Water Resources Podcast - Sergii Skakun
[00:00:22] Bridget Scanlon: I'm delighted to welcome Sergii Skakun to the podcast. Sergii is an Associate Professor with a joint appointment in the Department of Geographical Sciences and also the College of Information at the University of Maryland. He is an investigator in the NASA Harvest and NASA Acre programs.
Today we're going to talk about his applications of remote sensing to quantify spatial and temporal variability on agricultural production in Ukraine and Western Russia. His work consists on the impacts of climate extremes, droughts and floods, and also climate change, and also the recent full-scale invasion of Russia in Ukraine. And, so, I'm delighted to have you and I haven't previously interviewed anybody about Western Russia.
it's great to have you on the podcast to focus on this. Thank you so much for joining.
[00:01:19] Sergii Skakun: Thank you, Bridget. Of course.
[00:01:21] Bridget Scanlon: Sergii when I was looking at background on Russia and stuff like that I was really clueless. So maybe you can talk a little bit about giving background on the evolution of Russia and the former Soviet Union, population and things like that to give the listeners some background on those topics.
[00:01:39] Sergii Skakun: Yeah, of course agriculture was always since many, many years, a big topic. And it was first of all, economically important, but also socially important. A lot of people, one way or another, involved in agriculture. If you're going back before Soviet Union was formed, those were like a small holder farmers.
Actually, I found out that my great-great-grandfather was in Khmelnytskyi oblast. He was a farmer, but then he was actually jailed for counterrevolutionary activity. And then when the revolution happened, Bolshevik revolution, of course, and they collectivized everything.
And that is how we started having in Soviet Union, “kolkhoz”, collective farming. And then my, his father, his son, was jailed for the anti-Soviet agitation, something like that. I mean, a lot of Ukrainians, of course, have this kind of story with their ancestries just saying something or trying to preserve the private property back then.
Actually, in my personal opinion the conversion from where those farmers had the land and then Bolsheviks came and everything become collective. I usually call it collective irresponsibility because those were usually very ineffective enterprises. Nobody really cared because it was nobody's. And that is what we had in Soviet Union for majority of parts.
So the main players at that time, it was Republic. Now it's separate countries. Of course, it was always Ukraine with their famous soil. Very fertile soils. Actually, there are well-documented facts that during World War II, when Germans, Nazis, would come and they would cut half of the most productive lands, and they're going to ship back into Europe; of course, Russia and then Kazakhstan.
There are a lot of important crops, but wheat of course is one of the major ones. The Soviet Union and the Soviet Union collapse and conversion to the more market economy in many places those cohorts were either privatized or just were dissolved.
And a lot of smallholder farmers started in some countries. For example, Georgia, you can just see if you take Corona satellites from 1960s, you can see huge fields and compare to the contemporary data. You can see very small fields. And then of course there was a huge transition.
Some of that period is characterized that some of the land was actually abandoned and there are some research documents on that. And of course, as the economy started getting better, because those early nineties were very terrible economically speaking, mainly in Ukraine.
Even from personal experience, if you had some relatives living in villages so you can grow something that it was a big deal because at least you can supply for self-consumption primarily. But then as the economy grow and the enterprises and agriculture started to be more advanced technology, et cetera, now we have pretty advanced enterprises.
We still have some of the crops, very important ones, especially locally like potato has been still produced 90% by the local farmers. Whereas things like wheat, it's mainly produced by large agriculture enterprises. There is still going on a lot in terms of land because for a long time, land was very sacred thing in terms of giving access to outside investors or investors from outside.
And there is a lot of programs involved to make it land as a commodity. This is where we are right now. Everything that's going on in Russia and Ukraine now because of that impact on the exports. Russia is the largest exporter of wheat. I was looking at the different numbers before full scale invasion. The export was going to a lot of middle East, African countries, Ukraine goes also for mainly northern African countries in central. So everything that's happening in this country in terms of wheat then it populates to the markets.
Other countries that is important here, is also Kazakhstan. They mainly plant spring wheat, which is it's less productive. If we estimate wheat production in Ukraine Russia when the Soviet Union collapse, the good one was around two tons per hectare. Now we are talking about four tons per hectare. Whereas like spring wheat, it didn't change a lot, and in Kazakhstan it's still around one ton/ha.
So if you compare spring wheat with winter wheat, we're talking about factor of two to four in terms of production with the same area. And again, given, as I said before, huge impact on the export. Everything that's happening, whether it's droughts or things like armed conflict and invasion of Russia to Ukraine.
It echos for the markets and it's extremely important to monitor. We're at the place where nobody really can go in Russia in that sense. USDA foreign agriculture service, they used to do the field campaigns both in Ukraine and Russia.
But then, when the situation became political, let's say, unstable, it was too dangerous to go to Russia. So now, one of the biggest sources of information, of what's going on a repeated basis, you can use all of the social media data some intelligence data, et cetera, et cetera. But satellite data becomes a very much integral part of that.
It's very similar too. When you talk about Landsat and you talk about LACIE (Large Area Crop Inventory Experiment) program, where one of the biggest applications was to look what's going on back then in Soviet Union in agriculture. Whether there is something going on, whether it's better production, worse production. So right now we are getting with a much better means of satellite data. We can get information on what's going on in agricultural sector in those regions.
[00:07:27] Bridget Scanlon: It is great to get your perspective Sergii, having grown up in Ukraine, in Kyiv. And then learning about the history and everything. And, I think in 1922, was when the Soviet Union formed and then late 1920s, then the collectivization.
And there was a lot of trouble related to that.
[00:07:45] Sergii Skakun: I'm by no means a historian, I'm just from personal perspective, but that feeling of private property, and I know that, sometimes , in US, they take it for granted. But, there is all the things that's going on with it. It's just killed this feeling that it's yours. After those Soviet Union occupation of Ukraine, basically for 70- 60 years. It is just killed that. From historical perspective of what Stalin was doing is that he was taking anybody who was conscious about identity and identifying them as Ukrainians. They would put them in Siberia, whereas in their place, they would put in Russians from that area.
And that is where the part of the problem is, they claim, that they cannot help Russia, they cannot do this and that, and it's just historically it was done basically artificially. By that time, including Stalin, and that is where some of those problems start.
Those seeds were planted back then.
[00:08:45] Bridget Scanlon: Right. Of course, in ‘91, then with the dissolution of the Soviet Union, Russia and then the 14 or 15 other countries, there was a lot of change going on then. And as I think you mentioned, abandonment of land- it was a very difficult time.
[00:09:01] Sergii Skakun: Yeah, exactly. You were in survival mode so to speak. Those early nineties it was just horrible in terms of what you can do, but at the same time we survived. Now of course situation, given the war, before the annexation of parts of Crimea. It was going at least in the right direction in terms of having a goal to join the European Union and things like that. They were like on the positive side.
Unfortunately, after Soviet Union collapsed, there was still, for a long period of time, one of the largest party was the Communist party and they were all pro-Russians. They would not let any different, topic or different ideas of being more pro-European.
That also did not accelerate to the point to join like Baltic countries. I was always a little bit jealous in a good way of Baltic countries that they were able much more quickly to join NATO, and then to join European Union.
I mean, it's much more complicated. Maybe I'm painting it a little bit more simplified ways because it's a very complicated process. But yeah, that was the early nineties.
It is what it is right now.
[00:10:12] Bridget Scanlon: I was really clueless about the population of Russia. Until I started looking into it, I had no idea. And so seeing that 140 million people in Russia now, I thought it would be much bigger. That's like half of the US population and it's such a huge country.
Of course, most of them live in the European side, west of the Urals.
[00:10:31] Sergii Skakun: And it's declining because of various factors. And if you're talking about agriculture, yeah. Maybe GDP wise it's small fraction. Of course, you cannot compete with oil and gas industry. I was just looking that the labor force in agriculture declined over the past 20 years from around 12, 15% to 5%.
It's very difficult to evaluate just numerical agriculture because how do you evaluate food security? It's such an important thing that a lot of depends on that. And even though a product GDP wise itself, it might be fraction from social standpoint it's a huge. Even though Russia gets billions of dollars exporting to those countries, unfortunately, they now they use all of this money to convert into the weapons that they terrorize basically Ukraine. I've guessed that export will not happen even under the sanctions, but it is the state of the art right now.
[00:11:29] Bridget Scanlon: Cereal production, is a basic food and very important for the food security. When I talk to Australian economists or people like that in talking about cereal production, they say they're always moving towards higher value crops, vineyards and other things. So we'll all be drinking wine, but we won't be eating any bread.
The profit margins are pretty small for cereal production. Ukraine and Russia, top cereal producers. I think Russia accounted for 25% of global wheat exports in 2017 and 18 in one of your papers, and maybe Ukraine, about 10% prior to the invasion.
So extremely important for global food production. You've done a lot of work on winter wheat and spring wheat and wheat production in those countries. Maybe you can describe it a little bit.
[00:12:17] Sergii Skakun: There are so many dimensions and when you start digging into that, some of these fascinating facts. So winter wheat, in our latitudes, you start planting that from September, for previous year. Then it goes in dormancy.
So the main idea is that you would have a good snow cover. Which, first, will protect from the winter temperatures, and then it's going to feed with the soil, the moisture. That will be very strong, and that is why usually the quality, the protein level, and everything that goes with that for winter wheat is higher than the spring. So that is why it's preferred.
You still have maybe a little bit more risks because the longer growing season. Usually, they would prefer because you can get more revenue, more return on investment. That's why, for example, in terms of if you're talking about Russia, the winter wheat, majority of it goes to that export like you said, around 30%. Which is account around six, seven depending on the price. We're talking about up to 10 billion of dollars, depending on the year.
Then majority of the spring crops because it's not competitive, even on the market, the price is lower. So usually they use it for self-consumption. In Ukraine, majority of wheat is winter wheat. There is small fraction of spring wheat, but usually within the crop nomenclature it's a winter crop.
The governments in both countries. They monitor what crop patterns are. So they even have recommended how much of the entire cultivated land. So that's why you're right that there are more profitable crops.
But at the same time, they would monitor. I don’t know to what extent they can dictate what farmers plant. In the kind of market economy, but because it's really a part of national security, it's important to have that. Otherwise yeah it becomes an issue.
I already mentioned that through those, after Soviet Union collapse and collapse in both country, winter wheat saw major improvements from around two to four tons per hectare. A lot of this associated with the better practices, better machinery, better seeds, but at the same time, what's happening on meteorological side.
We almost don't have wheat in certain snow in certain regions. I remember when I was a child and we always have very snow winter. I'm not even saying- when my parents were children, they didn't talk about snow- it's how much snow. It's one or two meters, something like that.
If you look at the present time, there is very little snow. You would think that okay, there is larger likelihood of having frost. And of course, with temperature rises and droughts becoming more frequently, you would think how it's possible that two converts to four tons/ha.
Apparently, this is what is happening, and that practice is to certain degree mitigated. But you were right saying it's still, it's not like the margin of profit increased considerably. They would rather plant, I don't know, sunflower, or grapes or some fruits. Things like that, that, that would get more profits on that.
And then throughout these years, if Ukraine stayed more or less. There were some changes, but in terms of land use, just because, there is as much cropland area as is in Ukraine, but in Russia it's huge increase.
Some of the oblasts, for example, Rostov oblast in 2000, it was less than millions of hectares, whereas right now it's almost close to three, 2.8 million ha. And some oblasts saw three to five and that was done intentional because they want to expand that. There was a program signed back then when Medvedev was president, they wanted to increase production. That is why they were able to become one of the- 20 years ago they were not a top exporter- now they became a top exporter.
And then if those trends more or less, keep right now- of course it's not linear, but our estimate that they can easily like, do around 50 - 70% more to make it sustainable. To get to that, having two around two within the four-year cycle, having two winter crops. At the same time we have situations like 2010 where there was a huge drought in that region.
Russia had a 30% decline. Kazakhstan had 30% decline in production year to year. Ukraine had around 17%. what happened then they signed that they put an export ban for the wheat and of course, market. I think like price in couple of months increased, like twice. that is where the time when, one of the system that NASA Harvest and Acres contributed, GEOGLAM (Group on Earth Observations Global Agricultural Monitoring)was formed. It's global agriculture monitoring system.
Actually two systems were formed. It's a GEOGLAM and then agriculture market information system with the idea is that. If we see that something might happen in terms of production, especially in those countries, that's important for the export. The chief economist, they gather and try to reach consensus in terms how to deal with that situation. In order to reduce the potential impact on markets, because of course, again, we are talking about some of those countries that food insecure countries in Africa, that Ukraine and Russia export to. It has very profound impact leading or increasing at least the likelihood of potential of famine or people not having food.
One of the main- of course not the only one, but one of the main sources for data in GEOGLAM, it's satellite data. Again, because of those strengths, that advantages, that it brings. It's repeatable, it's consistent. We can extract a lot of the variables about vegetation and including crops and try to know where it's grown, what kind of crops it's growing, and then how much it's going to be growing. And ultimately to gauge the feeling of how much of production will be. And of course, those countries, given current situation, we built quite mature expertise within our department and in general the GEOGLAM community.
And of course after full scale invasion it's even amplified just the need for that.
[00:18:30] Bridget Scanlon: You mentioned the low margins in wheat production, even for winter wheat, where you've got two to four tons per hectare. Then you have a four-year cycle. Maybe you plant two, two years and then you might have another crop or you might have a fallow period. But you've also done quite a bit of research on sunflower, and I think one of your papers I thought was very interesting. Showed the value of remote sensing, because oftentimes we make these basic assumptions.
Oh, there's an invasion, or then the sunflower production's falling off a cliff, and so it's really nice to be able to quantify these things. Maybe you can describe a little bit about the sunflower production and how that fits in and how much more profitable it is.
[00:19:09] Sergii Skakun: Sunflower is a fascinating crop and it has been always big deal in Ukraine. Usually we use it for the sunflower oil, which we use for dressing, we use for the cooking and all kind of sorts.
The good thing about Sunflower for farmers it is profitable. Whereas the downside that there are some crop rotations nomenclature that you should follow. Because usually sunflower on average is very demanding in terms of soil content. So is this harvest everything? It might have potential impact. So for example. Sunflower by Ukrainian law in the same field recommendation that you can plant only once per seven years. The reason for that is just that other crops would allow to rest the land. Actually, land if you think about, it's almost like an organism. You cannot take everything.
You have to make a balance, right? And the same balance comes to the crops. There is an entire domain of crop type mapping with satellite data. We use both methods and those are trying to, again, to do something global that we can, but at the same time, we are not trying to just throw into the washing machine, just get something.
We really try to understand phenology, then we know what drives that model. Ultimately, and one of my PhD students, Abdul, he actually looked at the SAR data or synthetic Aperture Radar data that operate in the microwave. He found some features about helio-optic behavior. Actually sunflower, it follows sun to certain degree until the head becomes too heavy that it's "okay, I'm not moving that head anymore."
And he found that certain features that you can use. And in such a way, he built some model, so he understood phenology. We were able to map and then, look at when the full-scale invasion happened. We looked at the, what's going on in Ukrainian government controlled regions, temporary occupied by Russia, and the movement of hotspot.
We saw that there was a movement from the frontline, which of course makes sense where the farmers would plant more in there. We didn't see a lot of drastic changes. Of course, there were some changes, and now we have to, when we do all these calculations, one of the advantage of satellite data, now we have to do it separately for Ukraine control regions and temporary Russia occupied regions because the sources of information are different. We were able to do it with sunflower.
We continue that work to incorporate other crops. Not only winter crops that we already were matured and sunflower and how we expanding that work to crops like maize and soybean. And to see how we can use similar approaches, just different data sets and different phenology features in order to do those estimations. Because we work in, within not harvest, we work directly with the Ukrainian end users. Whether it's ministry or statistical agencies. Where we try to develop approaches and compare their estimates because they're also limited in data for Ukraine controlled areas. But then they don't have any information coming from the occupied by Russians. But we know that agriculture is happening there.
The big question is where it goes we still don't have a good answer to that because it doesn't go into the Ukrainian statistics. It does not yet go to the Russian statistics. If you're talking about balancing all of that, we are talking about billions of dollars, even just grains.
And Russia basically, literally- I'm just going to call it- they're just stealing that crop production and just use it. Probably it goes in black markets and not accounted and people making money out of this.
[00:22:51] Bridget Scanlon: One of the fascinating things that I found, Sergii, when I was looking at your research and stuff and reading many of the papers, was that the impacts of climate change on cereal production in Russia and north versus south and intensification versus expansion and how you did that with remote sensing and also the potential.
Most of the times we hear a lot about all of the negative impacts and not so saying that we are pro climate change, but some places will benefit and it seems like Russia is potentially one of those areas in cereal production, wheat production, is one area also.
[00:23:31] Sergii Skakun: Those southern regions, there already was a quite intensive use of winter crops in terms of yield and increase in temperatures. Projections usually that they're going to start having lower yield. We shall see how it's going to, and how much the technology and everything that's related. Fertilizers, can they compensate or not?
But that rise in temperatures which led you to, goes up north. You absolutely right. That actually benefits because the season becomes longer. Rather than planting spring wheat in the very short cycle, they can start planting winter wheat. if you look at the maps and the data, it increases with the latitude.
Some of the places that 20 years ago, they never would plant winter crops. Now they're doing. And a majority of that winter crops is winter wheat. I should make a remark that I'm saying winter crops rather than winter wheat, because it's very difficult from satellite data. We don't have mechanism to reliably dissect wheat from other cereals like barley, oats, et cetera. Especially at global level it's extremely difficult. But that increase in temperatures and longer season really allows to expand on that. This is what we saw, and we think that is going to continue.
And in general I have another PhD student, so he works on the entire globe. It's still in the making, so to speak. But hopefully he will publish soon about how that winter crops moving in terms of spatiotemporal globally. And we see that, some of the regions actually.
You have to factor, it's not like it can move up north to anything in terms of soil productivity, right? So this is can be constrained factor. But for that, we would need good data on soil types. Some of those exist, some don't exist. Unfortunately, soil Is extremely difficult subject to investigate. A lot of the data that we have on soil they're very old. It's very difficult to measure productivity very frequently. But we will need the data if you want to see what are the constraints where the winter crops can go up to the north.
This is one of the potentially limiting factor that tell you how much it can go. In case of Russia, actually, it can just go in the areas where it was spring wheat. Because we know that might not be the most productive lens, but it's still productive. You can grow agriculture commodities and if situations will, besides that, at some point they will say, yeah let's do winter wheat instead of spring wheat.
You have similar areas around the similar areas, but you will gain around factor of two. Maybe not too fully depending on the types, but you can get better production and that is your potential to even further cement for the export.
[00:26:14] Bridget Scanlon: Those papers that consider both intensification, which you said accounted for about a third of the increase, and then expansion about two thirds. It'd be nice to describe a little bit the change detection that you did with MODIS satellite data.
Taking the early data and then the later data and being able to evaluate those changes.
[00:26:35] Sergii Skakun: MODIS, which is, this is a sensor that was installed on Terra and Aqua satellites. I consider this one of the masterpieces of NASA scientists and engineers and everybody involved because they will launch those two satellites were launched in 2000. Think about that. And they still fly 25 years later. the good thing about MODIS, so we have twice, two observations per day.
So usually Terra am and Aqua pm and of course spatial resolution, its ranges to 500 m to one kilometer. For example, if I were to do something in South Eastern Asia or Africa, that resolution will not help. But we can go back 25 years ago. We have enough temporal coverage to actually capture the signal from winter crops.
And this is what we did and in landscape in Ukraine and Russia in such fields that big, so we can capture with that spatial resolution. We looked at what was going on in only in four year period, in 2000s. We applied the same algorithm when we produce the maps.
One have to keep in mind, that maps can have errors. That is why our community develops statistical approaches in order to get unbiased estimators for the area. We take the map, we make them stratification depending on the area of your interest.
It can be, let's say changes. It can be gains, it can be losses, and then we throw sample unions that in an ideal world, you want to go and see what was there. Sometimes it's impossible even when you would be able to go, but you cannot go with the time machine in 2000. You build that expertise through the field measurements where we can, whether it's in US, whether it's in Europe, et cetera, and we know how the signal from the satellite looks.
Then you build the skills and expertise in terms to do photo interpretation. We have the sample units and then several experts try to really say whether it's winter crop or winter crops. If we disagree, we try to reach some of the consensus. in such a way, we built basically statistical models and statistical estimator of how much land changed.
It's similar to any survey when it's not feasible to do the entire census, right? You take this sample and here the role of satellite data is to improve efficiency. Efficiency improves in terms of whether precision of your estimates, so how much you can reduce the variance of estimates. If you're talking about the changes and you have something, let's say 10% plus minus 20%. Statistically you cannot say that it's actually, but with the stratification, you can say it's 10% change plus minus 0.1. Satellite data guide us, whereas then we use statistical approaches to estimate with the corresponding uncertainties.
And having the maps will allow us to do all the fancy stratifications. How many times per four year cycle, it was planted, if it's our interest, or what are the gains, losses, et cetera. And we use MODIS for that. We could not use Landsat because in 2000 LandSat was available, but usually Landsat, it's good enough for estimating cropland.
Our colleagues from other groups in our department, from Matt Hanson’s group. They did global cropland using Landsat. But unfortunately, in the early 2000, even 2010, before the sentinels, unfortunately it's very difficult globally to map specific crop types.
Maybe you log it in some cases, yes, but for the entire Russia, there is no alternative to MODIS, unfortunately. And that is how we used it. It has lower spatial resolution, because everybody is saying we need to see all the details, et cetera. But nothing can beat MODIS.
This is if we're talking about 20 years ago. If you want to have something with the high temporal revisit cycle. With Landsat and Sentinels, we can have it since basically 2015- 2016, for the past around 10 years. That is a new era where we can do much cooler stuff.
[00:30:25] Bridget Scanlon: It really is fascinating that you could take that long-term record then and see those trends, intensification and expansion, and then link it to climate and then make projections that you think it will continue and get an indication of the future capacity. Maybe still another 50 to 70% increase that you project.
You mentioned the difficulties of dealing with the soils. The soils in the south or in Ukraine, are they black organic rich soils or they're very fertile? with the permafrost melting in the north, is soil structure and all of that an issue?
And how do you compare, expansion in Russia versus, say Canada? Some people say we don't expect that much northward shift in cultivation in Canada, maybe because of the soil limitations.
[00:31:09] Sergii Skakun: That's exactly when I was mentioning about the potential constraints. This is what I've read about Canada. There is potential, very similar In Russia's situation. I think Russia has more potential.
You're absolutely right going back to Ukraine, it's black when you just see it is really black. Soil is a big factor in that. Now there is always a lot of debate about how you treat soil. For example, we're talking about till or no till.
In US there is a lot of no till. It's indicated that it's positively impacts. Traditionally in Ukraine and in Russia, we all do tillage. There are pros and cons. I think there are still no big consensus, it's very difficult to convince them to do no-till for some reason, they continue doing tillage.
Returning back to your original question, soil is extremely important. I think those projections at some point, especially we are talking about modeling the impact of the agriculture. It's one thing when you project, and it is important to know how yields would react, but ultimately land is another factor that you can grow as much. In having those good projections, what's going to happen with land productivity and composition in those higher latitudes?
I think still it needs a lot of research done. I don't think we have clear answers on that. Maybe some PhD students will take on that. Especially at global scale I think a lot of done at regional scale, which a little bit more feasible, but at global scale, especially when you don't have access to high quality ground data.
And yeah, that's still big unanswered questions.
[00:32:48] Bridget Scanlon: You mentioned, Sergii, the drought in 2010. I think it impacted both the winter and spring wheat. The embargoes on exports or the bans on exporting grain.
Russia has now a food security doctrine, or something, and I think maybe that impacts some of those things. You hate to see people having a food security issue at home and then they're exporting. The potato famine in Ireland, they were exporting a lot of food to UK, but, so that goes back a long time.
[00:33:20] Sergii Skakun: I was looking, just recent numbers in terms of labor force that I mentioned. They put basically their right now economy on the military rail, so to speak, in order to terrorize Ukraine.
And they now have the shortage, but at the same time they want to increase production of major commodities including wheat, steel, nature of infrastructure. In terms of where to grow that, and meteorological can just actually it favors them. It's a question about, to find enough people, enough investment in there, and they can do that.
My personal opinion and prognosis that one, they get to the critical mass. In terms of how much they can dictate things on the export scale stage, I would suspect that they would be using that as a leverage. Saying "we just going to do the export ban, and you figure out what's going to happen with those countries." As much as they will be expanding that, increasing that, they would have more leverage through that.
That is why when the full scan invasion happened, there was a lot of talks at the United Nation level to limit that export of wheat. It's said that they literally used that billions of dollars to then fund their military stuff. But then everybody was scared how it's going to impact.
This is the world that we are living in. I wish they wouldn't earn those monies. And I really don't want to make this podcast political, but as Ukrainian, I cannot just easily talk about that and address the elephant in the room.
[00:34:51] Bridget Scanlon: You mentioned earlier, the many countries in Africa, Kenya, and Egypt, and Mozambique in all of these countries, depending on imports from Russia and Ukraine. During your career, you have seen the explosion of Earth Observing Systems and, its value and, getting it into the normal realm of statistics. In getting its acceptance in decision making and policy making.
I was listening to some YouTube stuff, they're talking about NASA Harvest. Early on maybe you had a single satellite that you were dependent on Landsat or whatever. Now, you mentioned Landsat and Sentinel harmonizing those. It is usually they have one or the other, high spatial or temporal resolution, or high temporal, low spatial, But you're able to bring all of these things together.
Sergii, you were one of the early adopters of machine learning. You've been doing it for 20 years, and so you were way ahead of the game. That's really nice to see. So, allowing you to bring as many different data sources together to quantify what's happening and then quantify uncertainties and increase your reliability of the results.
Maybe you can describe that evolution a little bit.
[00:36:00] Sergii Skakun: I mentioned about MODIS it's a great sense of great satellite, but of course it's limited in spatial resolution. Now, we are using all available data, and by all we are using both Landsat 9. But together, Landsat 9 gives around eight days revisit cycle. You add Sentinel 2. Currently we have A, B, C and it's gives you basically two, three days revisit cycle. And I'm talking right now about open and free satellites, like MODIS.
We are not yet at the level of MODIS where you have two observations per day, but still two, three days. It's much better than in older 2000s. We would just have one Landsat with 16 days and no way you can get. Especially for example, winter crop map and it's usually you want to have a good signal in the early spring, but that region it's always cloudy. That is why you need to have that very high spatiotemporal resolution. And now with the Sentinel One, which is SAR satellite, which allows to penetrate clouds.
For a long time there were a lot of research done on SAR, but usually those were like not operational missions. And now with Sentinel-1, actually just two days ago yesterday, they will launch Sentinel 1D. It will add to the spatial resolution temporal resolution because one of the satellites unfortunately has malfunctioned. Kudos to European Space Agency that they are providing this data. Without those data, it would be difficult to survive.
Having those multiple satellites, of course you want to make sure that they're harmonized. From end user perspective you're going to treat it as is data coming from the single source. So that is what NASA did a very good job on organizing that harmonized Landsat and Sentinel two product.
We have that good temporal resolution, sentinel ones are, it's a different beast. But then you can have features, like I mentioned about sunflowers, sometimes unexpected. Depending on the crop phenology, we use either combination of those or specific satellites. So of course then we can take advantage of that.
In past 10 years, we had a huge boom in commercial satellites, like Planet Labs, a three meter resolution. You have daily, almost daily data of course it's commercial domain even though for research purposes within NASA, commercial satellite data we can use for research and see the benefits. If you want to then move signs to operation, that is where you start to say, "okay, what are the tradeoffs, what are the benefits, how we can quantify that?"
Do you really always need that three meter resolution? Sometimes you don't. If you're talking about big enterprises, probably 10 meter will be enough If you're really talking smallholderfarming, like in Southeast Asia, I mentioned before, Africa, you really have to go one meter.
It's all depends on the tradeoffs and also observation jargon. It's all about tradeoffs. You correctly mentioned that it's usually either you have high temporal resolution or high spatial. We don't have everyday data at one meter or half meter. I think technologically, it's possible, it's just will be a flood of data and the question becomes, "okay, how we make use on that?"
With the new methods of AI and the mail you can even more optimize now very popular foundation models. Google recently made publicly available AlphaEarth. The main idea there is that you already have pre-computed something that would allow you to even more make it efficient in terms of whether it's training data or pre-computed features. This is new and interesting stuff.
Again, you still need to do some good research in order to actually prove it. One of my recent research studies was on the impact of satellite based maps on the error estimation. And there is a very nice map behind that, but also, we clearly know what efficiency mean.
I already mentioned about either variance, precision of estimate or sample size, because we need to collect those samples, whether it's through ground measurements or through photo implementation. And it takes time and effort. So we can even put in dollar value, economic values, and see. You would spend, I don’t know, hundreds of thousand dollars to make a little bit improvements in the map. But for area estimation, it actually, the improvements will be marginal. And you would not get very little gain on that, on the investment.
Whereas if you get very good improvements for very little and something was pre-computed already and you invest less, but getting better accuracy, so better efficiency that we can estimate that. And implementing all of this and end user side, yeah, there, there is that last mile to go from science to operation it's always very difficult. You have to build trust and you need to prove that what you're producing is what expected by the end users. I think within acres and higher waste we are doing some good job on that.
And to educate, but also to do some research. And some of those technologies to implement on their site so it can benefit their decision-making processes.
[00:41:04] Bridget Scanlon: Really fascinating and as you said, the 80-20 rule, sometimes we get stuck making small things and it makes it making a career outta something that's just going to give us a marginal. Improvement.
Our guest today was Sergii Skakun and he's a professor at the Department of Geographical Sciences and College of Information at University of Maryland.
I'm extremely grateful to him for educating me on Ukraine and Russia, and food production, and remote sensing. And I hugely admire the work that they're doing in keeping track of all these things. Global food trade and food security and all of these things are very important.
So thank you so much. Sergii.
[00:41:44] Sergii Skakun: Thank you, Bridget, for having me. Thank you.
