, and they couldn't find their way to the beach because of these things. So there are also other things, it's not only the social vulnerability as such, but also the institutional setting of countries, I would say, that also makes some things even worse than they are. Thank you. Thank you. We have one more question from our in-person audience. Hello, so I would have a question for Silvia. So I was wondering, I would be curious whether you could test your approach already with the people from the government, like that you already know if it would be helpful for them as well. In Portugal we have direct contact with civil protection and some national authorities, and actually with some companies that have some lands that they want to protect. And these kind of models were so good for them, and they use it so often, that we actually started to do this to other areas like Mozambique, where one of the companies has some lands too. So just to give you an example of how this can be used and was used, we know that in 2018 we had an area from the south of Portugal in Monchique that presented higher values of danger for that day in July, in the end of July, that were higher than the rest of the country. And the civil protection started to put some more airplanes in that area and more trucks and be prepared. Unfortunately an ignition started and it became a very large wildfire, however they were there. And I think they used this information, they told us that they used this. So yes, they say that it's very useful and we are trying to do this for Pantanal too, and probably to other areas then. Thank you, and thank you for the question. Thank you everyone. As we are now nearing the close of our press conference time slot, I would just like to say that we don't have any other questions coming in both virtually and in person. So I think we can wrap up today's session. Thank you so much for joining us for this briefing. If you're struggling to connect with any of the speakers for interviews or comments after today's press conference, please drop me an email at media at egu.eu. We have five more press conferences lined up today and the day after. So please be sure to visit the media.egu.eu page for more information. Thank you once again. Thanks.