Skip main navigation

Course Introduction

Welcome to AI for Earth Monitoring!
7.7
Dallas Campbell: Hello, and welcome to this course on artificial intelligence for Earth monitoring. I’m Dallas Campbell. And I’m going to take you through this course as you discover how cutting edge artificial intelligence and machine learning technologies are helping to advance Earth observation science and the benefits that this has for citizens. The course will provide you with an overview of the different types of AI and ML, and the fundamental techniques of working with AI algorithms for analysing satellite base Earth observations and in situ data. You’ll also be able to work with hands on tutorials using Jupyter Notebooks. So you can see for yourself how AI works. We’re going to cover four thematic areas so oceans, atmosphere, land, and climate.
55.9
And in each of these topics, you’ll be learning from experts who use AI to discover more about Earth and to solve some of the fundamental problems that we face as a society. This is a Copernicus course developed by EUMETSAT in partnership with ECMWF, Mercator Ocean International, and the European Environment Agency.
81.6
We’re going to start by talking to Dr Mark Higgins, Training Manager at EUMETSAT and Paolo Ruti, EUMETSAT’s Chief Scientist. Mark, Paolo, lovely to see you. So I want to just quickly talk a little bit about AI. And obviously, we’re familiar with the term AI in daily life, people think about, you know, Terminator killer robots, or they think about self driving cars, or Alexa, whatever it is. But when you guys talk about AI in terms of Earth observation, whether it’s satellite- just satellite data, or in situ data, what do you mean, what do you use AI for?
113.9
Paolo Ruti: We mean to, to get the data closer to the people, they live on the street. So if you have the storms coming, and you want, you see the storms from the satellite, and you want to integrate that information with other information that you have from Radars, and you want to do it quickly, very quickly. So that’s where AI and machine learning makes sense to us.
137.6
Dallas Campbell: And is it just because there’s there’s such an enormous amount of data, you need AI? I mean, presumably, the amount of data that you’re dealing with is vast?
144.9
Paolo Ruti: Yeah, that’s, that’s what in the in the recent years, or in the last decade, that was an exponential growth of the amount of data we are getting from the satellites or from any other sources? So this is certainly why we need to look to the data very fast. And the human brain is, is, he needs a bit of help.
167
Dallas Campbell: So speed is one thing, volume of data is one thing. But is it also about using AI to make sense of all that data to actually kind of give you answers to questions that you’re looking for?
178.9
Mark Higgins: Yeah, I mean, it’s one of the tasks we have in understanding the system is to find patterns. And AI and machine learning really help us to do that. And that could be patterns in terms of, is there something here that is extreme? Is there something here that is dangerous, or unusual? In the climate realm, there’s quite a lot to help understand, what are the particular sensitivities of the Earth system? So of course, we have loads of data on the whole Earth system. And we can start to say, well, what are some of the really sensitive places on the Earth to climate change and things like that? So that’s, that’s another growth area of AI and machine learning.
216.2
Dallas Campbell: Yeah. But I mean, just on the volume side, could could you make sense of, of the data that you’re dealing with, without AI, for example, is, is there a way that the human brain can just sort of deal with it somehow?
228
Mark Higgins: Not with the shear rate. But there’s, underpinning our world, there’s a whole ton of algorithmic stuff, people processing the data, using algorithms and computers. So we’re very I mean, we were talking earlier, so this goes all the way back to Ada Lovelace and sort of 1814 when she’s developing algorithms to compute particular series of numbers. And just that heritage, what what she started is the root of what means that we can exploit our data. And the AI, machine learning helps us to do some stuff quicker, but also gives us robust ways to find patterns that we can explain. And finding patterns in the Earth system really matters.
265
If we’re going to understand, not just climate change as one big whole thing, but exactly what difference is it making? Where on the Earth exactly is more sensitive than say other places? What about the Arctic environments? What about places in Europe? What about places in Africa? And sort of being able to really pinpoint these using AI to find those patterns is really helpful.
286.6
Dallas Campbell: So really, it’s a pattern recognition and obviously human beings are good, are good at pattern recognition, but the the AI is a tool that kind of helps us refine that, I suppose.
296.9
Mark Higgins: Yeah, I mean, I guess as humans, we can find patterns. But the AI is a little bit more robust, we can get numbers and statistics to say, is that pattern really there? Because of course, as humans, we’ve lots of biases, confirmation biases and seeing things that we want to see, which of course, is rooted in our curiosity. That’s cool. But the AI and machine learning also helps us to have statistics and robust processes for, are these patterns there, aan you prove it? Which of course, coming from a science perspective really matters.
324.5
Dallas Campbell: To say, I have no biases, by the way, I’m completely completely bias free. I always think confirmation bias is very, very easy to see in other people and very hard to see in yourself. You mentioned climate change, obviously, very important, what are the other factors that AI is being used to help society generally or life on Earth, generally?
347.1
Paolo Ruti: We can look at the air quality, air quality as being one of the big issue, I mean, people living in big cities, certainly, and you have a lot of sensors, I mean, nowadays, because of COVID, COVID, there’s a lot of people with CO2 sensors going around and measuring the amount of CO2, we have a lot of data from satellites, on different air quality, PM 2.5, etc. So are we n a way to give the inf rmation that could be rel vant at the street level. And and this is the scale prob em, so you have information on t e large scale, then you need to zoom.
379.6
And this is where this kind of tools they could be extr mely useful.
387.6
Dallas Campbell: It seems to me that kind of in a way, almost all everything we do now in modern life is is somehow and increasingly so reliant on on Earth observation data, I’m trying to imagine what things are going to be like in 10 years time, perhaps what are the new revolutions that are coming up?
405.4
Mark Higgins: I think for me, the thing and Paolo just started alluding to it there, this integration of scales, so my little sensor at home that is measuring something is measuring something that connects to something we can see from satellites, so the sensor based world, we’re going to have more data coming to the same place. But the other thing is really the extension of, we call it cross disciplinary research. So I come from a physics background. I know I know the observation really well. But there’s people out there who know tons about people and systems and how humans live. And actually bringing these two perspectives together in interdisciplinary research and services.
439.1
I think that’s the thing that I’m really excited about. For me.
443.9
Paolo Ruti: I think yesterday they awarded the physics Nobel Prize. And then the main reason was for, for understanding complexity. So I think that’s that’s certainly a big challenge, not only in physics, but also on the way we we measure and the way we use the data, in 10 years time. So are we be able to, to provide a forecast on CO2 emission, which is a big problem for Paris Agreement.
471.3
Dallas Campbell: Actually, perhaps you can just give me a quick rundown, maybe of just, just what are the benefits that we get here on Earth from from the use of AI and Earth observation data.
483.3
Mark Higgins: Integration, benefit number one, the fact that you can bring data together from different sources, and bring it towards a problem. So as Paolo mentioned, so monitoring carbon dioxide potentially in the future, forecasting it, there’s some things where we’ve got established processes already for using the observing data, but they’re really complicated, and AI machine learning we can really speed them up dramatically. So we can get all of the data from the satellites into the models much, much quicker. And that’s an important thing. And also parts of those models can run faster. And that means you can have faster services for people, faster warnings, which then can lead to better action, and also just understanding the whole earth system.
524.2
So, we use the phrase Earth system, but we’re talking about our own life support system. We live here, it’s home.
530.4
Dallas Campbell: And presumably, you know, when we talk about the Earth systems, we’re talking about incredible complexity. And really, the only way we can deal with understanding that level of complexity is using vast amounts of data, which is, you know, being processed and being understood by AI.
547.2
Paolo Ruti: I think we need to understand that AI is driven by our brain. And this is one important, one important aspect. So there’s a lot of understanding on our- from our perspective on the physics of our planet, how storms is moving and why at a certain point theres a lot of rain. And what certainly we need is to use AI to make our understanding much more useful. And we need to I mean, this is to go to the street level, to really give information to the people where they need it. That’s the tricky, important aspects.
589.6
Dallas Campbell: Okay, so here’s a question for you. Why should people do this course, what is exciting about this course? What do you think are the main benefits of a course like this?
600.5
Mark Higgins: So the core sort of thing that really gets me going about it is its openness. So it’s a, it’s an open course, everyone can come. And there’s other bits of openness in there. So the data that we’re showing you and you can get to play with, it’s all open, you can have it, you can play with it. So if you’ve not come across any Earth observing data before, or the kind of Earth observing data we have in this course, you’re gonna have to find out what it is and find out how to get hold of it. But also, we’re going to show you lots of how you can actually do some of this AI machine learning stuff.
627.3
And you can have all that code, you can take that away and do what you want to do with it. But also, we’re going to show you how to- the meaningfulness aspect of it. So it’s not just a throw data into thing, outcomes thing. There is there is thought that sits in there and like Paolo says, the ability to understand the numbers and understand what they mean is also something we’re going to cover quite a lot, this sort of explainability alliance. So those are kind of the main things that got me excited about this course.
658.5
Paolo Ruti: You will have experts on AI and machine learning, you will have experts on air quality, you will have experts on ocean dynamics. So, these different expertise will come together to help people, but if you don’t know anything about machine learning, should you come - Yes, because maybe you want to understand how to use this course in your day to day life, in your business. And so, the showcases will be an important element of this MOOC.
687.9
Mark Higgins: Some of the principles that we’re going to teach, they apply across all of the disciplines angled, so it’s, it’s gonna be a useful thing and it’s gonna be a lot of fun, too.

Welcome to Artificial Intelligence for Earth Monitoring!

In this course you will discover how cutting-edge artificial intelligence and machine learning technologies are helping to advance Earth observation science and the benefits that this will have for citizens.

The course will provide you with an overview of the different types of AI and ML and the fundamental techniques of working with AI algorithms for analysing satellite-based Earth observations and in-situ data. You will also be able to work with hands-on tutorials using Jupyter Notebooks so you can see for yourself how AI works.

The course will cover four thematic areas: oceans, atmosphere, land and climate. You will be able to see how AI and ML are being used in these areas to develop a new generation of Copernicus Earth observation products and services.

The main topic videos are the backbone of this course and you can re-watch them as much as you need. For further context and more detailed explanations, you can also read the introductory text provided with each video, explore the optional further reading links and take an in-depth look at the information about the data, imagery and satellites provided in each topic.

The course videos begin with Topic 1a in Step 1.6. Over the next few steps we have provided more information about the course educators and how to get the most out of the course.

If you talk about the course on social media you can tag your comments #AI4EarthMonitoring

This article is from the free online

Artificial Intelligence (AI) for Earth Monitoring

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education