 Okay. Now thanks folks. I think we'll kick off now. We might just have a couple of more people coming through. But just to welcome you and say good afternoon and welcome to the virtual EGU this year. Obviously, it's a virtual annual meeting of the European Geoscience Union and this year we've got an impressive 14,000 abstracts and 16,000 people from across the globe participating in the meeting. So my name is Erin Martin Jones and I'm the is EGU press conference assistant, and I'll be hosting today's press conference, which will include a question and answer session following the presentations by our five speakers, which will allow members of the media to ask your own questions. So once the last speaker's finished, please write the letter Q in the chat box to ask a question, and I'll call on you directly in the order that the questions are asked to ask your question. You're also welcome to type your questions out in the chat, and I can read them out for you if you'd rather. So hopefully this won't happen, but if for some reason Zoom suddenly quits, we will restart the press conference. I'll give you all a few minutes to rejoin the session. So likewise, if you've got some home internet problems, you can rejoin the session, and that's completely fine. Someone will let you back in. So the abstract and other documents relating to the press conferences are uploaded to the document section of the online press centre, so that's media.edu.eu. So you can please check in there for more information. So I'll introduce all of our five panellists now to make for a faster transition in between them. And so obviously this press conference is titled Improving Food Security New Techniques, and our speakers are firstly Dr Shradd Shukla, who is an associate researcher, University of California at Santa Barbara, United States. Next up we have Gabriela Jim Oroz, Nelbray, who is a research associate at the Vri University in the Netherlands. And then we have Dr Andrew Smeralt, who is a postdoctoral researcher at the Karrus Institute of Technology in Germany. And then we have Manon Bayard, who's a researcher at the University of Oslo in Norway. And lastly, we have Christina Madrid Lopez, who's research associate at the Autonomous University of Barcelona in Spain. I don't know in Fabio if Christina has turned up yet. I can't see her in the list. Not yet, no. So we'll hang on and we'll leave Christina to present at the end. Hopefully she will turn up. But if not, you can find her slides on the virtual EGU press centre afterwards. And I'm sure Christine would be happy to chat to journalists after this. So we'll run in that order then and I'll hand over to our panellists who'll be speaking for roughly five minutes each. And then after all of our panellists have spoken, we'll open up the floor for questions from journalists. So should we start then with with Shradd, Dr Shradd Shukla. Great, let's do it. Thank you. I assume you can see my slide now. Yeah, we've got your slides. Yeah. Okay, great. Well, thank you very much everyone for being here. Thank you for this opportunity to speak and a good morning from Ventura, California. So this presentation is going to be about a agro pastoral water deficit forecasting system that we are developing for West Africa to support food insecurity early warning. And this project is supported by Serbia programme, which is a joint initiative of USAID and NASA. In the region of West Africa, we are working very closely with our partners at AgriMet, located in Nimainichir and SILS. So just some key messages that we're hoping to convey from this presentation. Several countries in West Africa are prone to food insecurity, as many of you may know already. At least in the last 10 years, there are several countries in West Africa that have reached crisis level of food insecurity. What that means is that once the level of food insecurity reaches to crisis level, then emergency food assistance is needed. And drought, of course, leads to or worsens food insecurity. And this is why drought forecasting is an integral part of food insecurity. And this is what we are doing in this project. Our project and our work specifically shows that latest sub seasonal climate forecasts from multiple global climate models can help improve drought forecasting and also provide drought forecasting information at a spatial and temporal scale that is most relevant to the decision makers in the region. And as I mentioned before, we are currently building a 21st century water deficit forecasting system for West Africa, which will focus mainly on agricultural and pastoral usage. So another context here, and this is just a map that uses integrated phase classification IPC, which classifies food insecurity into five different categories, ranging from minimal food insecurity to famine condition. And what we are showing here is the worst condition of food insecurity that any of the West Africa country has reported in the last 10 years. And also the frequency with which different parts in West Africa have reported these different classes of food insecurity. So we can see that in general, there are a lot of countries which have reported crisis level at least once, which means the food assistance, emergency food assistance has been needed for those countries at least once in the last 10 years. And these are the some of the countries like Mauritania, Mali, Burkina Faso, and Niger and Nigeria and Chad, where we are focusing our work on. And that is mostly because those are the countries which are typically, which have typically reported food insecurity more often. So another context for this, we conducted an extensive survey of climate service providers from national meteorological and hydrological service agencies in many of the West African countries. And we found that there is a general need of going beyond course scale seasonal forecast to some to a sub seasonal forecast, which are able to provide us drought information at a better spatial and temporal scale. So to keep keeping that in mind, we're working on a water availability forecasting system for both agriculture water use and for pastoral water use. And a key kind of tool for developing this forecasting system is against sub seasonal scale forecast of both rainfall and even pretty demand from multiple global climate models. So this is just an example of how these products look like. So on the top is a product that's useful for agricultural water usage. And what this map shows here is that the places where we, which are in green, there's a higher likelihood of above normal crop production. And the regions that are shown in brown, they have higher likelihood of below normal agriculture production. So these are the kind of the maps that we are providing in the region. And also in terms of water level forecast or pastoral usage, this is another map that comes from Fuse Net Water Point Monitor. And again, in this map, the places that are in green, they are indicated to have good level of water. The places that are in red, they have near dry water level. So as an example, we are using multiple climate models for providing sub seasonal scale forecast. And these forecasts go out to the next 30 days and they are updated every week. So for example, this is a forecast that was provided on the last Thursday of April 22nd. And here is just a map that shows that these forecasts are skillful, especially in the critical months of June to September. All the regions that are in dark red colors are the ones which have higher skill in the region. Finally, another important result that I wanted to share is that what we are finding is that there is a stronger connection between estimates of a monthly available moisture and seasonal vegetation health in West Africa. And that the connection is stronger in those regions of West Africa, which are shown here in red color, they are the regions that have experienced food insecurity much more often than other places. So you can see that the correlation basically between both of those variables is stronger in the places that are in red and generally weaker in the places that are blue, which just means that places which are less food insecure, this connection between climate and vegetation is weaker. So with that, I would like to acknowledge all our partners from different institutes. And I would also like to welcome you to follow Climate Hazard Centre and my Twitter handle for more updates. Thank you very much. Okay, thank you Dr Shukla. Again, if anyone has any questions, please sort of save them for the end once we've heard all of our five speakers. So next up, let's move on then to Gabriella. Dr Nobre. Yeah, let me start by sharing my screen. And I hope you can see my presentation in full mode. Okay, great. Then again, good afternoon everybody. It's a pleasure to be here. My name is Gabriella Nobre and I'm a researcher at the Dry University Amsterdam. And in the next minutes, I would like to give you an overview of the project that I've been coordinating that is called the Forecast Space Financing for Food Security. And this project was conducted in partnership with the 510 initiative from the Netherlands Red Cross with the Climate Hazard Centre that Shred just gave his overview at UC Santa Barbara and the Kenya Red Cross with also funds from the Global Facility for Disaster Risk Reduction and Recovery, the Foreign and Common Wealth and Development Office and the Centre for Global Disaster Protection. Before I start going into the project, let me give you an overview of the problem that we tried to approach. So as you may be aware, when an extreme event happens such as floods and droughts, an increased number of humanitarian organisations provide affected population with cash in order to support livelihoods facing critical levels of food insecurity. The problem is that these humanitarian assistance, they often reach the population too late when their basic needs are already deteriorated. However, the impact of these hazards, they can be reduced when forecasting formation is available to trigger anticipatory action. Therefore, we believe that there is a growing opportunity for humanitarian organisations to trigger and also to implement cash transfer within the window of opportunity between the issue of forecasting information and the materialisation of the event. So our F4S project worked within this window of opportunity by providing tools and evidence on how this early action could reduce the risk of food insecurity. And now a project we did that by investigating three pillars. That's what I'd like to talk to you today in the presentation. The first one was forecasting. The second was through a better understanding of the local context. And the third was through an understanding of the benefits of acting early. And I'd also like to highlight that the project had Kenya, Ethiopia and Uganda as case studies. So an important feature that I've just mentioned of our project was the attention we gave to better understanding the local context and the livelihoods of the people that we are trying to support in parallel to developing these forecasting systems. So for these, we carried out household surveys to better understand both their communities and also their local knowledge on early warning signs of food insecurity. And we found that early action is often adopted by the communities once they have the information of an upcoming hazards. But once hazards hit the community and these impacts are felt, most of the households implement some sort of negative coping strategy. For instance, they reduce the number of meals. We have also observed in terms of local knowledge that there is a large variety of local knowledge that these communities have and that they also use their local knowledge for guiding their decisions. Also, we have performed a choice experiment to better understand people's expenditure choices if they would receive cash transfer prior to a shock. So we did that in a sort of a game by playing scenarios in which some designing elements of a cash transfer program were tested, for instance, the payment format. And we observed that the expenditure choices can change depending on certain designing elements. So for instance, here on this graphic we see that there is a higher share of the aid that would be spending food expenditure if these beneficiaries would receive aid in small payments instead of a lump sum. So we think that information like these is important for understanding ways in which these cash transfer programs could maximise an outcome. As concerning forecasting, we focus on developing models that can predict key components of food security. And after we identified some food security indicators that are relevant for a certain geography, we came up with three indicators that we could potentially forecast. The first one was the shortage on calories in which we forecast where a certain percentage of a population would be likely or not to experience caloric shortage and also we forecasted forage discussities, which is an indicator that links well with livestock mortality in pastoral production systems. And lastly, we forecasted transitions in the state of the food security, which tell us months ahead whether the levels of food security is going to deteriorate, improve or remain the same. So in a nutshell, what we have observed was that with the aid of machine learning, but also in combination with local knowledge and relevant biophysical and socio-economical information, we were able to forecast up to three months ahead of a short shortage on calories for all the agricultural and agropastoral regions in the maps. We were also able to forecast up to four months ahead forage discussity for pastoralist districts in Kenya. And also we were able to forecast up to one year ahead transitions in the state of the food security in Ethiopia. So overall, what we think that's interesting about these is that providing this accurate forecast on key indicators of food security longer ahead opens up a wide window of opportunity. The last pillar investigated in our project was the benefits of acting early based on forecasts rather than exposed when it's often too late. And we did that by carrying out cost-benefit analysis. And what we have found with the cost-benefit analysis is that there's a larger range of benefits that can be generated if the financial support reached the communities earlier. So, for instance, in pastoral communities in Kenya, we found that each Kenyan shilling invested in early action yields 3.6 in benefits. And this is due to the fact that early payments allow communities to use the money to protect their livestock instead of replacing them. And in addition, we have also found that acting earlier can also be cheaper. So here in blue, we show all areas in which a higher reduction in cost per beneficiary can be achieved if this cash transfer is dispersed prior to a shock instead of after. So some key message of our project overall, we learned that weather-related hazards often lead communities to implement negative coping strategies. However, forecast information is often trusted by the communities. And we also found that, including local knowledge, we can also produce accurate forecasts of indicators of food security long ahead of a shock. So this combination of trust and lead time opens up a wider window of opportunity for implementing anticipatory action at the short and also medium term, which can also avert some of these negative coping capacities. We have also found that the design of X under cash can have an effect on people's expenditure and therefore is important to further investigate these co-designing strategies between institutions and beneficiaries. And lastly, we have also found that despite saving lives and creating a wider range of benefits, early cash can also be a cost-effective solution. Thanks very much for listening to my talk in the open floor for any questions, if any. Thank you. Thank you, Dr Noelbray, for that introduction. So we'll save questions till the end and press on. So next on the list is Dr Andrew Smeralds. If Andrew could come forward. Thank you. Okay, I'm muted, sorry. Great, we've got you now. Thank you. Hi, so we've been looking at the cost of producing more food. So in particular, what kind of by-products are produced in terms of nitrogen? So just as a taster here, the map is showing the cost of producing a bit more maize and the cost in terms of greenhouse gas emissions, where red means high and green means low. So sorry anyway, it's late, I'm not moving on. If you click on the screen it should work again. Sorry, apologies. So the background to this is that there's quite an industry in predicting what future food needs will be. So this depends a lot on the assumptions about diets and about food wastage and population growth and so on. But the consensus is that we will need more food in the next decades and supposedly maybe 25 up to 70% more by 2050. So how do you produce more food? Well, there's two obvious solutions to this. Either you have more cropland, you expand it, or you work the cropland more intensively that you do have. And so there's a consensus that the intensification is likely the better way to go because it seems to produce less biodiversity loss for fuel greenhouse gas emissions and also just there's only so much good farmland left available. So this brings us on something called yield gaps. So what is a yield gap? The yield gap is quite a simple concept. It's just the difference between what's actually being produced in a region and what could potentially be produced with current technology. So if you want to intensify, then you need to close these yield gaps. There's many different strategies, of course, to do this. You can mechanize more, you can employ more human labour, many other ways. But the one thing that's common to all of these strategies that you need more, you need sufficient nutrients for the plant to grow. And the main nutrient is nitrogen and this comes as fertilizer or manure or something else. So the problem with adding extra nitrogen to get more plant growth is that it comes with some harmful environmental consequences. One of these is nitrous oxide emissions. So it's a gas. It's produced by microbes in the soil and is responsible for roughly 7% of global warming. And if you look at the graph on the right here, you see this hockey stick like figure over the last 2000 years. So similar to carbon dioxide, nitrous oxide concentration in the atmosphere has turned up sharply in the last 100 years or so. And the main source of this is fertilizer or the main anthropogenic source of this. The second problem coming from all this extra nitrogen is nitrate leaching into groundwater. So nitrate is necessary for plants. It's one of the ways they take up nitrogen. But once it leaches out of the roots, sewn into the groundwater, it's both a pollutant. So, for example, in Germany, there's strict laws on how much is allowed to be in tap water. And it also causes what's known as eutrophication where algae grows and dies and consumes all the oxygen and then you get big dead zones. So, for example, here in the Gulf of Mexico. So what have we been looking at? So the question we already asked is if you start to close these yield gaps, how much of these extra harmful things will you produce? Or if I say this in a slightly more optimistic way, where can additional food be produced at the lowest cost? The way we've been looking at this is we have a computer model which tracks nitrogen and carbon water and plant growth and so on all on an hourly basis. We run this on a global scale and as you can imagine, this requires a lot of computer resources. So a couple of results. So I show you something for maize. So maize is one of the main global crops. So maize, rice and wheat produce a lot of global calories. So the first plot here shows the how big these yield gaps are. So basically red means that the yield gap is very small, so there's not a huge amount of potential to close it, whereas green means a very big yield gap. So plenty of room to produce more food. So this in itself is not a particularly new result. This other people have looked at as well, but of course it's nice to know that our model is consistent with what others have found. What we then add to this is say, well, if you do produce a bit more maize, how much of these greenhouse gases will be produced? Where again, red here means high and green means low. And one thing you can notice is the places which already have high yields or low yield gaps also would produce a lot of greenhouse gases if the yields were increased. So for example, in Western Europe or China or North America. And this is showing you that this greenhouse gas emission is non-linear. So the first unit of fertilizer you add is most of it ends up in the plant. By the time you get to the hundredth unit of fertilizer, a lot of it is being converted into other things. We can do the same thing for nitrate leaching. The pattern is not so different. One thing you might notice here is in India, producing more maize seems to cause quite a lot more nitrate leaching, particularly in East India. And one thing this is highlighting is that soil and climate are very important for this as well. It's not just helping the yield gappers. So how expensive is it to increase food production? So one thing we looked at is, well, let's imagine that we start closing these yield gaps and we do it such that food production increases by 25%. So at the lower level of what's expected to be needed by 2050. And let's imagine we just do this in the optimum way, so the way which produces the least of these nitrogen byproducts. If we do that, we find maybe 20% more of this greenhouse gas nitrous oxide, 25% more nitrate being produced at the same time. The good news then is that, well, it is possible to produce more food without increasing the amount of nitrogen losses per unit of food. So this is not completely obvious given that I claim this was a non-linear type effect. The bad news, of course, is there's no free lunch here. More food means more greenhouse gases, more pollution at a time when really we'd be like to be reducing these types of things. So then what is the implication you can take away from this? I think, well, closing yield gaps is clearly important, particularly because they're generally biggest in the places that are most in need of some extra food over the next years. But if you actually want to make a dent in greenhouse gas emissions, it has to be combined with also fertiliser reductions in ritual regions. So okay, this is basically what I'm going to say, so thanks for listening. Thank you, Dr Smarrod. So let's head on now to Dr Manon Bayad. If you can now come forward and share your slides. Yes. Thank you, Manon. Excellent, we can see them. Good afternoon everyone. My name is Manon Bajar. I'm a researcher at the University of Oslo. This week at EGU, we are presenting a study entitled Climate Viability Control Development of the Pre-Viking Society during the Late Antiquity in Southeastern Norway. This study is part of a project called Vikings. It's a interdisciplinary project involving geologists, biologists, archaeologists to better understand the role of volcanic eruption on climate and society. This afternoon, our main question is how did past societies respond to climate change because we need to adapt our agricultural system to present and future climate change to maintain and improve the food security and take part in the climate action. So how we do that? A way to do this is to learn from the past, to study past cases. So hundreds, two thousand years ago, look at the climate viability and how the society at that time responded, if they responded or if they collapsed or if they adapt and how they adapt. We found here first illustration of early adaptation of societies in the pre-viking age to climate changes and we also provide new insight into this society between three and eight hundred eighty, a period called the Dark Ages because we don't know so much about the society of these periods that is between what we know very well about the Roman, the antiquity and the Viking and the Middle Ages. To learn from the past, we are actually studying a lexediment course. Those are natural archives of the environment. It is very similar to ice course, but here it's not ice. This is the mud that is accumulating year after year in the bottom of lakes. So we have some layers to reconstruct year after year, the evolution of the environment. So it's a continuous record of the environment. It contains particles of soil through erosion and runoff and it contains so pieces of plants. It contains pollen from the vegetation and also DNA of plants and animals. And it also includes the information about the biological life of the lake that is dependent, for example, on the temperature. So we also have here a record of the climate in this sediment course. At the same time in the same archive that we can date with C14, we have a record of climate, vegetation and human activities. Here we are studying a lake that is close to the airport of Oslo in Norway and we reconstructed the temperature with the chemistry of calcium on the period between 200 and 1380. And we have one period at the end of the Roman period and also one period in the Viking age and in the middle age. So here you can see that the Viking age started with an increase in the temperature. And between these two one period, we have a colder period. I refer that the dark age is cold period so it was much colder during this period. And then we reconstructed the agricultural practices by counting first the pollen of cereals that we find in the sediment. So, for example, here we have pollen of rhai, wheat and barley. And I like it in the orange period where we have more of these pollen of cereals, meaning that we have more cultivation of cereals during these periods. And finally, we reconstructed husbandry activities by measuring the concentration of sodaria that is a fungi that is developing on the feces of animals. So, if there were animals grazing around the lake, then the feces will be washed to the lake with the rain and then we can find this fungi in the sediment. And I like it in green periods where we have obviously grazing activity. And what you can see here is that there is an alternation between a period with more cultivation of cereals and a period with more grazing activity between 200 and 800 AD. And if we compare that to the climatic reconstruction on top, we can see that when it was warmer, we had a cultivation of cereals and when it was colder, we have more grazing activity. Here it's a little bit warmer, we have more cultivation of cereals, colder, more grazing activity, warmer, more cultivation of cereals, less grazing activity, and so on until the Viking Age. So, we show that we have a dominance of cultivation of cereals, but also hemp during one periods and on the contrary dominance of grazing activities during cold periods, suggesting that already 1500 years ago, the society before the Viking Age was able to adapt their agricultural practices to climate. If you want to know more about our project, you can use this QR code or follow me on Twitter to know more about our activities this week at Egypt. Thank you for your attention. Lovely, thank you, Manon. Okay, and last but not least, let's head over to Dr Christina Madrid Lopez if Dr Lopez is around. Yes, I'm here. Excellent. So, you're able to share your slides, okay? I think so, let me see. Good, yeah, we've got them. Okay, super. Perfect. Okay. Hi everyone, thank you for being here today and thanks for the invitation. I wanted to tell you a little bit of the word that we developed at the Universitat Autónoma de Barcelona and the Sostenibria Research Group, where we examine how sustainable can urban agriculture be. So, we depart from the issue of developing urban agriculture in a world in which urban food demand raises in between 50, or is suspected to raise in between 50 and 60 percent and where long food chains have environmental issues. And then when there is this tendency in urban areas to develop agriculture because it provides not only food but also some of the environmental and social services. And at the same time, we have agriculture being the main water user and being a resource that is under highly competition with other uses in urban areas. And most recently, the European Water Framework Directive has been analysed and has been detected that the coordination with urban development, it's a little bit weak. And so the aim it has about improving or reducing vulnerability of water bodies, especially important in urban areas. So in order to try to answer this question and to have in mind these issues, we analyse a region in Barcelona, which is a metropolitan region in Barcelona, in which we have about three million inhabitants, but we welcome about in between four and five million every day for work. And nowadays we cover 10 percent of the food needs, of the fruit and vegetable needs in the local area. And we find ourselves at the moment developing a new urban development plan. So we do this study within the European Project Urbag, which is led by Professor Gary Alba. And I leave you here a picture of my colleagues, Joanne and Sergi, with whom I'm also working in this research line. So in this project, we combine or we integrate climate and hydrological models with models of land use to assess how green infrastructure in general and also agriculture can change the local climate and how the changes in the local climate can influence urban agriculture. So what is new in this particular study that I would like to introduce today is that we are doing a georeference analysis that is usually not the case in here. It has a monthly resolution and it also connects the vulnerability of river basins with the locations where the water is being used. So just for not being very deep into the details, feel free to email me if you need more information. But what we do is we map the agriculture and the related water extraction points. And then we associate those extraction points with the river basins or the aquifers. And then we use that geographical connection to assess the changes in the vulnerability status of the river basins or the aquifers. And just a couple of highlights, the type of results that we are receiving. This is, for example, a comparison between four scenarios where a zero is the current scenario where agriculture is 8% of the surface of the metropolitan region. And then one scenario that would be reducing agriculture if the current urbanization trend continues. Then we also have two scenarios, one of them would be to revert that trend and the other one would be to use the full potential for agricultural production in the land of the metropolitan region. So do you see that, for example, this graph is showing water use by month. So the water use peaks or starts to peak in April. And that is not only because of the weather changes, but also because of the type of crops and the crop calendar that we have in the region. And you see how, for example, because of the type of crops produced in the different scenario, even though we are reducing the area, we are increasingly increasing water use. So taking into account the type of crop and time actually matters. If we assess what does this mean for the local water resources, we will go from a situation in the current situation right now is that we have most of our river basins in a very bad status with the rest not being that well. So we do see that they are all in indexes for three, four and five, which are the worst ones. And this is for the agriculture that we have right now, that is 8% of the territory. But when we go to up to 20% of the territory, for example, that means that potentially, if we do not consider to regulate the type of crops that we are growing, for example, and the time when we are growing them, then we will reach a situation which is actually against the principles of the water framework directive. So that is to say, the take home messages that we have for the moment because the study is an ongoing effort that urban development needs to consider the crop calendar and maybe create a list of about crops, for example, for urban agriculture, in the same way that you have a list of requisites for the buildings and the type of buildings that you can put. And then that we have already tools to constrain the requirements of urban development in terms of considering agriculture and, for example, the water framework directive, vulnerability indexes could be one of those tools. And this is the little bit that I wanted to show you. This is our session, which is next Wednesday. If you're around and you want to join, you're very welcome. And here you have some details on how to join the network of the URVAC. It's a project that is going to be running until 2024. And here you have my email if you have any more questions. Thank you. Thank you, Christina. And yeah, thank you to all of our speakers and their fantastic presentations there. So now I'm going to open up the floor for questions. And as before, if you have questions, feel free to either type a cue in the chat box and we'll head over to you and you can read out your question. Or I can also read out any questions that you might have as well, if you'd rather type them in. OK, so James has a question for Manon. So, James, if you'd like to unmute yourself. Yeah, can you hear me? Got you, yeah, yeah. Thanks for the talk. Yeah, it's just it'd be good to hear a bit maybe about the process by which you extract these cores from the lake. And also, I know with ice cores there's an issue sometimes where when a glacier starts to melt, you get mixing between the different layers. So if you're talking about a lake sediment, is that a challenge to try and unravel the mixing that might have taken place between the different layers? So for the lake sediments, so the sedimentation, so the deposit of the mud started when the lake started to form, actually, when the lake started to be a lake. And this happened in Norway, Scandinavia, after the melting of the glacier, so 10,000 years ago. So after that, we don't have the glacier any more here in Norway. And then we have a continuous record without any glacier advance that can disturb the sediment here. OK, thank you, Manon. OK, so we have a question from Sarah, Sarah Darwin, and Sarah says wonderful talks and a question for Dr. Nelgray. I'm wondering if you can go into more details about the monetary support for communities. It looks like a great return on investment. What is the money usually used for? Are these temporary stop gaps like for one season or more long term solutions? Thanks for the question, yeah. So let me let me see if I understood your question. So the question is about how usually is the money spent either by the donor or perhaps also by the community. In terms of the donor, I think that this idea of forecast based action in which you can support the community months ahead of a shock is relatively new. It's well, let's say the steering of the community started by around 2015 in really recognising that this forecasting information is useful and can be used for better planning of these anticipatory action activities. And well, if you look at the records of donors spent, most of these money is usually spent at the recovery phase. So when the communities are already heated by a hazard being that a flood or a drought, these money is usually spent for recovery activities. And what we're trying to advocate with forecast based action in also with our project in general is that there is this window of opportunity for shifting some of this money. We understand that, well, the impacts, we can't reduce all the impacts with this type of anticipatory action, but we can reduce some of them and some of them they can be given to the communities early in advance. So for the communities to implement type of activities that I could name, for instance, if the money reached the community months ahead of a potential delayed staff of the season, the communities can perhaps use the aid to have access to more drought tolerant inputs for the agricultural practice. So in fact, these communities, they can try to better prepare the agricultural systems to accommodate some of the shocks. If the shocks that they already happened, let's say, within the growing cycle and the communities have they already planted, what these aid could do, for instance, is to reach the population early in advance so that the population can reach the market at more affordable levels, because usually what is observed, once they already experience this duration on their basic needs is that there is a very, it's very difficult for the communities to have access to the market because the prices are already so increased that on top of not not having a good agricultural ear, which often sustain the assets and the built these economical buffers, these communities, they often have these assets deteriorated as well. And on top of that, there's experience with the fragility on the market where the prices are usually higher. So these people can, the communities of beneficiaries, they can use these aid in order to try to reduce some of these buffer and reduce some of these impacts. So this is for instance some of the examples that we have learned from interviewing the community and trying to understand how would they utilise these aid if they would receive these aid and support early in advance. Thank you, Gabriella. So we have a question coming from Andrew. The natural dioxide over the last 2000 years, out of the modern period, does it reflect old historical agriculture emissions or is it a real natural background? Yeah, so there's lots of sources of nitrous oxide emissions. So from wetlands, I mean from natural lands, from forests, from wetlands, from things like this. The thing is over the last 2000 years it's been relatively balanced. There's sources and there's sinks. So the stuff coming from your swamps or whatever is being balanced out by nitrous oxide that is then being taken out of the atmosphere by reactions with ozone or something. So this is where you have a constant level. What's changed over the last one to 200 years is an intensification of agriculture and particularly the harbour bush process, the process by which you make fertilizer. So this has basically put a lot more nitrogen into the system and this is mostly put into agricultural fields of course and it's this additional nitrogen in the system which then gets, which has then changed the balance between the sources and the sinks. So the sources are now stronger than the sinks let's say so there's an increase in the atmospheric concentration and if the source stays constant then of course this will rebalance out at some higher level but at the moment the sources have been constantly increased as more fertilizer gets used. If that's an answer to your question. Yeah, sounds good. Okay, thank you Andrew. Any more questions to round up? Okay, yeah. One for Dr Shukla. What sort of information do you need to make sub-seasonal forecasts and how do you use this information to inform groups and communities like those Dr Noel Gray works with? Great. Thank you very much for your question. So there are many different approaches that can be used for sub-seasonal scale forecasting. The approach that we are using uses many different global climate models and I think in scientific language we call those a dynamical climate forecasts. So what we basically do is take these five or six different models and each of those models they provide different numbers of scenarios of what climate may look like over the next 30 days. So for example with these five six models we can get up to 60 different scenarios of what next 30 days would look like. So what we are doing now is using those climate forecasts to then drive models that can provide us estimates of available water for crop usage or available water for pastoral usage. So basically trying to translate these climate forecasts into something that is more meaningful to the decision makers and also the final end users like farmers and pastoralists. One thing that I didn't get to highlight in the presentation before is that we are working very closely with our partners in the region of West Africa, mainly and also we are doing several rounds of capacity building activities with national hydrological services in several countries. So what our approach to making sure that this information is actually making it to the final end user which are farmers and pastoralists. Okay. Dr Shuklau, we seem to have lost you. Can you hear us? If not we might have to come back to you if that's okay and we will head to a question for Gabriella. Can you say a bit more about the machine learning aspects of your forecasting tools? Yeah, it's a pity that the shred just dropped because it would be a great compliment to their question actually. So I shred was mentioning this. Sorry shred, we lost you. I don't know if we want to maybe give the floor for shred to finish his question. Do you want to go back shred? Yeah and then we'll jump to you. Sorry about that. Now I would just think that the one final point. Oh no, you've broken up again shred. How about now? You're okay right now. I'm sorry, yeah you're really breaking up. So yeah we've been really lucky with Connection so far. Do you want to put something in the chat shreds just to round up if you want to? Yeah, if you're able to. I'll do that. Okay sounds good. Thank you. And we'll just head over to Gabriella's related point which will probably be our last point before we wrap up. Thanks. Sure. So well shred was mentioning the dynamic type of forecast which comes from these well global models with different realizations of the climate. Actually our approach by using machine learning is taking more of a statistical based approach. So you can think of machine learning as just a type of forecasting tool as you would use for instance with simple linear regressions but in our case we have used try to explore the use of machine learning. So what we feed this machine learning with it's a combination of many data sets. So we have some core data sets that we try to extract information. For instance we can play around with precipitation, precipitation accumulated with a soil moisture with a greenness measured by satellites but also with more infrastructure type of indicators such as the distance to the markets that people have or informations from the previous seasons. And then we can see as a these are machine learning as some kind of bucket where we feed our models. And this machine learning we do have the flexibility to train the model in order to maximize a outcome. So we can train until we find a range of combination of parameters which we can tune this decision. These machine learning to a outcome that we desire. So in our project we have obtained at the end three different machine learnings because we were targeting three different indicators of food insecurity. So the difference as Shred was discussing from his approach and through our approaches because his approach he used let's say more dynamical type of forecasting and our model is more statistical based types of forecasts but that at the end they all try to capture these information on looking ahead for shark in the future. So thank you Gabriella. So I'm afraid we're going to have to wrap it up shortly as I'm heading over to the next press conference but Shrad has just dropped a comment in the chat and I don't know if you have time to quickly read it through but you can also follow him on Twitter Shrad's contact details are there and likewise for all of our speakers today please please feel free to drop them an email. I'm sure they'll be more than more happy to answer your questions so thank you again for your for the speakers for their time and for all of your questions and in the meantime if you'd like to head over to the next press conference which is called Scientific Sloothing, very exciting title that will begin right now I think presently so I'll hop over there and any more information that you should need please have a look at themedia.edu.eu website. So thank you for all coming. Bye.