 So we, yeah. So thank you everyone for joining and this knowledge cafe on research communities and climate action and being open to drive change. So I'm just going to give a brief intro on what this session is and what it's part of. My name is Ben cat. I work for open air, which is an organization based in Europe, and we're running this series of events and this particular session is part of a series of things happening this week. But before I go any further, I just want to give some housekeeping rules. So, as you will probably just notice this event is being recorded. Please make sure that the microphones are off during the speakers when they're speaking. We do have the chat function and please use that introduce yourself interact with other participants and also but please importantly this is where you can ask questions to the speakers. We will give the opportunity to for you to speak as well so please raise your hand if you would like to do so and then we can do it that way. The presentation and the recording will be updated in the event page as well and at the bottom you can see a series of hashtags that you can also use to tweet about this event. So I'll just give you a brief intro of what this is all about. So this is part of the open access week 2022. And open access week has actually been running for a number of years now, since 2008. And it's an international event. Many organizations around the world participate. And each year there's a different scene. So open access week was actually, I should say as well set up by an organization called spark. You can actually find out more I'll paste some links later on in the chat window that you can actually follow up on. And I should also thank Irina Kuchma is actually very closely involved in organizing this too. And indeed there's lots of other people I am not going to try and mention all the names right here. The theme of this year is all about climate justice. And what exactly is that. Well, we've invited expert researchers to discuss this theme and about the research that they're conducting. So we're trying things like oceanic atmospheric and earth sciences and how does impacts upon trying to better change what is happening with climate change, monitor it and try to provide solutions and climate justice in itself means also. We need equity across the world to because certainly in low middle income countries for example. We need a more hard hit perhaps and we need to try and regress the balance and so this idea of climate justice is also being looked at. But to actually be able to do this, how do you do it and open science is a big part of this equation we believe this idea that the data generated through research should be especially public research or should say is made open to all. So I want to foster this change in the attitude that researchers conducted and open science is a big part of that. So in all the cases that you'll see following this the research will be reusing and also providing data to the global community and they will be using technical infrastructures to actually facilitate this openness as well. So we have four speakers for you, and I will introduce them right here. And then let them speak in order but we have beyond back of her from the sea scale project and through them from reliance. This is Ella de Oliveira from EGI script EGI ace project, and Professor spirit on raps on my net case, who is from the name is project of these projects are based in Europe. And you will get an idea of the research that they're doing and how this is plays a part in open science. So without further ado, I will ask the first speaker please be on to share your screen and take away and thank you for listening. When sick while I get my ducks in a row quickly. For some reason it doesn't want to allow me to share. I have to quit and reopen I'll be back in two seconds. Sorry, apologies. May I suggest, and if you are ready to speak that maybe you would like to go now. Would that be okay. Yes, if you want me. Be on the spot already. Yeah, okay. Okay, sorry. Take away your apologies for that. It seems that I needed to enable a lot of my computer to share the screen but hopefully you're seeing my screen right now. Yes, thank you. Okay. So sorry for that technical glitch and glad to be back. So hello everybody. My name is beyond back a big I work at a research institute called Delta Iris in the environmental hydrodynamics and forecasting department, and I'm going to try and give you a perspective on the importance of open access data for climate justice from a more ocean climate dynamics research perspective which has been the sort of the work that I've been doing for most of my my research career. So, so I'm going to I thought it might be useful to give you a bit of an overview of the kind of research that I do which which is both computing and data intensive. And I thought to start out with what I did for my PhD, which was modeling the music scale scale variability in the ocean so music scale variability is essentially all of these meanders that you see in the ocean currents and the little eddies that are moving around. The key topic for my PhD was investigating different numerical methods that improve the accuracy of the model solution so here in this example in the top here you see satellite derived currents. And you see that there's a very well defined current along the southeast coast of South Africa. The satellite observations but then when you try to model these in sort of state of the art numerical model, you find that it gets these exaggerated rings, which is not representative of reality. These two bottom figures here are essentially the same but they describe authenticity which is essentially spin in the ocean. So if it's red, the currents are spinning in a red in a anti clockwise direction in the southern hemisphere and blue is the opposite. So what you see here is is the representation of spin or vorticity based on this image here at the top, and then when we apply in the model what we call a fourth order momentum addiction scheme. So just a more complicated way to rep to model currents, we find that we are able to represent a more explicit ocean current, which is an improvement over this situation on the right here. Now having said that if you're implementing more advanced numerical methods, usually these come at some computational expense so they are more computationally heavy, but they do come with advantages. Another way that we can improve our model solution is by just increasing the, the resolution of the model so going from example in this case, 10 kilometer resolution in the horizontal to for example double that five kilometer resolution. And we did this, and we found that, in fact, actually what we see it doesn't really solve our problems in this. This is on the right we have this lower resolution model simulation you still see these exaggerated eddies, propagating into the Atlantic following very similar trajectories, as was the case in the previous image. And then on the left is like a model with double the resolution and you see a lot finer detail in the model solution but in fact the problem of these exaggerated eddies is still retained. So improving the model resolution doesn't always solve all of our problems, which was the main message I wanted to bring here. But having said that if you double the model resolution, you're now making your model twice as heavy essentially in terms of the data that is generating and from a point of view of running the model it also takes a lot longer. I then I then moved from sort of understanding the numerical methods to try and improve the model solution to looking at ocean data assimilation. And this is a method where we basically use ocean observations to constrain the model errors. And what it basically does it improves the timing and the positioning of in this case music scale features so here you have a nice example. If you focus on the top here. There's a blue Eddie or a blue dots that's propagating south this is a cyclonic Eddie, which means you would expect features to move clockwise around them. And these black dots that you see here are drifter observations these are actual observations, and you notice that they go right through in an Eddie which is not what what should be happening they should be going around it. And when we apply data simulation in this case at the bottom here we've assimilated along track altimetry we find that it positions the eddies correctly so that the drifters are actually moving around the features as as you would expect in reality. So now we're faced with the solution where we have, we've got a computation the expensive model that produces a lot of data, and now we're also trying to take observational data and ingested into the model to correct for errors. And so these challenges around making the data. Talk to your to your numerical code or your data simulation code is challenges around standardization for the difference numerical date the model that you have the data that you simulate. And I thought it might be useful at this stage to give a quick example about the sort of scale of the problem. To do that I thought it might be useful to just give you a quick run on on what data simulation actually does. So here's a very, very simple analogy. So the problem that we're trying to solve in this case is we are we are lost at sea, and we want to find our position. So we have a model that can estimate the speed of the boat. So by, and when then we have observations that we can use to estimate our position from the stars, but both of these estimates have errors. So now, our assumption is that the truth, the green the green bar, but here so where we are actually positioned in space and time is somewhere between the observations that have error and the model that also has error. So now, finding that truth or that analysis is basically what we do is we apply a waiting factor to either them to both the model and the observations so essentially, we make some decisions around which one do we trust more the model or the observations. And then effectively you assimilate those observations into a model to give you an improved estimates based on that waiting. And your, your accuracy is essentially been a choice for of K you need to basically fiddle with this parameter K to minimize the variation of your true of your analysis or of your truth. So then your analysis is essentially a combination of your model and your observations so remember computationally expensive model runs produces a lot of data. And there's a lot of data that are being ingested into the model itself so it's quite a computationally and data intensive process. So that was a very simple analogy but if we sort of convert that into a sort of more realistic application, typically what we what we find is, we have in a in a model system we have about 10 to the eight variables so that's a lot of variables. So the number of observations that we have to constrain those variables are like less than 1% of the entire domain of the model. So then, in order to constrain this model with such few data or such few observations, we need to calculate what we call a multivariance multivariate covariance matrix, where we basically figure out what is the relationship of one observation with the rest of the model grid. And to get there we have to, at each point in the model, project the model into the observational space, then construct a covariance matrix between the model and observations, then weigh this covariance matrix according to the model and the observational error covariance and then multiply by the difference between the two. And in that way we calculate the analysis and then our relatively simple, a very simple equation looks something more like this where you have an analysis which is a matrix, a model field which is a matrix. Then you have a model covariance matrix at the observational positions, which is a very, very large matrix, which you then have to divide by two other fairly large matrices, namely the error covariances. And then you multiply that by another matrix with the innovation. So you see these are a lot of complex computationally intensive calculations, matrix calculations that need to be done. So, and that's just to constrain the model at one time step. So whenever you run the model, you run the model, say for argument's sake, for seven days to produce some forecast, you let the model diverge in terms of its error. And then you stop the model, and then you run this whole system to bring the model back, the model error back to some, to some state, and then you rerun it again. So now when you combine models and observations and climate research, you need to do this over long, long, long time series. And so we need to be able to run our model for, you know, 50 to 100 years to really understand the climate dynamics. And here's an example of the kind of output that you could see. This was actually derived from satellite observations. And what it shows you is the, the energy in the ocean, and it's long term trend. So you see areas of red is where the ocean currents have increased, the mean ocean currents have increased in terms of red. So you see here north of north of Madagascar and along the east coast of South Africa there is an increasing trend. And then along the, but then the changes to decreasing trend. And overall, if you look at the variability, so these is the eddies. So the red areas just indicate where the variability has increased over the sort of satellite record. To really. So this image was calculated based on satellite data. And satellite data typically only covers the surface of the ocean but you really want to understand what happens in the interior in order to really quantify the mechanisms that are responsible for this. And for that you really need to combine these data similar to ocean models with into two observations and satellite data to really quantify the processes that are contributing to these long term trends. So you start combining vast amounts of data from different sensors from different platforms, there's satellite platforms there's in situ platforms. And then you're combining those with you know computation computationally intensive models that in turn also produce a ton of data. So about goodness 1215 years ago when I started doing this research. I was still able to kind of download the data to my local computer and do all of the processing there and then create a result. But, and this is exactly what I've been what what we've been doing in the past this is kind of how we analyze data in the past we, there's a whole bunch of data available. They are different files and typically in different formats that you download to your computer, then you have to get them all into a generic format and then you can do some sort of analysis and create a result. Now, one of the challenges that we're faced with these days is that the satellite data for example is becoming increasingly high in terms of resolution so here's an example on the on the left you have Sentinel to a which comes at 10, 10 meter resolution as you can see, it's, it's a little bit blurry compared to the you know point eight meter resolution from planets data, but even 10 meter resolution is a vast improvement over this data which was 25 kilometer resolution in, in, in, yeah, in space. So this is a huge improvement in our sort of monitoring system and we've even got higher resolution data available so not the problem that I've described in terms of numerical modeling data simulation and analyzing satellite data is sort of exploded in terms of the resources that you need in order to do this it's no longer possible to do this on your, on your local machine. So, this scenario, where we, where we bring the data to the compute is, is fine when you're working with sort of order megabytes of data but it becomes sort of increasingly difficult almost impossible to move to sort of order gigabytes terabytes or petabytes. So what that means is we need to think of a, we need to do things differently. So here we, we, and that's quite evident in the earth observation space at the moment where, where people are instead of downloading the data to their local they are bringing the analysis the model or the processing to where the data is living. So you basic we make use of API is we send jobs to, to the, to the data that is hosted on some sort of compute. And then we just retrieve the results which makes some life as a researcher much easier. There's a bat. Oh, and this is kind of, this is kind of what we're trying to achieve in the C scale project which is the Copernicus EOS analytics engine. And basically what we're trying to do is we're trying to federate all of the European compute and data providers into sort of a harmonized system so that that I as a user can use this platform. And I don't know whether I'm computing in Portugal and Portland. My user experience is the same I have all of the tools and data, readily available for me and I don't have to spend any time dealing with complicated setting up complicated computing infrastructures in order to do my processing. Now, this situation is exactly what we're trying to trying to build in the C scale project because, especially specifically for Copernicus and earth observation data. There is no single European processing back in that can really serve all of the data sets of interest to the scientific community. If you look at the sort of ecosystem of platforms, there is a huge variety of them, and it's really difficult as a scientist to kind of figure out which one serves my needs. They all have different approaches or different technologies that they use is them standardization that that different standards that they adopt as well in terms of the data that they serve. The challenges here are vast and navigating the system is quite overwhelming for us researchers. So, in C scale, the really the objective is to support us open science and what we plan to do is deliver a federated computer and data infrastructure, which provides a seamless user experience for analyzing, you know, computing Copernicus data. The idea is to have access to optimize data and which is both low, low level as well as the high level analysis ready data which is what the scientists typically are interested in. They want to use the analysis ready data to basically just feed into my data simulation pipeline. And where these analysis ready data aren't available we want to be able to provide on demand solution so that users can generate these data where they aren't available. The technical infrastructure is being co designed tested and piloted by the research communities and I'm one of the researchers that is testing this infrastructure to basically help the infrastructure providers build a platform that is useful for the community. So what does this really mean for the end user. So at the end of the day, we want to enable users to quickly and easily generate meaningful results so that they don't have to get bogged down and technical infrastructure details to get their data and processing and analytics to work they can just kind of arrive at C scale request resources and then they can focus on doing the science which is what they're good at and then the infrastructure providers can focus on the technologies which is what they're good at to really support the scientists focusing on science. And here what we're talking about is abstracting complexity away from the end user and providing homogenous access to resources. What this all means in the, in the sort of open with what this means is that from a data point of view it's also incredibly important that the data is openly and then easily available. And that the standards are consistent across the different platforms that you're working on. This will all help a researcher really focus on the science and I have to always convert things into different formats etc etc to to get to results the amount of machine learning is a very nice example most machine learning experts become data wranglers where they spend a vast amount of their time just making sure that the data is in the right format so that they can put it into the machine learning pipeline. That's essentially that what we were trying to overcome for them. So, here are the key project results from our from our project so we aim to deliver a a computing and data platform which we call federal data, which gives you uniform access to a federation of computing and data providers that are all harmonized and follow consistent standards in terms of how they operate. There's a processing platform that allows you to send jobs to the providers instead of instead of having to download the data locally. There's a metadata query service that we're building which allows people to find the data that is also an incredibly important aspect. Sometimes it's very difficult to find the data that we're we're interested in. There's a lot of projects are building catalogs. And, you know, if you look at the European landscape now there's almost these catalogs of data sort of scattered as outcomes of projects all over the environment. And now you also know as a researcher needs to make heads or tails of the different catalogs and what they do. So the query service is aimed to overcome that challenge. Finally, we're building workflow solutions in this project, which is essentially easy to deploy workflows that support monitoring modeling and forecasting applications so these are like building blocks to enable application developers or developers that want to you build a data processing pipeline that these are reusable building blocks that basically fast fast track researchers to go from having nothing in place to having something in place. And then finally we have a CCL community page which is where you can come and engage with us in the project. And that was that was it for me. Thank you very much for your attention. Thanks Bjorn. Only two minutes over time, I think. So that's good. And there's a technical. Yeah, it'll be, you know. Thanks very much. And we'll take questions at the end please. So if I could just move on to and please. Yes. Thank you. I think you should see my screen. So as you said, I'm on and I was working in the department of geoscience, geosciences at the University of Oslo, but I just moved three weeks ago to similar. And in my talk, I will mostly focus on why open science matters to me and how I practice open science in my work and along the way I will try to link to climate justice. So why does open science matter to me. So first, because I want to search find and access relevant resources. So I'm a climate data scientist so publish papers are not really sufficient. I write a lot of programs, I develop new algorithm and of course I analyze a lot of data and very large amount of data. So what's very useful for me is, for instance, to find a Jupiter notebook that shows how I can use data I'm interested in. So typically, for instance, Copernicus data, which are very, very large, because I don't want to start from scratch, because I want to be able to build upon something that is already available and it saves me a lot of time. So second, because reproducibility and reuse matters to me. So for instance, if I find a Jupiter notebooks, this is really great, but I want to make sure it can rerun. And I want to check the results so to make sure I start from like a sound basis from scientific point of view and technical point of view. So then, this is the most interesting part of my job. I can start to create some derivative work. I can apply the same analysis but over different data sets or over different geographical area. So for instance, here you see there is an analysis of Italy, but maybe I want to do something about Spain, but still using the same methods, maybe later I want to change the methods. And here collaboration is very, very important. So I never ever work alone and sharing what I do when I'm doing it, it saves me a lot of time because it can reduce potential issues such as, well, it works for me on my laptop but it doesn't work for others or vice versa. So that's the second thing. And the third one is because I want my work to be reduced and cited. So I don't want to retain my research work. I want others to benefit from it as soon as possible. So not when I publish, like maybe one or two years later, because it takes time. So for this I want to register my work early, for instance, like on a web portal, which is very user friendly. And of course my main objective here is to get more citation and potentially to initiate a new collaboration. So now let's focus on how do I practice open science. So I said in my previous slide I'm sharing while doing. So what I want to do is I want to register my work very early and I do it in what we call a research object portal. So it's a web application online. And I want my research work to be found by others. So it's live. So I can add things when available but I can also remove things that are no longer relevant. So at the very beginning, what do I share? I share IDs and we share IDs and we collaborate to refine them. So if you want to understand climate change and mitigate its impact, you need to work with a very interdisciplinary team so it cannot work alone. So we can use tools like Mirabord and we can discuss and improve the IDs. And then at some point we will discuss on the task and who can do what. And we can also identify potentially missing expertise. Do we need additional collaborators, for instance, to fulfill the work? And this is where we start. So when we start, we usually start from the data and we want to analyze a large amount of data, very, very big data like Copernicus data. So we want to make them available to everyone. So we use online tools and like for instance Amazon Object Storage and we can use EGI Data Hub and Adam platform. We have all tools available developed by others. But we also want to share the tools we use for analyzing the data. So not only the tools we use, where we get the data. And from this tool, we can access the data sets and we can analyze them. And we also use online tools such as the EGI Jupiter notebooks and the Galaxy Climate Science Workbench. And we can create very complex workflow and generate new research results that everyone can reuse. And finally, we want to communicate with us and so we can write papers like everyone else, but we can also write software papers, but we can also write blogs and tweets. So like different types of communication to make sure this is a bit more spread. So let me show you an example. So I'm mostly working on the atmosphere part and recently working on the impact of COVID-19 lockdown on air quality. So it's still ongoing work. And it's also a topic that can relate to climate justice to some extent. And what you see in here is my research work has made available live on this web interface, which is World Research Object Portal. So here you have like a title, which is not using any jargon or not too much. You have the author, you have the list of everyone who has collaborated on this work. It has some keywords, so it's easier for others to find out what it is about. And it has many discovered metadata. So keywords and sentences that were automatically extracted from all the research work I deposited here. So this is very helpful because my work is more likely to be found, so to be reused and maybe to be extended. And anyone can reuse it even if it is not finalized. So you can, for instance, fork it and you can get your own copy and you can start to create some derivative work on your own or with other collaborators. So here you see the different resources and I try to organize them to make it easy for others to reuse. So you can see some links to bibliographical resources and I can include videos, newspaper, not necessarily only scientific papers. In the input folder, I will put everything that is related to the data set I have used and the data set will get also plenty of information about the data where you can get them. And this is also associated to the tools you can use for analyzing the visualization of the data itself. So for instance, here if you click on this small cube, this is what you will get. It says you can open it in add-on platform. So add-on platform is an online tool which you can use to visualize data and a new window will pop up and you will see the data which is this data. So this is nitrogen dioxide from Copernicus data quality forecast and you can hear you can zoom, you can change the geographical area. So you can change the date. So you can already have an idea of what I have been using and maybe what you can reuse in different contexts. And this is the same for the tools. So here you have a tool folder and the tools, so the data analysis part of my work, you can also get it. So for instance here when you click on this link, it will say it will open the resources in the EGI notebook and the EGI notebook is fully reproducible. So you can rerun it, but you can change it so you can update it and you can make your own work. So if I want to summarize, I share my work and I try to share it at all stages and because I really want to facilitate reuse of what I do by others so they don't have to reinvent the wheel. There is one thing I haven't mentioned which I think is very important. When you share and when you communicate to increase, for instance, the collective benefit, it also comes with additional responsibilities. So when I share and I think I should say before I share, I need to think about will it harm someone if I share this data or this analysis. So that is what I want. I want my research to be fair, so to be findable, accessible, interoperable and reusable. But I also want to think about the care principles. So the care principles were initially defined for indigenous people and this is all about collective benefit, authority to control, responsibility and ethics. And the responsibility of the researcher is also to understand how to share, what to share and when to share. Let's take here an example. This is where Poucher, we are using science papers to target newly discovered species. So for instance, if you share the location of and then just species, it's potentially harmful. So it's, I guess it's like for sensitive data and medicine, you don't want to keep the data to everyone. There are some protocols to respect and this is quite important. But you can share the methods of the data analysis and you can also provide for instance, randomized and anonymized data sets will facilitate the reviews. So now when it comes to ethics and responsibility, I think we can have recommendations or policing level. But I think this is also where community of practice comes into play. So it's also very important that we as community organize ourselves. So we discuss and grow together to do better open science and to be a bit more, not only open but to care. So this is what we are trying to do to try to achieve with this environmental data science book. So it's an initiative led by Alejandro Cacastro is from the Island Touring Institute and the idea is to onboard others. So we want others to practice open science to create reproducible scalable and shareable environmental data science. So, and this is also useful for climate justice. So the idea is to define together to discuss on guidance and best practices to increase collective benefits. So we provide templates for, for Jupyter notebooks and we all try to agree together on what would be a better way to do open science in this context. So typically we recommend to have a title to provide some tags, some keywords to give the context without using jargon to list all the contributors. We also recommend the work to separate like the data access from the analysis to make it easier to reuse the same work but with different data set or different data analysis. And of course we encourage everyone to cite what what is used in the analysis and not only the paper but also the software and the tools. And we also want to encourage collaboration because if you want to mitigate climate change, you need interdisciplinary research. So we also have this collaborative review so to to ensure reproducibility so we have both a scientific and technical review. And the final work is is archives in Zenodo so we want to notify everyone that okay we will no longer work on this live Jupyter notebook or live research object. Then we publish it and those are can also continue the work and create new derivative work. So we can also create training materials so we organize co-working session and discussion around open science, but with very very concrete examples so it's very like bottom up approach. And that's it. That's the end of my presentation. You are encouraged to join and contribute and you can contact me here with my new contact details. I tried to highlight how I contribute to open science and to some extent to climate justice and I know that it's very very minor contribution when it comes to climate justice. I mean it's mostly through sharing my work and working open to empower others but I said in the title every little help so thanks a lot for your attention. Thank you very much and right on time. Well done. You know, we're all part of minor cogs and a bigger overall aim in trying to drive change so all research builds towards that. Thank you and if I can move on to Annabella please. Hello everyone. Good afternoon. I'm going to share my screen. I hope you can see it well. Yes, and if you can go full screen. Okay, okay. Okay, so I'm a coastal engineering, a researcher from the National Laboratory for Civil Engineering, and I'm going to talk with you. I'm following very well and presentation and very concrete example on how open data can contribute and help people address climate change impacts on the coast. Open Coast what I'm going to talk about today is an on demand open coastal forecast tool to support the daily management and the emergency at the coast. This is one of the thematic service from it is a very large project that bring in a lot of researchers from many disciplines together with the people that develop the core services that make computational resources as we see previously available for everyone to do their research in a very simple and user friendly way. Oh, and I forget to say I'm a part of the hydroxyl environmental department of the neck. Okay, so why, what is the motivation for those work and how does it fit in in a climate change context. We all know that our coastal regions are experienced very severe fragility and probability associated with climate change. So, our coastal managers and the community in general need to have some support in anticipating others events, and being able to make emergency actions in the in the proper way to protect the people and assets, and they also need information about the coast for the daily activities, and also for the public for they are selling the swimming all those nice uses of our coastal coasts. Finally, knowledge of the coast is also very important to guide the management to minimize climate change and other risks, and for users to be able to intervene when it's necessary. So what is our vision what is behind the development of this tool is to develop a tool that is user friendly can be used by everyone and provide coastal forecasting at the systems that the users won't so it's not a tool for a specific coastal system. It's a tool that can be applied in any place from the coast all over the world. And it is part of the effort towards developing coastal digital teams. A mechanism that is being promoted in Europe to make the capacity to simulate and test what if climate change scenarios available to everyone in a very user friendly way. So in a very short definition, what is open calls open calls is a platform that builds on demand for consistent for a coastal area of interest and it generates all the information that we need at the cost. Water levels, wave parameters, velocity, solubility, temperature, and also water quality, it is based in models similar to those that Bjork had talked about that are run in an automatic way every day, and they are based on the representation of all the relevant physical and chemical processes. So why why is is open calls different from what the other forecast systems that have been around for a decade or two. Typically forecast systems are tools that are very complex to set up and to maintain an operation in a daily way. And this would make, for instance, poor regions of the planet, not being able to have to take advantage of these tools for their daily activities. So the golden open cast is open calls is to make these tools available to everyone in a very simple way. So you don't need to be an American model or a coastal engineer in order to be able to use open calls and to set up a forecast system for your system of interest. So by doing this, we are simplifying and hiding all the complexity of the systems. And through the use of something that is very important here. We shared the European computational resources that are necessary to have a robust tool, and a tool that runs every day without any intervention from the user to user just needs to go inside the tool when he or she wants to look at the results and predictions for the day. It's basically everything very simple to use. From a more research point of view, we also made these two very user friendly for those people that want to build on top of it. So it's very flexible on the four things that you want to use and I will talk with you more about this. The process is physical biogeochemical that you want to use, and those are the medical models that are behind it so giving the user all the freedom that he or she wants to produce their forecast. Finally, we will also have a very flexible architecture built on top of the European Open Science Fund that allows the system to be implemented not only where we have it in the cloud, but also in a proprietary way in one specific institution where there are special issues maybe at that time. So just a very brief overview what we have in this tool. We have three manipulars. The first one, the user configurates the forecast that he or she wants to operate. The banner at the top shows the several steps so the user is led the step by step with a lot of help in the procedure if it's needed. And I'd like to talk about if it's like going to the supermarket. So you pick up everything that you need from each shelf and in the end you just have your full card. And the nice thing about this is that this is open science. It's open software, open access and open software tools that everyone can interact, use and also build on top of it. So we want to contribute towards coastal forecasting as everyday tool for everyone. So here the forecast manager, this is particularly interesting for researchers where they want to experiment numerical parameters for instance, and they can have multiple forecasts at the same time. And it's very easy to do using this manager. Finally, this is not useful at all if you cannot look at the results, download the results, download the input files. You cannot put fewer where the user can look at the maps probe over the results in a very straightforward and easy way. So why, why open coast is relevant in this context of open data. So this is not just data, it's data and numerical models. In the sense that we run the models but we provide automatic comparison with real time data, for instance, from the Sentinel images processed for for instance, the water level, or for the automatic comparison with the data from the system. So the users know that the forecast that they are producing the quality they are going to get in the end. Right now we are using these all over the world, we have users in all continents, and what is interesting, we don't have just the researchers using open coast. So many, many managers that are looking at this tool is a very important way to understand their system, and to take adequate management decisions. So why, how are using open data here, besides making all of these work available and open to everyone. All of this work relies on open data from a series of providers. And this is very important because without this open data that we use to force our models every day. Open coast won't be possible so it's very important when we build a workflow of procedures that all the information that is already there that someone has invested time and money to build can be reused by others and we are very good example of reusing the data from global and regional providers. For instance, forecasts from the atmospheric fields, from GFS, Meteo Galicia, Meteo France, the global time levels from Fez 2014, and as I said previously the amount of data stations that can be get in an automatic way and introducing to our system. All of this is done automatically so the user just has to say what he or she wants to use. One specific forecast provider, one specific comparison with field data. All of these again brands on top of the again open access computational resource that they ask European Open Science Cloud that allow us to set up random simulations and the archive the results available to everyone. So, another important aspect on open data, and you have heard it before. Another good data can be used if it doesn't follow standards, and it's not developing a way that computers can go and grab it to produce new knowledge so open coasts and the other automatic services in GIAs are all calling the standards and standards in their community and by doing this we allow a very easy way to produce new features. For instance, right now we are building the capacity for the computational grid to be built in an automatic way so the user doesn't need to have any information at all just once to say, I want to make a prediction in the North Sea, and they start here and it gets there. So, and this is possible only because it's all built in this workflow of open functionalities procedures. Finally, we would like also, and in this context of open date and open science to allow users to be able to share their applications to other users. For instance, if one user works in the North Sea, and another user is interested in the output of that application of open coasts, he or she will be able to access it if the original user allows it. So to summarize what we have here, an example of our open science and open data is really fundamental for us to build tools that can address societal challenges. So it's an open access where people can go and build their systems, it's all based in open source software, and we hope that by making, creating this tool and making it available to everyone that we can allow people over the world to understand how their postal systems behave, I'll be to evolve in the future, not only due to climate change, but also due to human interventions and their consequences. So I'm leaving here just two links for you to browse the first one is the access to the open coast service. And the other one is built beyond the open coast, it deals with all the automatic service in EGIS. And I also welcome you to go and browse through the very nice tools they have there. And this is all I have for you. Thank you so much for your attention and I'll be glad to answer any questions. Thank you very much, Annabella. Excellent. Yeah, and if there's any questions, please do post in the chat window, please. And let's move on to our final speaker. Professor Samanakis, please take it away. No, I don't know if you can give me the screen. Yes, please. Do you have your presentation? Yeah. I'm able to share your screen. Okay. Anybody can see it? Yes, thank you. Let me expand it if I can. The project that we run is the Aneansis and Novel Services for Emerging Atmosphere and Water and Space Challenges. Basically, climate is a theme that concerns the atmosphere service. But I'll carry on and speak about underwater services and some of the services of the atmosphere that do not have anything to do with climate. In the present situation, I was responsible for Section 3, the Atmospheric Research Services. They are services that anybody can use and they can get their own data or use data from databases and treat them with the research services that are available in the Atmospheric Research Services. And Service 1 in our case is the greenhouse gases flux density monitoring service and implementation. At the front end of the service is a simple user friendly interface that accepts the data and guides the user to obtain the desired flux densities and energy balance results. It is possible for the user to upload data from his own database or from any other database. I have to say here that calculating gas greenhouse gases flux density or in this case water. Water fluxes or any kind of fluxes, mass fluxes is not a trivial case. It's not trivial science, it's not trivial thing and it's very complicated. And of course I'm taking a step back in what has been said so far is how to obtain this data. I had discussions mainly not only on how to get this big data, treat them and share them. In this case we go a step back is how to obtain this data and then treat this data and then make it available to open access. In this case, in our case, what we used in the years we used many on the left, you can see the projects were obtained data. And in the middle you can see the method we used the actual physical method that we used to treat the data. And then you can see where we deposited the data and how big the data were. I'll have to stress again that the way to obtain this data is very expensive instrumentation is very expensive. It needs well experienced scientists or even PhD students won't do the job well. And we use two methods to obtain flux densities. In our case for greenhouse gases, the eddy covariance micrometer loads for method and the gradient method for heights dynamic gradient method. We deposited the data also at the European flux database cluster in Italy, and there is an American flux database. Also, one can obtain data from there or compare his the way he treats the data from the American flux database or you can use one can see a greenhouse gas fluxes from Japan stations at the bottom. So the thing here is that the way we deposited our algorithm in the onions, one very simply can go and use his data or her data and obtain fluxes of greenhouse gases. Why we still need flux of greenhouse gases is that as this well understood from the research community is that this data are missing all over the world. What we get from the IPCC on climate change are data calculated from emission from inventory emissions and data, a real data of greenhouse gases emissions are scarce or very difficult to obtain. The eddy covariance method is so well used by our group. We started using it very long ago and our first expedition was ace to over the Atlantic year 2000. So, bold then that we use the eddy covariance method from the aircraft. We obtained data on. In that case obtained data on. And determine the losses. You will see later. We also use aircraft data from after step model is using the variance method. We flew over after some three or four different heights and then fluxes of pollution in Athens. We also have a station and use in the station medical virus. But all data are available. I have to say they are treated for standard deviation deviation. And what we like to do also is give out error propagation on our data. This is an example of our station. The total 125 meter hour and here one one of the examples is that one can see that sensible heat, let it heat and net all wave length radiation. And to simplify matters is that using this heat fluxes we can see much fluxes water vapor or CO2. And between the surface of coordinates and the atmosphere, and hence we may have a little transpiration that may be useful. With a particular sector. Also, we can have a footprint of a particular products and so on and so forth. The key focus here is that these data are not very easy to obtain. And I go then to the second section of the aminesis, the second section is concept of monitoring atmospheric perturbations and components in active electronic regions. The University of Milano, because it has data, it uses the data in the sense that they want to look at and correlate gas or other emissions from faults. And so on at night and earthquake activity along the same folks and looking at the kinematic type what type of earthquake they may have. They obtain the data and they have the data, and they're trying to see if they can predict. And of course this goes for Mount Etna and Nakamene active volcano is something, something is very nice result, but opposite it is the Nakamene which is active volcano. Very small activity by still active, and we want to see generates earthquakes. I mean, our Portuguese collaborators are trying to give us an idea of the urban air quality estimation and monitoring and forecasting. This is the city of Porto, and they're trying to see what happens with air pollution in Porto, and, and then with their algorithm, trying to see which are the hotspots and from the hotspots direct the local authority to redirect the activity of cars and public transport, and they use real data also from places in Porto, and with this they built their own algorithm. And these things I had to say, but I would have liked to have spoken a lot more about our work on climate change since we worked since 1990 on climate change, but I think my time is too short. And I believe that sharing this data since the data is not a lot of data on greenhouse gas fluxes from the earth, let alone from the oceans. It's an area where data have to be said, there are so few, and of course I'm a strong believer that using greenhouse gas fluxes and anything that's concerned with climate, climate change greenhouse gas, concentrations in the atmosphere and everywhere is not correct to use emission inventories that go out there and measure the fluxes and calculate correctly what are the predictions or what are going to happen to the climate of this planet. And a great example is that we think that's a lot more. I can answer a lot of questions. Thank you very much. Thank you so much. Great. That's brilliant. And. Okay, well, we now have a Q&A section of this panel session. Feel free to ask any questions in the chat as I've posted there, but maybe to get things rolling. I have a few questions that I might want to ask and one of the things that you might have seen repeated through at least two of the talks I think was something called the EOSC or some people call it the EOSC, which actually is the European Open Science Cloud. And it's essentially a federated series of data and services and so forth that are across Europe. I see that some of the researchers, guest speakers here that they've been using this as well. And I invite anyone, perhaps one of the speakers to describe a bit more about how this is actually helped in their research. Is there someone that can actually crystallize how this has been of real use and importance to their research, please. I can start and then the arc follows me. Thank you. Oh, what I presented today is a tool that needs to run models every day. And in order to run the models in its computational resources every day. And we need to make sure that the resources are there they are working. And that we also have access to what is called core services. So what is the EOSC, is the EOSC is a federated infrastructure that congregates computational resources and data repositories from all over Europe that work in a seamless way so looking from the point of view of a researcher. I'm able to run my service on top of these infrastructure and using their, what is called core services for instance to authenticate my users. OpenCos is an application that requires a user to have an access and we need to make sure that this is a valid user, a valid researcher or a valid customer manager wanting to do good things. So we use EOSC not only for the computational power but for all the services they provide to help me make my platform work every day. I made a very practical presentation of EOSC. Brilliant. Just follow up on that and maybe someone else can jump in as well but in terms of. If you would like, I could say a few words about Neonias and EOSC. I believe the services of Neonias and EOSC are free to everybody. So, even though they may not be specialists in what they are trying to do, they can run the services of Neonias and ask questions, come back to us and explain to them what's happening with these services. But we can also get data via EOSC that otherwise would not have been available and try our own data if I can say that. So it's a dynamic exchange of data of algorithms and services for the whole community. So think of it as a dynamic place where you can get data, try your data, correct your data, correct your algorithms and discuss with others about the algorithms you are using to treat your data. So I find it very helpful not only by putting our own algorithms there for everybody to use but also for getting data and exchange ideas and make corrections to our algorithms and their algorithms. Excellent. Thank you very much. And Bjorn, do you have your hand up as well? Yeah, exactly. I just wanted to also make a comment. So I sit a little bit on the sort of interface between being like a scientific end user of EOSC but also sort of partially contributing to the development of the capabilities in the European Open Science Club. So what I really value about the European Open Science Club is this access to cross-disciplinary services or tools. For example, in the Earth observation community or the Earth science community you could argue has got relatively less experience when it comes to, for example, big data. And the high-energy physics community, for example, has been dealing with big data issues since the early 2000s when they needed to federate all of the computing infrastructures in Europe in order to deal with the big data coming out of the Large Hadron Collider, for example. They spent a good 15 to 20 years now building all of these tools that actually manage, for example, big data logistics that allow for exchange of data that allow for federated computing. And it's these core services that are really enabling us as the sort of end users to leverage them in other applications, for example, federating around Copernicus data. And there's a question in the chat from Julia, which is, we do need to spend time to learn these services, but typically in my experience we have a lot of support from the providers of the services and the infrastructures. And that's also a really nice added value of working in this environment is that you get to collaborate with the experts in things that you might not necessarily understand. Thanks. Thanks, Bjorn, for picking up on the question. I was going to bring it up. Yes, and you touched on that, that a big part of what open science is, and I hope will also really benefit climate justice is all about interdisciplinary research. One of the big things about open science nowadays in modern research, I should say, is there's no real siloed fields of research like you would have maybe, you know, a few decades ago that you have biology, physics, chemistry. Now there's a lot of interdisciplinary research which is being generated and maybe someone can also reflect on this. I don't know Bjorn if you wanted to reflect more on that and maybe some of the other speakers. If I may, again, please. At the start of the whole story of Nair Nios and Nios, we found ourselves isolated in the sense that we are atmospheric scientists, my group, at least. So you have an algorithm or many algorithms to calculate fluxes and from aircraft from stations in the earth on the ground but obviously not from ships, although I tried a lot for 80 years but we found ourselves isolated. We met IT people, information technologies, and we managed through an interdisciplinary thought and way to persuade them and they persuaded us that it's a very good thing if we gave them the algorithms. We will try and put it up on genodo and make it publicly available by interacting for the past three years, we had ups and downs, but the interdisciplinary there is very, very clear information technologies and as we managed to work together and produced algorithms that are usable by everyone, not strictly by our group that we understand, we made them available to the public. It's a, we are very excited by this working together with information technologies because I couldn't understand anything about information technology and they couldn't understand anything about the work we are doing. So, it was very exciting and it was very successful. That's why you're very pleased. Thank you. Great, that sounds like a great success story. So, and you had your hand up and there's a question for your course and in the chat as well. But maybe you also want to pick up on this. Yeah, I just wanted to mention for the interdisciplinary aspect that the way we used to work and this is exactly what was mentioned mentioned before the way we used to work is we were working alone in the lab. And we were working with close collaborators that were presented to us, especially when you are a young researcher. Like now the world is open to us and it's much easier. I found, I find to approach even like new collaborators in the open science framework. So for me, this is really very important. This is exactly what he mentioned we are really supported when we are trying services. Or when we are trying these tools. Well, in my lab, when I'm struggling, I'm alone. And when I don't find any solution. I fail alone. Nobody knows about it, but I fail alone. So that's what I wanted to mention about this. For answering the question, if you could have one interface in which compiler services you mentioned, how do you imagine that. I'm not sure this is exactly fully mature, but for instance, what they do in reliance with this research object portal is maybe closer to what I would expect like one place where when, when I have something about the research, I know what tool to use, because it's very hard for me when I read a paper to associate a paper to something that is concrete. So they say, oh yeah, you can find the data here but can you just link. Can I just click somewhere and I get the data can I just click and I can visualize the data. So this is probably what I would like. I'm not sure we can have one service that fits everyone and I like the multidisciplinary aspect of many different services but we need to find a way to associate a service to a research work. I'm not sure I answered the question properly. Sorry. No, thank you. Thank you. And, Annabella, do you have any reflections as well on this or any of the previous questions. Actually, I would like to answer with the question that was given to me. I think that what we are looking in the future, at least in the oceanographic and cultural communities is as digital twins digital representations where we can interact with the infrastructure to get answers. I think that's probably where we are going to go to be able to bring every service in one specific way in place that the users can interact with the science with the data with the model results to ask questions and get answers. So I think that's probably what we're looking at in terms of information technology working in an interdisciplinary way with the other sciences. Open Coast is a little bit of work towards that direction in the sense that what we have done could not have been done just from coastal engineers alone they would have the capacity to build the IT infrastructure and the services. And, likewise, the IT people wouldn't know about the processes and the physics and the biogeochemistry to build a service that would be useful for society. I think that most of the science that we see these days are inherently interdisciplinary. And that's the way that the science is progressing these days. And the completion of resources and the impression of the chaos are great facilitators towards that direction. Although it takes a little bit of time to be able to take full advantage of what is there, but it's very, very rich and the support from the people that work there is very, very good. That's our experience. Well, you actually have a question as well, Annabella. You spoke about a supermarket. How should this be organized for you. Supermarket is like an image of the process that the user from Open Coast is used. I think it should be organized. Open Coast is just an example. I work in many fields. I'm responsible for the IT research division at the hydraulic departments so I work in many areas related to hydraulic and environment and the supermarket vision is a way to make sure that we have a very clear workflow for users. They don't know what to do. They know what they want, but sometimes it's very difficult when you give people a new app, a new portal, something new for them, not to spend too much time trying to understand how to use it. I think this supermarket vision should be a usability question. Talking with the users to make sure that you are doing the right thing. We are a very an applied research institution. We work very close with our users and I say 50% of our work is work that the user comes and says I want to do this, help me. So I think it's very important for any research that is being done to be successful and it's not only nice to write papers. It's very nice to make sure that the people we were thinking on when we build something that they are going to use it. I measure success by that. Papers are very important also, but I think when we are able to put our research, people that work in applied research, in the end of the users and they use it, I think we did a good job. Absolutely real world changes that happen as a consequence of research. Yeah, that's absolutely. Yeah, I wanted to also quickly touch on this notion of an interface to compile all services. I think some of the ways that we're going is we're we're building all of these containers, you know, like Docker containers or a singularity containers that kind of host compiled versions of our software in which you as a user you can just pull to the environment that you're working in so you don't need to really compile things anymore, and I certainly see that there's a direction that we're going in that in that way, and just also jumping so jumping on what I was saying about digital twins. I think they're what we're thinking about is this sort of interactive modeling exactly what you presented with your open coasts idea. So being able to set up models interactively run different what what if scenarios for example put a barrier here run another simulation and do the same doing this in an interactive way I think that'll be the big challenge that we will need to in the next couple of years or decades. And what I think will be very intriguing is this idea of sort of community modeling approaches using application programming interfaces, as opposed to for example source codes. I think that's an interesting notion to pursue. Thank you for that. I, I just had a question and go back to something that I think both Professor up some and I can set and also I think anabella you touched on it and maybe others as well but that's the idea of expense, the cost of a lot of the equipment, the data, etc. How open science in terms of climate justice and equity, how this is a real driver towards that and do you have any of any of the speakers, any practical examples of this and how this might have benefited in low and middle income countries etc that yet the cost is being reduced for these countries in terms of the research that they can do embracing open science and so forth. I'm not allowed. I have to leave soon for traveling. So, but I can say that then we obtain not to obtain we use two methods for our work. And the second method or chronologically the first method. It was introduced in the late 50s, early 60s from air hardware in the game. It has become now so cheap to run using very cheap instrumentation and with the correct algorithm. In some cases, calculating fluxes of energy greenhouse gases at one and so forth using mass balance and that several countries or countries who are not able to use for things like that should be able to use. So, we had that in mind what you said it was in our mind also and even people who are in the agricultural sector could use them very, very cheaply. And so, yes, we had that in mind, and we hope that people at least in the cultural sector will accept this fact and try and use it to obtain the carbon footprint of their products. It's very, very important that trying in trying to obtain the carbon footprint of their products. They use less energy less. They use less phosphorus. They put a lot of less things in the field, they use a lot of less diesel, and they're trying to counterbalance all these things by the CO2 and draw down to their plants. So, it's another way of saving energy for the planet and they have products that they are acceptable to the public as taking care of the environment. I don't know if I'm understanding all these things. No, no, no. Perfectly, perfectly, thank you. And we actually have, okay, we're actually at the top of the earth, but let's just run with this and then end, but we have Bjorn and Anne also wanted to speak please. What you just mentioned sort of made me think a little bit about the developing world sound from South Africa originally and one of the things that that I always find interesting or something that we should probably think about is as we are moving to the cloud and all of these distributed computing infrastructures access requires good internet infrastructure. And that is something that the developing world specifically or particularly in Africa, they use cell phones predominantly. Then we have to think about how do we make those services available to those end users who do not have the infrastructure to to connect to it so that's another another thing to take into consideration specifically you're on talking about climate justice, I think. That's a great point. Yes. Internet isn't always available in same manner everywhere. Absolutely. Yeah, I think this is, I mean what you mentioned, Bjorn is also what I wanted to mention about the internet access. So we having like workflows that someone can submit and run like in the background is very important. And I also wanted to mention maybe like the people and community aspect where now most of the publication are done by Western countries and most of the researcher are usually from Western countries and they are like imposing their views of how data should be analyzed and what should be done even in lower income countries. I think open science can try to not to reverse or at least to have a more bi-directional discussion on how things should be done and how things that can be tackled in other countries. So to open the dialogue and not only come and say, you see, I've done a great research, now you take it. I think this is very important for climate justice. Excellent. Thank you. Okay. All the hands are coming up. Annabella and then maybe we will conclude, but please Annabella. Okay, so just a very quick remark. There are many users from low income countries in open cause. And usually what they ask that's not there is the ability to have this in a continuous way what is necessary to have this in a continuous way. The reason is that data either open or not open is very scarce on a local basis. They have the satellite data of course like we all have, but local monitoring networks are very scarce. Sometimes they were destroyed, or sometimes it won't have any monitoring stations at all. So they look at these type of systems supported and validated with some data as a way to know their cost of systems. So it's somehow the ability to have all these networks of knowledge and resources, global resources or open resources makes the science available to everyone. I think it's very important and that they they have very different requirements in the sense on how they use these type of tools completely different from the western countries like to answer. It's a very good experience and very challenging. Brilliant. Thank you so much and to all the speakers and we've come to the end of our time and thank you again to the speakers for your stimulating talks and for this discussion and I hope we can continue this discussion at some point in the future. Thank you to everyone and hope to see you again soon. Thank you. Bye. Thank you. Thank you very much everyone. Goodbye. Bye. Thank you.