 Okay, I think we can start. Participants are still joining, but let me first wish you a very good morning, afternoon or good evening, depending on where you are logging from. It's my pleasure to welcome you to this course session of the Glossalon Soil Spectroscopy webinar. My name is Isabel Verbeck from the FAOS Global Soil Partnership Secretariat. So today we're going to Brazil. I wish we could all go there. So our speaker of today will present the effort and successes from Brazil in building a soil spectral library. Before starting, I would like to remind you that the session is organized in a webinar format in which participants can activate their audio and camera. The meeting is recorded and recording together with presentation will be uploaded on the Glossalon web page. I encourage you to post questions to the speaker and his colleague that are here to support in the Q&A box, which would be moderated by my colleagues. In addition, you see also a chat box where participants are right now writing some very nice message, and you can use this chat box for interacting between participants. For any technical issues, please write to me on the chat. I will be very happy to help. So, before the presentation of today, I would like to first give the floor to my colleague Yi Peng who will provide you with a bit of background on the Global Soil Partnership and the Glossalon. He will be also moderating this session. Yi, you have the floor. Thank you. Thank you Isabel. Hello everyone. My name is Yi Peng and from Global Soil Partnership FAO. In the next few minutes, I'm going to briefly introduce you who are we and why we are organizing such webinar. We are Global Soil Partnership. Normally we call it TSP. TSP is established in 2012 to position soils in a global agenda through the collective actions. Our main objective is to promote sustainable soil management and improve soil governors to guarantee hairs and productive soils. All our activities are downscaled through seven regional partnerships, also supported by our partners. Our activities are also running under the guidance of the Intergovernmental Panel Technical Panel on Soils. We call ITPS. As you can see, we work with a wide range of topics. For more information of the GSP, you are very welcome to visit our website. Above all these topics, we also have different networks. For example, Global Soil Laboratory Networks. We call Glossalon. Glossalon is established in 2017 to build and strengthen the capacity of a laboratory in soil analysis and to respond to the need for harmonizing soil analytical data. In 2017, Glossalon started to work on wet chemistry, focus on training, harmonization of standard operating procedures, and the execution of inter-laboratory comparisons. Last year, we launched the Glossalon Initiative on Soil Spectroscopy, also international network on fertilizer analysis. For more information on these two initiatives, you are also very welcome to visit the Glossalon website. As I mentioned previously, our main focus of the Glossalon Initiative on Soil Spectroscopy is national capacity building. So we plan to have some webinars to train the colleagues and the lab's research institute around the world. Probably some of you already joined the previous three sessions of the webinar. The first two sessions focused on giving a basic introduction about this technology. The third session, giving a bit of future perspective of this technology, also to address some future challenges. All of these first three webinars video recording are online available. So you are more than welcome to watch online, also share this information with your colleagues and friends. So after, in the beginning of this year, I had some communication with different countries. And one of the most often asked the question was how to build a spectral library and how to use it. That is why the fourth and the fifth session will organize the webinar on the topics of the Soil Spectral Library, the experience from Brazil and France. The last webinar of this series will be talking about something about the measurement. So, the first talk about the Soil Spectral Library will be given by the Professor de Maté from Brazil, University of Sao Paulo. The reason we invited Professor de Maté to give this webinar, one main reason was Professor de Maté was one of the first Soil scientists to study research on the Soil Spectroscopy in our community. And the other reason was Professor de Maté spent the last 25 years successfully gathered all the institute stakeholders to build this joint effort of Brazilian Soil Spectral Library. As many of you probably already know, the main challenge to build a national soil database is to gather all the institute together. Because, as you know, many of the country have more very similar situation, which is the same, the different institute under the same department or under the same ministry, even the different group under the same institute that they are collecting the data, but they never share the data, never share the soil share the information that becomes the major barrier to build a national soil information system or database. So, now we are going to listen the experience from Brazil, how Professor de Maté successfully also historically brought all the institution and the soil labs, including commercial labs together to build this joint effort. So, Professor de Maté, now floor is yours. Please. You can share your screen. You have to unmute yourself. Have you seen the screen? Yes. Okay. Thank you for all. Thank you. I would like to thank FAL and the Glozlum and all the team for this invitation. It's always a great pleasure to be on this position to exchange experience. The next request is I'm going to talk about the Brazilian soil spectral laboratory and our experience on that. So, this speech, this lecture is a bit different from the other ones because I'm not going to stay so much on the statistical part. This was that other moment. I divided in some topics. The first topic is a contestualization. And I would say that I would invite everybody please to come and hear like a history. It's a step by step history on how could we reach that. So, first of all, I would like to start showing about the science, science and food production. I will introduce you one of the main important pedologists of Brazil. That is José Ruiz de Maté. He's already retired. And it's one of the most important ones in Brazil. During a lecture, he was talking about the challenge to feed the world and that in the next 50 years, we will have to produce the similar volume from the last 8,000 years for the next, the last 8,000 years. The important thing here is that, along his life, the base for productivity was to know soil types. And inside this complex, the main engine for him is mineralogy. His great merit was to put pedology for soil management and food production and not pedology itself. So now we are on the next step. Our evolution is to make pedology for pedometrics and to put spectroscopy together to make, to go to food production. So what is happening in the world? It's a lot of pressure. And I'm talking firstly about Brazil. Regarding Brazil, the pressure to go against lands is very, it's great. And we know that places that we don't know, we don't want to go to preserve such as Amazon, because we have lands to do. But look at these numbers. Brazil has about 7% of the area already agriculture, but it's available more 11%. And we are the third largest agriculture exporter in the world. And first on a lot of projects. But the numbers says that to sustain the world's needs, Brazil probably will have to produce 20% more after he has all the area used. So I can tell you how to do that without geotechnologies. We have to use this, we have to evolve. And on this matter comes soil. Soil analysis is the basic information for everything you say about soil. Now, an important point that we see on farmers, mostly, is that they are more, they look more at the first layer, the surface layer. So they ask for mostly for nutrients, for fertilizers. But soils are knowing are not only that for soils are deeper. It's complex. They have other things important to make the fertilizer act. So it's dynamics needs texture carbon mineralogy again, and death to be understood. So soils. In fact, soil function as a complex organism. And how can I imagine it if I don't know it. So this is the point I have to know the soils to manage it. Okay, so the third part is the consequence. This is the consequence when you don't know soils degradation, false diagnostic incorrect management. So why to still keep focus on fertilizer fertilizer I'm saying that it's important. But I'm saying that something is missing and going towards. Yes, we have the solution on the soil analysis that are that are the soil spectroscopy. But before that, let's look at some numbers of soil analysis. In Brazil, it's estimated about 6 million, 6 million soil samples per year soil analysis. They are still growing. The world demand is about 600 million per year. So the question is, can we keep the demand with environmental quality. Well, we have the natural resources for wet analysis. These are the questions that we have have to have to make looking at these numbers. The solution is spectroscopy. Why, because it's quick, easy, clean, and a method for soil analysis, and it has physical detection. It's strong scientific background, but it has a little issue for the persons that do not know it. It's an invisible information because you really cannot see it. You have to understand it. So you cannot see this one, but usually the wet laboratories can see what they are doing. So this is the issue. So how to explain to them that. Okay, so make it visible. Make it visible is what I call spectroscopy and basic field information. Yes, I had to make a lot of field information to show spectroscopy. Let's look at flow chart. So I started on IRL photographs and what what well this part I call the human learning and then go to the boreholes to see the soils and afterwards go to the pits and see what's happening in depth, and then come back to the relief and link all of this. This is starting pedology. Okay, but now comes the soil analysis. And when comes soil analysis I don't stop. Then we go and see by feeling what's happening with soil analysis and what and I'm seeing that the field, and if it's, if it's, if it's a quarrel. And afterwards, put everything together. And then I can start spectroscopy. And when I start spectroscopy, I can link what I see in spectroscopy with the field to see if it's current. And afterwards, of course, keep learning on the sensor, keep teaching, and you can go to remote sensing that is the never next level to link with the laboratory. That's why spectroscopy is good, you can go further. And now, after all this, I go to the field inside the pit to look at the depth and look at the roots and really and link the roots with productivity that is linked with pedology. And all of this, because of today's, we have the machine learning system that put everything together and put also soil spectroscopy. And this machine learning, of course, will come out at the end for the human decision. All this sequence gave the strength to say to talk about spectroscopy. Now, now I'm going to part B there is the Brazilian experience on spectroscopy. The first idea, and the first issue. The goal was to develop the Brazilian soil spectral laboratory to show people with the relation with soil properties. But despite the technique had strong background, why it still remains in scientific field. I kept wondering why where was the bottleneck. And I took a look at this picture here, where our group always were working like richer searchers and trying to see where was the bottleneck. Well, in our experience, we saw that everything started on soil analysis in the wet laboratory, and we kept doing the spectroscopy but separated. Now, who really is linked with the field and with the farmers and the fertilizer guys and consultants is the wet laboratory. He has the confidence and researchers keep running around. And why, why laboratories were like that. So, in my opinion, we saw that the seed of everything was the wet laboratory and why the first document of them was on 1889 and they have about 131 years. And we expect to take out this trust from everything that is on field. So we think that we had to merge the sensor with the wet and not trying to go use the sensor directly to the farmers farmers won't buy it. We need to merge efforts with people that already have a lot of experience. So that's it. We prepared a three steps, three steps to reach society. Yeah, it took 25 years. And this is it. We started in 1995. We started. The idea was we let's prepare to demonstrate this to people. It was demonstration for scientists, because still many of them did not believe. And then we will go to the teaching we will teach it for persons, but this teaching has to be first for the wet laboratories and afterwards for the scientists that still were not aware about it, and then go to society. And the third step will prepare something for users to use and practice and then final put everything together and then go to the next step that will be on the real world. So these were our ideas. Now, how to convince scientists to participate from this Brazilian soil spectral library. The first step. The first thing was that we had to show what we did since 1993 I had to show everybody what we were doing. So we are here. I was still young and collecting soil samples and thinking, thinking, how would be a laboratory. And we made at that time a spectral laboratory test and measurements, and saw that they really could function. So I took this information and started to describe to the by bulletin's papers, lectures, nails and started to spread all to everybody. And I also had to take this on speeches, but to spread the idea of the BS SL, I could not stay only one community. I had to go to soil scientists, conservation events, soil management physics classification mapping, digital soil mapping, plant and soil nutrition precision agriculture pedo metrics, remote and proximal sensing, but not only that. I had to go and show the idea for farmers extensionists, consultants, cooperatives, and public policies. So it was a lot of energy to go and keep showing to persons keep showing for a lot of years. I had still to be despite the key that was the lab laboratories I had still two other key topics. The first was the podologists. They were the most critical ones on this. So I made an action. I invited a retired awarded respected podologists to come in our group that was legal lepch, and he stood on our team for one year, only looking at, and when he said it works and it's good, he started to also help to spread. So I went to the scientific fertility community and said to them, look, I will promise a course to explain what it is. And also, the Brazilian digital soil mapping group by Maria Delores gave a lot of support. So I was getting all the key persons to help on this task. So let's put to work. The first thing make it simple. No contract, no institution participation, only the main researchers put the key persons to spread the idea around the country. And almost nobody had the equipment so they didn't knew about that, but we will make the spectral for them. And they said to them that they could send students to learn and to participate so they got happy with that. So the key was trust to maintain the data. I will not give to nobody. Low cost. And you give you gain all game. So everything was based on trust. Even I can show the BSL flow chart and its impact. Now let's start at how it would how it worked. So as you see here in Brazil, these black spots would be like the owners or the collaborators that that wanted to participate. So there they shared their soil samples, and they delivered every everybody delivered to our university that were centralized. So we prepared everything we took the spectra, we put all things together we organized and, and afterwards, we delivered only the spectra for for the user only for only his data, and he got happy because he didn't have the equipment and he could start to learn. And this was so good that he started to spread the information to others. And today Brazil has a lot. So this is the first result on our two for clay 83%. And this was the impact on community new groups were created. And since we did not disclose the data, but we, we showed who had it. We encouraged the effective participation, we forced people to go to the another one to share and to talk. So external users can get the data. He can get in contact with them, and they can exchange the data. This is the first step to get the trust, and more young students got in on board and pedologists as well. The impact on all this is that we had about 40,000 40,000 samples 41 institutions. 61 soil scientists participated from all over the 26 states of Brazil. So this was the Brazilian spectral shop we were prepared and we convinced scientists to do it, but these are were not the regulators yet. So now it's the history of teaching for society, because now I had something to show to them. Now for teaching this is the history how it initiated traditional labs have years of experience in domain of the users. So although society starts, although society starts pressure along use of residues, mostly for organic matter, which will be prohibited in Brazil, and this is for organic matter. So it's a lot of money, a lot of residue, so they are thinking how am I going to do carbon. So this was the first trigger to say we have to find a new technology so spectroscopy, it has already in place here. But despite that, we also had another trigger. Some companies inserted in the market sensors to make soil analysis, but passing over the wet soil laboratories. So they said what what is happening here. So the promise was to analyze as you see here all elements. So the community got shocked is what laboratory community with 400 laboratories in Brazil is it our end. So what I do, farmers and research was also surprised is this possible. So to normalize communication and say to them, the limitations that advantages, and how it really works, we created the Brazilian program of soil analysis via spectroscopy. The idea was not to say that the sensor is wrong, but to say that advantages and limitations and how it works. And so we created the program, but we had the rule on the first course, the first course was to show how this near happens, and only wet laboratories could participate. We had about 40 wet laboratories from 10 states of Brazil participate only them. So they didn't do nothing about spectroscopy. And the program also has a lot of steps on the second step. People started to show all sensors but now other people started to come and see like scientists, pedologists, geographers, geologists. And on the third step we put agriculture inside to put in practice and on this stage, farmers, consumers, people from marketing started to participate. And now the pro base is a community that talks since from soil analysis until image interpretation and agriculture. So this course to put them to make them come was not easy. Yes, I had despite the trigger, I had to make everything by phone calls to I made more than 800 phone calls to try to tell them to come. And as I told, we had 40, 40 laboratories and about six, about 60 people, but 40 laboratories from eight, eight, eight, eight states was already good. This is how it happens this course. The laboratories, we gave them a protocol and they send to us their soil samples to our lab now it's different from the way it's from the spectral library. The labs send to us, and we measured it, and we prepared we see we see we prepared by by laboratory, we measured with a lot of types of equipments, and we made the predictions and processing, all of this was made by our group. Okay. So, five months afterwards, because we have to made about 7000 measurements. The laboratories came to our university presidential and we gave the first course, we stole them how it functions. And we presented the data that we got with their soils. Okay. And what we presented to him, everybody had the results, but nobody had new who was the laboratory they received a number. So number one, he knew that was him, but he didn't know who was number two. And he could see, and this example is for clay. He could see the R2 that that he had for clay, compared with the other laboratories. We also made a general model, and the local model for each lab. So I'm only going to present for clay only for demonstration. For clay, we had for the model variances here and the validation also some variances. These are we're all pretty minor ideas and we didn't make so much processing yet. But we also measure the same samples and other areas. And we try, I'm going to explain this, we tried to show how the port here, let's focus here. The percentage of laboratories that reached more that 0.8 R2 for clay on training. 82% of the labs reached 82 over 0.8 82% or organic matter was lower. But on the validation, it drops a little bit. So this is only to have an idea. We started to play with the data to see what was happening between the labs. And the conclusion was that local models were significantly better than general models. And the data were presented and the limitations and the advantages discussed. Okay. These are the positions of the participants along Brazil. And we are going now on the next one this year we already have people from other countries. And they score this course we already gave in Israel and probably will going out to Russia. Okay. Let's continue the teaching the impact on the on the people that saw the course. Look at what happened. The communities first impact on the exposure of the sensor equipment launched on the national chain as I said, when they got shocked that would we would replace laboratory analysis was now normalized. They got more comfortable. They do not a sensor is not magic. After the course, there was a complete this mystification of magic spectroscopy, and we are able to concentrate on real and documented data. They understood that spectroscopy is laboratory dependent. Now they, but they know that it's brought the spectroscopy is good, and it functions. But they only they only want to see the limitations and how to do it. So now they are questioning all around how to make it because they are seeking for this for it. They are questioning everybody. What's happened now how can I do this laboratory. So we are on the questioning part, and more, the regional soil quality regulators on wet soil laboratories are inserting now on in our their discipline spectroscopy. Okay, so we passed through the teaching, and now we go to the online service Brazilian experience, because now I can, I can show them, I can ask them to, to start to use it. So I'm going to show how we made this online it's a platform. You can see the platform on the chat. So we have the background. So the user will not see the background because it's not his interest on this phase. So here, all samples that came and have spectra will be inserted. But we also got people that sent a on a same a whole on a same fit. They send the soil classification. So we have soil attributes and soil classification. We can have more information. And on the front end, we will have the user that can choose on filtering the what he wants to see, and what he wants to do, and I'm going to show right now. So this is the structure. Okay, so how does it function. The online Brazilian soil spectral library. So here when you go to the web, you see services, and you can choose three types of services. The first service, and I think it's important, you can locate the owners and who has the data. So you can get in contact with him. You can see it here on the website, and you can get the map of the collaborators, and you can see his name, email, and you can contact and say, look, let, let exchange ideas, let's, let's work together. And at this time, you can get the data. So the data are not completely closed. They are half opened. The next step on one or two years. This will be completely disclosed, but not why not now, because first you need to practice you need to learn. Okay, now the laboratory has this near and MIR you can choose, and you can choose to see a soil classification how a fellow so you put fellow so let me see the spectra of a fellow so and it will show all deaths from a fellow so. Let me give you an example, or you can say I want to see a sandy texture. So, a spectra on mid IR, and you can see here, the spectra. It's automatically it's short on the screen. You can come here and choose the surface that the service that you want to make soil analysis, and you can click and say everything you want, we put granulometry chemistry and remember my neurology. So you request, you, of course, you will have to have a spectra, you will download the spectra, you will doubt the user download the spectra upload the spectra, the spectra will go into the system online. You will receive in your email a sheet that is the soil analysis report. This is already functioning. You can do it right now. You can send a spectra and ask for the results, and then you can compare with your real results and starts to see how our model is functioning. And more than that, we also prepared and we show the statistics to show the differences and accuracies between each element. Okay, so the summary and you know about that. If you cannot measure it, you cannot manage it. Okay. All right. So we're going forward. Now, what this teaching and what this thing it calls in the market in the chain of markets. Look at this now. This is happening on the same figure that I already showed what I call the soil sensing cycle, because the data sensing is all going around is that the wet laboratory that was separated from the sensors. Now they are together. Now, they are starting to come together. So what laboratories plus sensors now are a team. And I can tell you and this is my my feeling. If we use the really strong basic of chemistry of wet laboratory with the chemometrics from spectroscopy, we will have a greater result. This is the team and remember about the chain. Okay, but now look at what happens of the chain. The researcher when he goes around. He now it's easier for him to say that sensors functions because he have the laboratories trust the fertilizer guy also it's the same thing. And the farmer is the same thing. So we normalize everything at least at a private level, and also the sellers of the sensors. It will be easier for him to come and talk with everybody, because everybody now is understanding spectroscopy. So everybody have everything came from a seat. So important events from BSSL in 1993, and then learning and here in Brazil until 15 the first workshop on soil spectroscopy, and then until 18 the first symposium in the world Congress, the first remote sense is supposed to in the world Congress, then afterwards the first paper of the Brazilian, and the first program of teaching. And now the first service. Okay, so now I'm going to make it make some finally consideration regarding the technology and Brazilian experience. Let's talk about all this technology now. The near future on soil analysis, despite a lot of ideas that are coming on. I would say that the first is the hybrid laboratory, because, as I told it starts on the lab. The hybrid laboratory that started to show in this paper here. The definition is that wet soil analysis will work together with spectroscopy to reach the best quality, low cost clean and quick analysis, going from laboratory. And you can go to field as well. Okay, so there is and we saw in this papers and by experience that there exists larger variation between what laboratories when they measure the same element, because we know everything that happens. But sensors varies very lower, very lower difference between sensors, so we have to put these two together. More, the best model of spectroscopy was dependent when we got the best wet soil analysis. So it's the same thing. Okay. Well, let's continue on the way with soil analysis I'm going to show a short film about this and then I will continue how I see this is on practice, how you can construct an hybrid laboratory and this is only one idea. Okay. So what you see on this part is also all the same that is already done collect and ship. But when you receive and prepare the samples, you do not go to soil analysis, you go to spectral analysis. You go to the spectral sector inside the lab, then we'll make all preparation for all this time there are protocols that has to be used and after I will see on his speech. And now the samples go to spectroscopy and you have a lot of types of equipments and accessories we only giving some examples. So using the fiber, the fiber is more difficult to use not more difficult but it does not optimize so much the word, but it depends on what you have in your hands. So you also have the contact flow. So we are giving only two examples please. You have a lot of them in the market. Okay, and also have other mission this all can be more prepared for laboratories to make it quicker, make it quick. This was the first part, measuring the spectra. Okay. So now, let's go to the next part phase two. Let's continue you're in the lab phase two. So, as I told you, you have all soil samples that your clients brought, and you took the spectra from each one. Now look at this. Afterwards, you will have the leather sector, the chemo metrics one. The chemo metrics will select the main represented soil sample, but based on spectra, not soil analysis. And afterwards, the selected ones will go to traditional analysis, as everybody know. So here, only some samples went and who choose were from the spectra. Okay. So now what happens. These are these samples this on this side, what the clay content came from the wet soil analysis. But now it will go back inside and they have a respected spectra. Okay, so now we have the spectra from these samples. And we have the element, how much the sample had from the wet soil analysis. So what we have now is a model. In this model, all the spectra that from the other samples will come inside the model, and will spit out and will come out the clay content of all the other soil analysis. This is how one idea on how to play it, and will reach the farmer. Another sequence that I wanted to show is, and I call this, this would be the first part for you that are starting. You have to start to make the data. This is an advanced part. Imagine that you already have a lot of clients, and all clients send it to you and now you have a lot of spectra. So you have a laboratory database. Now it's a bit different. Now you can use another strategy. The client comes to this reception and brings one soil sample. So now you send the soil sample to the spectra sector. And the chemometrics will see your data set, and we'll see all the groups that are similar, and we'll try to see on, and, and he will try to see on which group, this unknown spectra fits. And if it fits, you will use the model, you will not need to make the wet analysis, but if it will not spit it fits, then you will have to use the soil analysis, and you start growing your data sets, and you deliver for your client. Okay. Now, some things that are important that people do not see from spectra spectroscopy for web laboratory quality evaluation like qualitative view. This is an empirical thing that we just saw. I was looking at this spectra here this is spectroscopy. When I saw the clay content, it said to me that it was very clay. So I put the, I put two spectra of a clay and a sand. And this was not, not similar for the clay. Absolutely not. So what did I made? I got the soil sample, same soil sample, remade, remade on wet analysis, and came the new result. It really was a sandy sample. So the spectra was right. Now, this could be what was only looking at, but just now we are creating, and this is still not published, a systematic spectroscopy quality evaluation where using the spectra, he creates clusterings. And by the clusterings, he sees on each clusters, he sees which sample is with a spectra to too much. It's like an outlier. He's too far from the pattern that he should have. So this we are doing as a thing to make more fast. You can run all your analysis and see, oh, this spectra is showing a soil analysis that is not correct. Let's remake it, but by spectra. Okay. Now let's talk about money because everything, everything goes about money. So spectroscopy to see your costs. We got all the data of these laboratories that are great contributors with us. And we got all this spectroscopy and look at what we've made. We made for each laboratory a training and a testing model using spectra and the soil analysis. But we use some numbers for training and testing and we were, we were increasing the testing, and we were diminishing the training to see what happened. And look what happened on this graphic here. Here is the laboratories. And here is the R2. And this is the size of the test. I'm going to try to explain on few words. This lab laboratory, number eight, what we saw is that when he, he can say 13% money if he use 0.85 of his accuracy. But he can say 50% of the money if he wants to use 95%, 95%. So the point is that if he increase the, he can increase the accuracy and, and goes lower. I'm sorry, but I think here is different. He can save 13 with a 0.95. Yes, 0.95. It's wrong here. And he can save 15 with 0.85, of course. So he can save more money, but with less accuracy. Look at this, all these laboratories here, they are all under 0.75. So please make this correction here for, thank you. So your analysis and productivity. So what are we missing? This is another point that few people talk about. When you go and farmers go and to use fertilizers and put fertilizers, I asked you, do you remember that on at what's happening inside the soil, it's not only the nutrients and the CC is based on the number of nutrients and the A, L and H. But in fact, we have a mineral inside here. And mineralogy is mineralogy is the engine of water, pH, fertilizer, roots, and sorts of minerals, microorganisms, soil aggregation, temperature, soil truck, it's a lot. It's too much. So in practice, users are with missing concepts from production. So mineralogy was forgotten because it's a very difficult thing to do in the past. But now we can do a spectroscopy. Look at this graphic here in Brazil. For grains, for soy, this is the production around the ears. Okay. And this is the use of fertilizers along the ears. Why is why is this happening? This is happening despite a lot of variables, because we still need more information about what's happening and the roots and then our soils, before putting all of this in practice and mineralogy can reach this. And finally, we come to the machine era. We are in the machine era. The machine era in its limitations. For one measurement, you can get a lot of information. And these measurements will go to the machine, the machine learning, and we will have a lot of elements. So we have to put in mind that these elements, main of them, most of them are reliable. For instance, now some of them are mostly, and some of them are still on research. We have to keep them in mind. But the machine learning is also population and models dependent. And in some case we cannot explain because it gets from inferences. So yes, we cannot see what's happening. Everything what's happening. But the laboratories can see what's happening on the equations. That's why it's difficult for them to see this. This is the evolution. We are in this area, and we have this challenge to go on with. So now I'm going to show something more to go to practice to people on spectroscopy. Okay, so spectroscopy can go also from the laboratory to remote and to the field. So let's take a look at this short film, please. So this, these are the data set that we have for this near and near from Brazil. And we created this real system to have the bear soil image. So we can estimate also the bears, so analysis from bear soil. This is field, but we can get a zoom from a farm. And this is not, this is physics to get played on content. It's not statistics. It's a detection of information of spectra. And this of course is to help to optimize. So we need soil types on this task. And afterwards going to the management zones and precision agriculture to use the important fertilizer and the variety to reach product. Okay, so what can be done users do not want to understand nor what make modeling. They want to make a platform where the user choose the population and the system makes them all the way. Advanced level, allow to do its own processing. This is more for use for researchers spectral laboratories can be made for the world continental or regional. Level chosen by the user's choice regarding accuracy scale and others. That's why it can be done. The type to use will be country soccer depth and as well and teaching courses at different levels from different countries is absolutely necessary to tell the advantages and limitations. The advantages, we all said, quick preparation can bring a lot of mineralogy back and will increase popularization of soil analysis. The limitation. Well, we can discuss some limitations as which elements are the best. And we have a lot of challenges yet. I don't know about models and spectral laboratory libraries, but the importance is that we can transfer this to field as well. This is the next step and all, despite his old also going at the same time. Finally, I would like I'm at the end. I would show a brief film as I was in this pandemic terrible moment. I was in the my nephew home and see a girl having a lecture and online. And this is the film of what I saw on her lecture about soils. The new generation, what they want. Okay, that's it. And finally, I will show a brief film of what we do in our team. And our team is closely dependent of a lot of persons around the world that help us. A lot on Australia on Israel on Europe on us. So we have help with a lot of people to make all of this. So I'm going to show what we all do to know soils from the basic to spectroscopy and to go to the farmer as a service to make food production. Only to say that field work for us is not only in one farm or two farms. We go along a lot of places in Brazil to see all the variances that we have and keep testing models on all places, testing and retesting on different environments and different biomas. The other thing that is important to state is this, this technology as pedo matrix. It looks the dollar G alive again. That's what that's my what I see. So students start with this but then probably can go and wants to know better about soil and go to the dollar. So it's a way to bring young students back again. Okay, so that's it. I thank you everybody. For this opportunity for following those alone and all of this that is on screen, and mostly for the young fellows that help me to prepare all this presentation. Thank you all. Thank you. Thank you very much, Professor the market. I think you did very, very good question presentation and I saw many participants. In the chat, they all say it's very informative presentation and they give a lot of learned a lot from your experience. And we also had quite a lot of questions. And we select a few of them can answer a live, maybe you can. There is a I think the front of Q a box you also can say can see it by yourself, but I will read it out because the recordings will not show the Q a box so for the later on if anybody want to watch the radio with recording so they will also know the questions. I think there is the first question is saying soil sampling is done by many laboratories and the person. What the error due to sampling from multi laboratory database, and it's addressed within the model, or yes, I think it's mainly talking about that is a random error from different laboratories or by the people. Yeah, I understand that that people collect differently. So it will be when this sample reaches the laboratory, it will have to be written how it was collected because if it's randomly and mixed. It's it this this can be something that the result will be for the for the random sample, not the single one. I agree as well. And there is another question actually quite interesting. I, I was trying to answer and I realized actually is a quite a good question, maybe you can help and then many other participants can can understand better. Okay, what is the best or what is recommended the sampling approach for the spectroscopy analysis for carbon probably also can something more general, or so he's talking about the, what is a recommended the sampling approach for constructing a show spectral library, should it be sampling by depth, or by the horizon. Well, this is a really good question. Well, let's say like this, let's say like this, we cannot make. If you have if you are research and you have kids, it's better by horizon, but but not all the horizon, you can split inside the basic right. But the point is, not everybody is going to open up it. Everybody is going to collect using using the other types of systems. So I would say that yes, you have to collect by horizon by surface, collect on surface. But on different depths. And, well, if we go directly to the question the correct is an horizon. Yes. The problem will be the technique part how to do that because nobody's going to pick all pits, because when you're doing with with auger, you're mixing horizons that will mix the information. The carbon is very important. So I would say that has to be by horizons. Otherwise, you have to begin to get only samples on the first layer. Now, another thing is on practice on practice for farmers, at least in Brazil. They do it 020 centimeters and use it because they don't have it will be too much money to keep collecting all 0505 1010 20 to make it so you have to differentiate scientific to practical reasons. Yeah. I think this question, we can discuss weeks and the weeks. Yes, many books. Yes, that's it. And there is another question. Also, it's general but it's also interesting to discuss is what are the limitations of the soil spectroscopy compared to the what chemistry analysis. Okay, this is the basic, the classic question. Point number one, you will not lose nothing using spectroscopy. You will only gain. This has to be inside. You will only gain using spectroscopy. The spectroscopy will bring a lot of advantages, like to see such as to see the quality to see spectra to see me in a lot very to increase information on your soil. This is the first thing. Second, carbon will be carbon credit for the world. Everybody needs carbon in Brazil carbon will want. To use more water residues. So it's another thing. So when you say what's the limitations I can tell you some limitations but you have much more advantages. Okay, so what are the limitations, the limitations are one. Another chemical ones like calcium, magnesium, phosphorus and nitrogen, the chemical elements are not directly detected by it. So you are model dependent by the machine learning systems that people are creating chemometrics. So, for instance, now today, it's a limitation, but chemometrics in machine learning will evolve. So what, what are the limitations that I can, I only see much more advantages, much more advantages than limitations. I can say that I see more limitations on the wet soil analysis, when you have where you have to use a lot of types of chemistry elements that comes from different companies, and you use different equipments, and you have a lot of stages. And that's why we have difference between laboratories. So think about that. So it's the opposite. What I see, you don't, you're focusing, it's a wrong focus when you say what's the limitations related with the wet. It's the contrary. It's what the limitations of wet regarding spectroscopy. One point, spectroscopy is not magic, and it will not do everything. Today, if you start, you will do clay, sand, silt, organic matter and CC, and if you want mineralogy, you will gain. And you will also gain regarding the qualitative part that I said, you'll gain, we'll gain two and wait a while about the nutrients and continue using the nutrients. For instance. I saw a few questions in the chat, some colleagues wrote some questions in the chat and this question is also particularly interesting. I mean I can answer it. But I think it's also too important to listen your opinion, not just this question this question is about the, is there a regulation for soil spectroscopy. In my mind, there is not a regulation, but I think it's interesting to discuss a little bit. Do we need such a thing actually. Yes, very good, very good question, because just now, and Dr. IO Bender will say that there is a group working in the war for one, create protocols and rules to make spectroscopy. And two, to get an ISO reference for user for the spectroscopy for soils. So, so the sensor will have to you have this ISO for soil analysis, and this is ongoing and probably will win on next year. So these things that you're asking for are already being prepared. So the idea is that at the end, you will have all the protocols and you have to think how we wet laboratory. And you will have the method, the rules and ISO, the regulator regulations, all said, and the types of sensors that you have to use, and all these stuff. Thank you. And there is another question is also quite interesting just posted. I think I remember this question was asked in in a chat. He said, how may, because this question is, is also often asked, I think asked a couple of times in the previous session of the webinar ask them. How many samples is necessary to build an acceptable soil spectroscopy model, where the model built for one country is applicable to other countries like a Brazilian model for tropical countries like for Philippine. I need a little time to respond that because I know how to respond. The point you have to think is spectral libraries is you can as you can, we will have a world spectral item. We will have it. And we will already have the Brazilian. So, if you have a sample in Philippines and want to use my mobile you can use it, and you will see the error. If you can you will use the world you can use it, and you can see the error. But at the end, at the end, these services will be chosen by the user. So, the idea is that when we have all this prepared the user can choose the samples that are around his area. And the samples that are around his area in the data set will use this model for you, because the best models are local models. And if you send samples from Philippines to use the Brazilian, well what we are seeing that on some cases got good and some cases not. So, I'm not sure about it, but if you have your local model, I'll be sure that you will have greater result. So, I have to make this point for the wet laboratory guys. If you have a wet laboratory, as I'm telling you in Brazil, start like this. You have the wet, prepare your spectral sensor, and all clients that come start to make the spectra and the soil analysis, and know from where he is. So, when he comes back with the new soil sample, you already have the spectral location from his farm. You know, so that's it. You have to think slow and then go greater. That's why the Brazilian laboratories are one to start because they will have the control. And it comes with 10 samples, but it's not good for him yet, then 10 more. So when I have a number of samples that it's good for him. Every time he comes, I use spectra. So the user will be able to use according to the accuracy that you want. For example, if you are in the country that does not have no laboratory analysis. Well, use the Brazilian model that was online. It will be something because we will send you an accuracy use the world one that it's coming to come here coming in. This is also dependent of the structure of the country and the structure of the farmer if he had the money that he has to make it or not. So it's all user choice. I hope I could give some ideas. Thanks. Yeah, well, the participants actually now are quite active and many questions coming when one question actually I also would like to answer a little bit on behalf of the gloss alone point of view. And also, Professor the mother also, you are welcome to respond a little bit because this is also specifically asked for the Brazilian library. There's a one colleague from China, ask them from Beijing China, ask it. Can we get the data of the Brazilian so special library or other spectral library to build a water model. So first of all, I would like to respond a little bit is to this colleague and also early other colleague is there are quite many free open data resources spectral library in this world of such as a Lucas data which is a European. And also spectral library is free available for everyone to download and free of the using and currently gloss alone is working with different our different partners, and we are trying to build a global with existing data, and also trying to provide different partners to share some of the information to build a better and bigger spectral library global spectral library, and we aim to provide a free online estimation service as well. That's our objective, but that will be a long journey to go that's and also in the meantime closer long is and ends to connect the laboratories around the world to stress in the communication between the laboratories. The answer is yes, we can try to courage and convince the laboratories to share some data, but it's not something we can do it just tomorrow but if you are any of you are interested, please write email to me, and I can try to help. And then I give the floor to the professor to talk about the Brazilian data. Yeah, the Brazilian data was created at first by trust as I said, it could not be disclosed they did not want it yet because they were working on that. They were making things on that, but we created a system now that is online. You can see all the persons that participated. And you can see the their email, you can contact directly with him. You can present yourself, and you can ask for the data and, and probably, and you know, this is good because you will get in contact with persons, because most of them are scientists, and you can make works together. Or they can always share with you. Yes, this is this is the idea of the Brazilian to encourage to create new. If we only disclose it. People go gets it, and that's it. But if we force them to go with the researchers, they will create more interrelationships all around the world. So this is the idea, you can get it, but you have to contact them, but we will, we will. This is the next step, it will be shared all of it opened after everybody has a time to use the online platform. Otherwise, the online platform loses its its its its idea. Okay, but it will be. Thank you. And another thing I would like to address is I first of all I apologize. In the beginning of this webinar I forgot to introduce our another two panelists to help you with the questions and they need and sorry. Could you please turn on your camera so our colleagues also can see both of you and thank you very much for this one and a half an hour, helping with the questions I think we have more than 50 questions, more than I think probably even 60 questions just a plus some questions from the chat. And then we, we really appreciate your help and support. And some of the participants also asked that if there will be a certificate certificate provided. Yes, definitely. We have a closed long webinar and I feel webinar we normally provided the certificate for upon attendance. It's just that we have a large amount of participants join the webinar, and every time we have a few hundred participants that and we have a different webinars so our team, our colleagues are working for this matter and very, very hard. We have some time to send you the certificate by the email but sure, you will receive the certificate by the email. I think in, we are in the end. This brings us to the end of the this session, you will still hear from us, and the receive shortly an email with a link to the recordings and presentations. There is three webinars already online so you are very welcome to watch online, and this one will be the recording will be online presentation will be online as well shortly in a week maximum two weeks I think you will receive the email as well. A very big thank you to our today presented a professor demate and the hundreds of the participants, I think we reached the approximate 400 participants, as I remember, joined our webinar, and we, I will now to, I will now populate the send the link to the our chat so you will be able to register our coming webinars to continue listen our experience and the knowledge is from our experts. Remember to check this wealth page regularly as another series of webinar on what chemistry, health and safety, equipment purchasing quality assurance and many other topics related to the soil laboratory work will be organized in next few months. So all the participants. Thank you very much for you for joining this webinar, and also thank you very much, Professor the matter and also our panelists, the two experts from the team of the professor from Brazil. And thank you very much for your contribution we really learned a lot and see you next time. Hey, thank you all. Thank you for the support. Thank you. Thank you. It was very good. Many questions. Okay. Okay. Yeah. It was very, it was a lot of work to prepare this presentation because it was, it was not prepared, it was not only get some papers and show them. So all presentations that we make are new are different. So I hope you liked it. Okay. Yeah, it was very, very good. So that's it, Isabelle. Thank you. Thank you. Thank you. Okay. Okay, thank you. Anything please get in contact. Okay. Sure, I will write to you shortly once the video everything is online and I will let you know and then we can talk something further collaboration. Okay, that's it. Bye. Thank you. Bye.