ChangeMAKErs - Dr. Nasem Badreldin

Welcome to ChangeMAKErs, a new MAKEManitoba podcast series highlighting research and innovation powered by members of the Faculty of Agricultural and Food Sciences.

In this episode, host Peter Frohlich, research development coordinator with the National Centre for Livestock and the Environment (NCLE) talks with Dr. Nasem Badreldin, Assistant Professor, Department of Soil Science.

Learn about his journey into digital agriculture, his research on soil health, crop yield prediction, and grassland resilience, and how advanced sensing technologies and AI are shaping the future of sustainable agroecosystems.

Transcript

Intro: 0:00

The way we grow and produce food is ever-changing, shaped by consumers and the climate in which we live and farm. Research at all points of our food system is essential for continuously improving food's journey from farm to table. The Manitoba Agriculture and Food Knowledge Exchange explores timely research innovations and applications that make our food system better than ever. Join us for today's podcast.

Peter Frohlich: 0:26

Hello and welcome to ChangMAKErs, a Manitoba agriculture and food knowledge exchange podcast series. My name is Peter Frohlich. I am the Research Development Coordinator for the National Center for Livestock and the Environment at the University of Manitoba. In each episode, we chat with an academic member of the Faculty of Agriculture and Food Science right here at the University of Manitoba to find out about the research they're working on and how this research is shaping agriculture and food production in Manitoba and beyond. We also get to know the researchers and get to know Thank you for joining us, Nasem. Nasem, before we get into the details about your research, can you please let us know why you decided to become a researcher and a professor and how you got interested in your field of research?

Nasem Badreldin: 1:18

Well, again, thank you very much for your invitation. Well, start back when I was a little kid. I was born in Morocco, and my grandfather from my mom's side are farmers. Most of the summer in Morocco, which is between, let's say, June to September, we spend time in the farm as part of meeting the family. And during that, I remember joining my grandfather, who is a very traditional Moroccan farmer. He's not using technology. He's not using sensors. He's not using any type of heavy machines. He's using the livestock and traditional way of agriculture. And I was joining him in everyday activity. And I found that his interaction with the farm and even predicting when the rain is coming or what type of crop he needs to cultivate or what is the expected yield without using technology, such fascinating skill. I was always asking him, how do you know that? And his answer is, he always gave me that, you connect with your nature. You read all the signs and you become more efficient in your farming. During that, I start to become more curious. See the bugs around, see the livestock around, see the interaction or using the livestock manure to fertilize the soil and then the soil become much more healthier or become degradated or something. Things accumulate in my mind that this is the field I would love to spend my life working with because it's fun.

Peter Frohlich: 2:53

So you got interested in the farm and then the work on the farm. And how did that lead you to study agriculture?

Nasem Badreldin: 2:59

There's another twist to the story. Once I, of course, seen these amazing things year by year, becoming a teenager, it become for me, you know, I start to realize, oh, this is a lot of work. It's not just about having fun anymore, you know, as a kid. It's now it's logistics and you have to manage seeds, you have to manage manage tillage, you have to manage irrigation. It started to become a more complex system. Then I decided to go for a law school and I went to the law school for half a semester and then I dropped during that time and then I joined a business school. Then I finished and said, okay, I'm going to give myself time in the business school and see. Because most of the time you're sitting in an office, you know, I didn't feel that I can be a good student in a business school. Then I started to realize, no, I belong to life science and one of the is agriculture.

Peter Frohlich: 3:49

So Nasem, it's wonderful to hear that you've developed a passion for farming in Morocco. Can you tell us some differences between farming in Morocco versus farming in Canada?

Nasem Badreldin: 3:58

Okay, thank you. That's a very good question. Well, let me bring you to Morocco first. In the 80s and 90s, the main cropping area was wheat and grain. Usually they are rain fed. They are not irrigated. It depends on the climate. And they always switch between barley and wheat, depending on the prediction of the drought. Brings me back to my grandpa when he decided before, you know, seeding that this is going to be a barley, not wheat, because he's expecting that there is a huge drought. Using the livestock for tillage, using the livestock for, I mean, the and everything for fertilizing and so on. Also, the farming sizes are a bit smaller. Household could be half acre in some households. The purpose of cultivation as well is different. It's not most of the time industrial for selling a massive amount of grains. Most of the time it's local markets or household consumption during the year. And adopting and using technology between the 80s and 90s and the early 2000s in Morocco become a little bit expensive. I remember my grandfather father's house has no electricity, at the time even charging those little small mobiles remember those old Nokia ones used to use they used to have the car batteries to charge them, but nowadays I see that they are adopting more western technology for massive agriculture production they start to realize definitely that there is a food gap they are taking a leadership role. I think in helping Africa's food shortage Canada is much more aligned with the development in the technology the farmers are well aware of the newest tractors and newest equipment combiners and what they can do and what they can afford. And there are organizations, they can help give some loans and support farmers. The other sides in Morocco, it's much more independent financial management and support. Some of the time, if they have a huge crop yield loss, there is no compensation. There is nobody is going to insure their crops. But now what I've seen in the past five years, the agriculture industry in Morocco is booming. They are investing a lot and they are pushing forward to have much more Western agriculture practices.

Peter Frohlich: 6:10

Is there academic knowledge exchange between Morocco and Canada?

Nasem Badreldin: 6:16

There is an MOU between the University of Manitoba and the UMXP, University of Mohammed VI Polytechnic, and we have exchange visits. I was there a couple of weeks ago, and the faculty, they were there last week, and there is an exchange. Basically, we are teaming up to work together to solve those bigger questions, such as the climate change impact to the crop yield prediction, as knowing in advance how much we're going to lose, how can we mitigate the loss, how we can use natural resources as most of the farms are using rainfall or rain-fed agriculture, how we can use the fertilizers more efficiently without wasting the natural resources, leaching all these kinds of extra chemicals that change the ecosystem balance in other sites. Again, Morocco has been also investing in the past 15 to 20 years. That's what I've seen in education. They sent so many scholars abroad. They lived in Canada. They lived in Europe. They've studied there. They've got their degrees. So they are really good colleagues. They are synchronized and have the same passion and the same quality and the same target. So I'm seeing Morocco is opening up to the West more, especially to Canada and more to the University of Manitoba.

Peter Frohlich: 7:26

Yeah, it's really great to see a global approach to food security between different nations. So digital agriculture, when did that concept begin from sort of more organic farming or maybe sort of simple type farming to digital agriculture?

Nasem Badreldin: 7:41

Well, I would say after I finished my PhD, I consider myself lucky. During my educational career, I have passed through different stages and meet different interesting people and joined different interesting schools in Africa, Europe, and here in Canada, and it starts to become clear to me after I finish my PhD, where I also finished from a different school, was geography. I have a PhD in geography, in physical geography. I was studying desertification in general for agriculture and environment. And then I start to realize, well, we are using all these interesting tools to predict, for example, landslide or soil erosion or desertification. How can we put it into action for the farmers, again, my grandfather, who they don't maybe have the same skills that my grandfather has. You have satellites, you have sensors. Can we support the existing farmers now to do more efficient work that become much more clear after my PhD, say around 2014 or something.

Peter Frohlich: 8:42

So what is your research area about and why is it important?

Nasem Badreldin: 8:46

I would call it my research task. Actually, instead of research area. I believe this concept with the new AI, the interdisciplinary world that we are living in, start to vanish the idea of that I am very specialized in this specific point. We are now working in, I would say, better than research area, research scope, a bigger thing. So my research task, actually, when I joined the faculty here, is to bring back three different groups. They are used to be seen separated apart. So my task is to bring them back, which are engineering, data science, and agriculture. I'm in the middle area where I use advanced technology and develop advanced data techniques to answer agriculture questions. I am, of course, starting with soil science because this is my field of research. So I'm starting with land degradation prediction using remote sensing, soil organic carbon prediction using spectroscopy, hyperspectral imaging, and so on.

Peter Frohlich: 9:43

Can you define land degradation?

Nasem Badreldin: 9:45

Well, land degradation is the destruction or the physical and chemical destruction of a soil health or soil status. And it is shown in different forms. Could be the physical loss of a soil or the chemical deterioration of a soil. Physical loss could be something like soil erosion by wind, water, or tillage. Chemical, I would say, I'm using very broad terms here, are not very specifically scientifically accurate for our audience. And chemical deterioration, something like soil salinity, for example, you see in the soil surface shining salt that is accumulating at the surface. That's a change of the chemical structure of the soil, change of the pH, for example, change of the mobility of the chemical compounds like the iron or the manganese in the different shallow soil profiles where there is anaerobic conditions. You see, some of them are visible to us and some of them invisible. My main task in the beginning of five years of my research is to work on the invisible ones, right? Something like you see a soil, you look at it, and then, of course, using equipment, using AI, and they start to realize, okay, this soil is in a sort of declining soil health.

Peter Frohlich: 11:02

Can you talk about the AI component? Because this data used to be collected using more proximal analysis, right? But you're using satellites, right, to collect the data, and then you're constructing the models to give you details about what's happening in the soil. Is that correct?

Nasem Badreldin: 11:17

That's right, but there are other techniques. We depend on the remote sensing on the satellites, but also we use different technologies around the satellite itself. Satellite, actually, they have a very specific purpose, and it's a final purpose. It's the mapping. But one of the stages that we work with is developing the AI. You know, the AI is a mimic of a human thinking. So you train that kind of AI to become much more smarter in predicting things using the features or the variables you are feeding them with. Once you become comfortable with the AI intelligence, or let's say the AI robustness, being precise and accurate and fast, then you try to scale it up using remote sensing now. Instead of using laboratory data, or you're using a plot data or a trial data, you are now going to the real world with real data, real time, and try to predict, for example, soil salinity. And we have a paper with some Egyptian colleagues predicting the soil salinity in the northern side of the Nile Delta, which is single because of the climate change impact. And it's not visible for the farmers, but for the satellites and for the trained AI become much more visible.

Peter Frohlich: 12:27

These are methods that potentially we can prevent some of these erosions, some of these degradation things by knowing sort of what's happening on a bigger scale.

Nasem Badreldin: 12:35

That's right. Actually, I'm trying to look at the invisible signs. The visible signs are as clear enough. I'm using, of course, the soil salinity. Soil salinity sometimes is very visible. You can see the salt accumulation and you can react based on that. There are some signs. before you come to that kind of conclusion, where it becomes very much expensive to retrieve, or I use the Latin word to say, mutatis mutandis, meaning changing things being changed. That's very expensive. And with the soil, things go slowly a little bit than any other ecosystem. Or you can destroy a soil within a few years, but to retrieve it back to its initial health, you need hundreds of years sometimes. And because of that, actually, I publish a paper with a group of data scientists from Netherlands about software sensors. the philosophy is how to measure things you can see. And we use three different examples, the voice of the animal to predict the disease, the spectral reflectance from the soil to understand the chemical concentration or the chemical properties of a soil, organic matter, nitrogen, phosphorus, and so on. So those kind of things become much more helpful, as you just mentioned, proactive.

Peter Frohlich: 13:40

So at what point does the producer, the farmer, get involved with your research?

Nasem Badreldin: 13:44

Well, they are very supportive and helpful to open their doors, you you know, and help us, you know, explore in their farms and exchange, because they have the science also, and they have the knowledge. Again, from my roots, where we started this conversation, I appreciate their opinion very much. They know exactly what's happening in the farm. And when I talk to a farmer, I start to say, point me where you see the decline in your crop yield, for example. Where are the areas are the most concerning ones in your farm? Because I know the farmer knows where the areas are. So we can zoom in and start looking and studying and exchanging that kind of knowledge. But again, as we're describing the satellites and the technologies and the complication, my research program, philosophy actually, accessibility and affordability is number one. So when we are developing things in our lab, we make sure that the farmers will be able first to understand the maps we are producing, because we don't want to just publish papers. We want to make change, real change in life, helping the local community, basically, to know, to learn, of course, but to take action. I want to give them the maps. But in addition to that, and this is the second phase of my research program with the accessibility and affordability, to see a better way that the farmers can reproduce and use the AI we are developing in a much simpler way so they can get the same result or nearby result with some confidence.

Peter Frohlich: 15:08

How does the farmer learn to interpret the maps?

Nasem Badreldin: 15:12

Well, we provide them with a clear indication of what we are measuring. For example, if I'm measuring soil organic carbon, say, okay, that's a percentage per unit that we are measuring gram per gram. And we explain to them the chemical analysis that we use to understand that this is a diluted or this is an actual measurement that they can see within the soil profile or the unit area or a crop yield, for example, that's another task. In addition to that, I feel that based on the latest interaction, maps could be the simplest way, colorful maps that shows degrees of stress or degrees of concentration. So they can also validate my research and findings as well. Because sometimes they can say, Nassim, this is not very much accurate. Because again, I'm using AI. And from there, I start again to revisit the analysis and see where we failed. And that's the part before we go for using satellite imageries and do the whole province kind of soil map.

Peter Frohlich: 16:10

There could be best management practices that are developed based on some of your research.

Nasem Badreldin: 16:15

Definitely. And our research findings is always open to the public, contacted by several farmers about the results we have and how they can use it. We sit with them and explain to them how this map has been developed or how this idea has been developed. We have a very strong collaboration with the potato growers in Carberry in Manitoba. We are developing different AI and digital ag-based research where we meet them regularly a couple of times a year, explaining to them the signs, but also, as you mentioned, how can you take this in your daily decision-making?

Peter Frohlich: 16:49

Some of your studies talked about doing work on grasslands. That's really important to livestock agriculture and biodiversity. What is the status of grasslands in Manitoba?

Nasem Badreldin: 17:00

This is a very controversial topic. As a matter of grassland dynamics, it's declining. The causality of the decline, it's a questionable thing. It's a debatable thing. But let me bring you back a little bit to the beginning of that kind of grassland work. We've done Manitoba grassland inventory. I started this even before I came here. I was working at the government of Saskatchewan as a modeling analyst, and I've developed the prairie landscape inventory before with a really amazing group over there. And we found that We have a really now neat algorithm or I would say most robust algorithm that has the accuracy to classify and to distinguish between three different types of grasslands. If you would like to understand the status of the grassland, you need to identify first what type of grassland are you looking for? So is it native? Is it tamed? Is it mixed? Because Canada is a very big country. Every province using a different definition of what's native, what's tamed, what's mixed. The first task we took at the time defined and finding a unified definition of what's needed. And after that, distinct different AIs and so on, and making sure that we have a comprehensive, or as we call it, a comprehensive accuracy assessment. It's not just statistics that tell us that our model is performing very well, because I can tell you based on experience, you can have 90 plus accuracy, but in reality, you have less than 20% accuracy, you see. So then we went back to the ranchers, to everybody in the farms, give them an application where they can point out where we failed, where we suck. Then we tailed back the technology again and, you know, look at it. And in the past year, we've been successful mapping the whole prairie. I will give you the number. 47.5 or 0.4 millions of hectares being analyzed using AI for every 10 meter by 10 meter pixel classified to see if it's native, tame and mixed. Now the task is, since we now know what's tame and what's native and so on, to do a change detection. Go back in time. So we use now Manitoba as a pilot study. We develop a neat technique using different satellites, using different AI, because within the AI, well, AI actually is not a very accurate word, actually supervised machine learning algorithms. We use different ones and we compare all of them and see which one performing better, why, what kind of features or inputs they are using to develop. And then after that, sharing it with the others, you know, as the growers community and other community and see what they are doing. feedback now we are in the second phase getting more ground truthing points field samples to enhance that kind of accuracy after we achieve this we will go in detail and do every year grassland but i have done it as my homework and we found there is a decline in the grassland there is a chance that we are losing i will give you now a rough number but i have to be specific honestly because grassland can be lost in different directions could be lost in cropland could be lost barren soil, could be lost as tamed grasses, could be lost as shrublands. The brush encouragement is also a considerable thing. There is 5% every year loss comes from brush encouragement in Manitoba. There is loss in the wildfire. The major loss is, that's a surprising finding. We have, of course, to do much more deeper research on this. It's not just the conversion of croplands because farmers being blamed to be the reason for losing. Actually, Climate change is not less than the farmer, you know, croplands conversion in losing the native grassland. I would say around 20,000 hectare every year loss of grasslands because of climate change impact.

Peter Frohlich: 20:44

It just makes sense to have the technology to see where the grassland is, what it is, what form it's in, or how to mitigate the loss of grasslands. So out of all the studies that you've done, and there's a lot, what is sort of the most surprising or the most interesting finding that you can recall?

Nasem Badreldin: 21:01

In which field? In the grassland one or in general?

Peter Frohlich: 21:03

Just in general.

Nasem Badreldin: 21:05

Well, there are so many, to be honest. And one of them is we had the assumption that if you're using the most sophisticated, the most sensitive technology, you get the most accurate results. Actually, this is not the case. Most of the cases, you could use a very simple technology such as, you will be surprised to tell you that there is maybe just the mobile camera can give you some soil property these information instead of bringing this kind of half million dollar piece of equipment to the field drone or whatever they are they are very expensive hyperspectral cameras and spectrodiometers very sensitive as well where you get the noise and the error instead of having a phone that can give you actually similar results or nearby results this is coming back to the accessibility and the affordability so we can help farmers actually have an application by just scanning or taking a picture of a soil knowing how much clay content, how much soil moisture it is, how much organic matter there is. That's pretty powerful. This is a surprising finding, and we have actually a paper starting asking the question, and that's the thing, starting asking again, do we really need to go that far with the complicated technology? We need it for our purposes in order to come up with simplifying tools. We need to go far to the complicated ones. But again, the simplified tools could be as powerful as hyperspectral cameras.

Peter Frohlich: 22:28

I agree 100%. These tools need to be affordable for the farmers. Farmers, I think, have a difficult time investing into a technology that may be good for the environment, but if it's not profitable, they'll have a hard time believing or adopting it. So this is a great finding. Beyond the research, can you tell us one or two things about yourself that's not related to research?

Nasem Badreldin: 22:49

Well, sometimes it's difficult because my work is my hobby. I'm having fun doing research because each time I'm exploring something new, something amazing. We are now in the crop yield prediction where I'm learning new things from scratch. But beside that, I really enjoy traveling. I'm born in a country and raised in another, educated in another continent. And now I'm living, I'm a Canadian here, contributing to the economy here. I found that it's super fascinating and super educational, enjoying when you travel. I've just had a vacation where I travel about 5,000 miles in one week, driving to the ocean city, just the sake of seeing Eastern United States and enjoying them in the Fourth of July, just to see how people celebrate such a day. That's a kind of enjoyment that I have in meeting interesting people, you and others. That's a fascinating thing.

Peter Frohlich: 23:42

Very cool. And that's it for today's episode of ChangeMAKErs. Thank you, Dr Badreldin, for chatting with me today about your fascinating research and about your path to become a researcher. And thank you all for listening. Join me in future episodes of ChangeMAKErs to hear about fascinating research being led by agri-food innovators at the University of Manitoba.