Hayley Evers-King

Hayley Evers-King

Marine Application Expert working to bring data from space to people working with, in, and on our beautiful oceans! Also likes to make satellites out of cake...

Location Darmstadt, Germany


  • Ah ok! For an introduction - I recommend these lectures - https://github.com/jmartinezheras/2018-MachineLearning-Lectures-ESA. We had Jose come and speak at EUMETSAT a while ago and I (as a non-ML but EO expert) found them great!

  • Hi @MasiasiRonaldMushikoma. Adding to what @CarlosFortunyLombraña said - we offer some detailed training courses at EUMETSAT - see https://training.eumetsat.int for details.

  • Not MOOCs but I can recommend checking out the general training opportunities at https://training.eumetsat.int. We use Python a lot in many trainings, at various levels, all focused on EO data.

  • @LiviuIoan I can recommend checking out the general training opportunities at https://training.eumetsat.int. We use Python a lot in many trainings, at various levels, all focused on EO data.

  • The notebooks are available here: https://github.com/wekeo/ai4EM_MOOC :)

  • Hi Martin, They link to the same playlist - the intro video is just the first one of the playlist. You should be able to see various others on the right. Note these are just short videos from previous training events - not directly related to the MOOC. Best wishes, Hayley

  • You get a free trial period with any of the plans for the virtual machines :) How long depends on the size of plan you request (more details here https://www.wekeo.eu/pricing)

  • We'll cover some use of Sentinel-2 for the marine domain in ocean week :)

  • Hi Iago. The WEkEO HDA had some problems this morning but should be working now, please try again. We have updated the WEkEO interface a little since I filmed the walk-through video some time ago, you can see your jobs in the data viewer now.

  • Hi Amit, Please try again, the WEkEO HDA had a problem but should be working fine now.

  • Hi Harlee. Great question! There are various ways you can look at coverage and/or predicted overpasses. It can vary a lot by satellite and sensor (because of the ways they measure, and the tradeoffs needed in missions design). You can find information on coverage for the Sentinel-3 data in our user handbooks (here's an example for S3 OLCI, see coverage maps on...

  • Hi Acelya, I can tell you a bit about Sentinel-3 and how it measures colour! The Sentinel-3 satellites have the OLCI (Ocean Land Colour Instrument) on board - https://www.eumetsat.int/olci. It's a radiometer and it measures light reflected from the Earths surfaces in 21 narrow wavebands (so around 21 different wavelengths), mostly in the visible part of the...

  • Unfortunately for validation of optical data we can't match up with the satellites when the weather is bad :( The data collected can still be useful for algorithm development though.

  • Completely agree Philippe! I think we are starting to get better at it (I couldn't have imagined a MOOC like this when I started my PhD), but we need to keep working and it's something I'm personally trying to champion. There are more and more people from the field on twitter these days - so that's a good platform to hear from the industry.

  • Hi Johnson. It depends on the way you are conducting your satellite validation, and the situation (oceanographically) that you are measuring. For example, in very dynamic coastal regions where properties vary quite rapidly in time and space, we might want stringent criteria (of the order of a few hours). For validation of merged/lower resolution products,...

  • Correct Jenny. Uncertain times at the moment.

  • This is a very salient point, and is the reason why we tend to use multiple sources of data for investigating the carbon cycle. Here's one project I've been involved in recently that used in situ, satellite, and modelled data http://www.oceancarbon.net/

  • We did - beautiful place! I love that they have an underwater trail that you can follow :)

  • I do often wish I could have eyes as sensitive as the radiometers aboard our ocean colour satellites :) Generally we're only good at detecting really drastic changes in the coastal zones.

  • You might all be interested in this hot-off-the-press paper on the topic of collecting ocean data through recreational activities: https://www.frontiersin.org/articles/10.3389/fmars.2017.00351/full

  • There's some examples from a UK perspective here: http://www.spaceforsmartergovernment.uk/defra-eo-coe/

  • Lots of new areas of application - here's a recent project that will bring socio-economists and satellite scientists (as well as many others) together: http://www.pml.ac.uk/News_and_media/News/Supporting_coastal_communities_in_Southeast_Asia

  • Generally yes, they can have big impacts on our activities, but also on near shore species who can't move too far - have a look at the lobster walk outs that occur on the west coast of South Africa. I've also seen some reports of impacts on seals and whales during particularly large blooms.

    The shellfish industry can be substantially impacted as you say - I...

  • GM is correct :)

  • Hi all,

    I'm away in Ghana this week - so my internet connection is not very good - I will get a more detailed reply for you soon!

    Best wishes,


  • Hi Sheila,

    The depth to which an ocean colour satellite can "see" varies, depending on what's in the water. They are passive sensors, so they are receiving light that is reflected back from the ocean. If there are lots of things in the water absorbing or scattering light, the depth to which light from the sun will reach decreases, and as such the depth to...

  • Unfortunately Bilko was only designed for windows...but I did put together an install guide for Mac using wine...you should be able to follow a similar installation process on Ubuntu/Linux. Guidelines are here: http://www.learn-eo.org/software.php

    Alternatively there's a version of this tutorial using the SNAP software that's just been very recently...

  • Not an easy task in a large and complex ocean - but some interesting studies pointing to possible sights based on found wreckage: https://phys.org/news/2017-01-oceanographic-analysis-potential-site-mh370.html

  • Hi Hardeep,

    You might need to try a different browser. For me it works on google chrome.


  • More content on algal blooms coming up in the following weeks :) Always happy to talk about these topics as they are a particular focus of my area of research :)

  • It's all about the algorithms! That said, some are quite simple - empirical functions that use two measurements of light at different wavelengths, where as others can get really complicated (I've worked with some that are attached to full radiative transfer models, and use different models of particles to estimate components of the ocean colour signal).

  • Yep! And it's not always easy, we have fairly good protocols developed for measuring POC for example, but extracting the phytoplankton component of that is a little more difficult.

  • Hi David,

    Since this video was filmed Victor and I have actually published a paper on the topic of POC estimation from space :) It's open access so you can have a look at it here: https://www.frontiersin.org/articles/10.3389/fmars.2017.00251/full


  • On uncertainties in particular - there's been some very interesting work done under OC-CCI project, and also under recent programmes to drive towards quantifying uncertainty using best metrology principles - I'm involved in one project on that and you can see some of the early stage outputs here: http://amt4sentinelfrm.org/International_Workshop There will...

  • Hi Murdoch,

    There is a vast amount of literature around these topics - probably far too much for me to summarise in 1200 characters ;)

    Thankfully much of it these days is open access - you might like to read some of the articles in a recent special issue of the Frontiers journal, which features a lot of the work presented at a recent Colour and Light in...

  • Hi GM,

    I'll try and boot up Bilko on my laptop and see if I can double check this for you over the next few days.


  • Lots of data and plots for the time series here: http://www.westernchannelobservatory.org.uk

  • Agreed Murdoch, though I will say it wasn't the most interesting day oceanographically on Plymouth Sound when we filmed :) If you'd like to see some data in more detail from this area, check out the Western Channel Observatory (http://www.westernchannelobservatory.org.uk/) they have data in real time from the buoys which Quest (in the video) goes out every...

  • I think it's meant to be ocean, but this is actually the first time I noticed it ;) That said, these things do influence the atmosphere too - coccolithophores play a really key role in the carbon cycle, particularly in the carbonate equilibria system which regulates how much carbon the ocean can absorb from the atmosphere, and the acidity of the ocean too....

  • So glad you love Ocean Colour Kenneth - I too am constantly astounded by the beauty in this topic! It's a pleasure to share that with you :)

  • Hi Patricia,

    Yes definitely - there are things satellites can't easily see or estimate that we can measure in situ. But we can often relate the measurements each one can make, to the other to get better insight in to spatial and temporal variability. e.g. satellites can give us an idea of the temperature/productivity of an area, which we might relate to in...

  • Hi David,

    We have a team here at PML that works on aircraft data (part of http://arsf.nerc.ac.uk/index.asp) - incredible data but as you say quite expensive to collect!

    One project I'm working on is looking at drone mounted sensors to support our validation efforts - all this new technology is really cool :)


  • Hi Veloisa,

    Often we won't be able to use data in those conditions - it can be a particular problem at high latitudes in winter.


  • For some current validation activities you might want to check out the blog from the AMT cruise that Gavin (in the following video!) is currently working on http://blogs.esa.int/campaignearth/2017/10/24/amt4sentinelfrm-comparing-measurements-with-measurements/

    This is part of the AMT4SentinelFRM project (which I'm currently sat in front of my computer...

  • For some current validation activities you might want to check out the blog from the AMT cruise that Gavin (in this video!) is currently working on http://blogs.esa.int/campaignearth/2017/10/24/amt4sentinelfrm-comparing-measurements-with-measurements/

  • ...make the same journey many times. To build up a time series.

    It's tricky work validation, but really important.


  • Hi Patricia,

    We know the orbital parameters of the satellite very well, so we can predict in advance when a satellite will pass over head to plan our field sampling. Here's one tool that's available to do that (https://earth.esa.int/web/guest/-/esov-software-tools-esov-ng-and-esov-classic-1652).

    We have a bit of a time window, depending on the parameter...

  • We'll still be members of ESA and EUMETSAT (that can happen outside of EU membership), but it would be very sad if we couldn't contribute to Copernicus. All a bit uncertain at the moment.

  • I've discussed this a fair bit on some other comments throughout the course, but in short - we don't know, it depends what the UK government negotiates. Potentially we can't be involved in programme development and service delivery anymore, which would be a terrible shame. We'll still have access to the data, because that's free and open to everyone. UK...

  • Or with the in situ data! I'm currently working on a project where we intercompare different radiometers, and there are many uncertainties to take in to account when making these measurements too - sea conditions, calibration, shadowing, deploymeny angles, sky reflectance...all just a few of the things that can influence different radiometer systems.

  • I don't know much about the field, but I can tell you getting ethics approval for these sorts of things is pretty complex and in depth :)

  • I love visiting the control room at EUMETSAT :)

  • Hi G M,

    An Argo float has a typical life span of 4-5 years. In that time it cycles between the deep ocean and the surface every 10 days or so. In the surface ocean it is likely to encounter stronger currents, in the depths, a much more quiescent (gentle!) regime (at least away from boundary currents). Consequently, the floats often do not travel as far as...

  • Hi Raquel,

    The task is asking you to have a think about how climate change might affect the area where you live (or you may like to think about an area related to your work or travel!). From that we'd then like you to think about how the different climate data sets derived from satellites (see for example the ESA CCI project webpage) might help you to...

  • Hi Johnson,

    I think Dr Mark Higgins discusses this is in the week 1 feedback video: https://www.youtube.com/watch?v=aULgs38MiKs

    Hopefully this will help answer your question.


  • Hi Fi,

    I've discussed this a couple of times across different sections of the course - so you might find more information in some of the other comment sections too. But in short: Copernicus data (including Sentinel 3 land (via ESA) and marine (via EUMETSAT)) is available free to everyone around the world. So we (I'm British) will still have access to the...

  • Hi Jay,

    This is actually not the easiest question to answer - depending on which data and downstream products you consider - but to give you some context Sentinel 1 and 3a/b should be producing around 1.6 TB a day, and Sentinel 2 a/b somewhere around 400TB per year - I believe that is just the raw data, not including the downstream levels and derived...

  • Hi David,

    This is an interesting one, as I can see that this topic would fit under two different streams of Copernicus - land and ocean related measurements. In particular I would imagine that this work would need the high resolution imagery provided by the Sentinel 2 MSI sensor for example (this is higher spatial resolution (10m pixels, only around the...

  • Hi Scott,

    You can see some other replies I made on this topic below. But in short: we should still be able to access data, but whether we will continue to take part in the development of the programme and provision of services around it depends on what the UK government decides. There's a lot of uncertainty but I hope we can continue to be involved in such...

  • Hi Anthony,

    On the poles question - they can go over the poles, but are usually offset such as to keep a consistent orbit such as a sun synchronous orbit - where the satellite passes over the same/very similar points and the same/very similar time.

    Re. Altimeters - in the full names of the instruments you'll see more specific naming i.e SRAL aboard...

  • I'm not sure which browsers you all use, but on mine (chrome) you can highlight words, right click and you have an option to "look up" which then has several other options for finding information out about a term (dictionary/wikipedia), and an option to google it as well. This might help, although I'm aware that the vast array of these acronyms might not be...

  • Hi GM,

    Much of the water extracted from aquifers ultimately ends up in the oceans. It's a relatively small contribution comparatively though - see more information here: https://phys.org/news/2016-05-groundwater-contribute-sea.html


  • ....chlorophyll concentrations.

    Level 3 is usually when images begin to be aggregated together in some way e.g. to create composites in space/time. This is beneficial if you need to work with data with fewer gaps, or if you are looking at data over longer time scales. You may also find at this level that multiple sensors are involved, creating merged...

  • Hi both.

    I agree this terminology can be quite confusing, it took me a while to get my head around it. You'll often see 4 levels of data referred to when it comes to satellites. What is measured, and what is done between the levels varies, depending on the individual sensor. But in general Level 0 will be the very rawest of data in the form of measurement...

  • Very true Lee - you'll see more on validation in subsequent weeks. I can tell you - collecting in situ data is pretty tricky too though ;)

  • Correct - the combination of wind stress and coriolis results in upwelling - you might want to look more in to Ekman transport if you want to learn a bit more about the underlying physics :)

  • Both you and Lee are a little ahead of us here - you'll find out more about phytoplankton preferences in the coming weeks, but I'll give you a bit of a preview and tell you that cold water typically comes from upwellings (as Lee above suggests), and upwelled waters come from depth where they've usually been sat gathering nutrients like ocean compost heaps for...

  • It's definitely hard work Anthony you're right! There is a lot of work that's gone in to putting the EO community in to a position to generate long term, climate quality data sets for the ocean. You can see lots of the work coming to fruition in programmes like Copernicus now - the data from EUMETSAT feeds in to the products available through the Copernicus...

  • Hi Lee,

    Yes you're right - in fact many satellite measurements are only representative of a surface picture (ocean colour is the same). To get a depth perspective we rely on other methods - in situ data (from ship based profiles, ARGO floats, gliders etc) and modelling, but the satellites still give a great synoptic surface view, and with a bit of...

  • Hi Lee,

    I did reply to Murdoch about this, but happy to reply to you too. As you can imagine it's an issue at the forefront of many scientists minds who are involved in the programme. In terms of data access, because it's open to everyone globally, we'll still have that. However we don't currently know what the UK government will negotiate in terms of our...

  • Great to see another Plymouthian involved :)

  • I can watch animations of the data forever, so I definitely understand :)

  • I'm pretty sure I can't remember them all, even after 7 years in the field!

  • Hi Kenneth,

    You'll definitely find some helpful content in the course and I'd always be happy to talk more about these issues - we do quite a bit of water quality research at PML and are always keen to collaborate.

    Enjoy the course!


  • Hi Murdoch,

    Access to the data is free for everyone, but as for involvement in the programme, we just don't know at the moment. Copernicus was explicitly mentioned in the recent government paper on Brexit and science, and of course I really hope we can continue to be involved as much as possible. It's been so exciting being involved in Copernicus at the...

  • Also an avid scuba diver myself! Hope you enjoy the course John! :)

  • Hi Murdoch,

    Maybe I can help!

    OC is sometimes used as shorthand for "ocean colour" a type of data we'll look at a lot more later in the course.

    MOOC stands for Massive Open Online Course - courses such as this one held online, open to all, and with many participants.

    EUMETSAT is the European organisation for the exploitation of meteorological...

  • Good morning everyone! Looking forward to chatting with lots of new people, and catching up with a few returning participants from last year! Lots of exciting new things to discuss in the ocean satellite world :)

  • Integrating measurements makes them so much more powerful. We couldn't possibly understand satellite data as clearly as we are beginning to without in situ data. And models allow us to experiment with our understanding of complex systems and provide tools that can be really useful for prediction. The more we can bring these techniques together, the better we...

  • Very true Richard, though I think economics in this sense is moving towards considering a variable environment as a key factor in sustainable resource use. Climate change (and it's potential impacts) has come up several times in our interviews with aquaculture stakeholders in South Africa over the last few weeks - satellite data and models can really help...

  • I'd love to update this video based on what Eleni and I have learnt during our trip to South Africa the last two weeks - particularly with regards to how we balance the requirements of different industries, the challenges of the environment itself, and try to help as many people's livelihoods as possible. It's been a real eye opener for me (as somewhat of a...

  • It's been a pleasure Richard - great fun fitting in answering questions via my mobile amongst jumping on to oyster harvesting boats, hosting a workshop, and learning about the economics of abalone! This platform has allowed me to learn a great deal about everyones varied perspectives on Earth Observation and the marine environment. Thank you very much for...

  • Hi Anthony,

    It's worth keeping an eye on this site: https://earth.esa.int/web/guest/pi-community/training for ESA courses.

    For ocean colour, other ones I can think of which come up yearly are the IOCCG lecture series (http://www.ioccg.org/training/SLS_2016.html - also other ocean colour opportunities advertised through there - sign up for the mailing...

  • It was Ivor - I also remember visiting there when I was younger :) Glad to see some Plymouth locals taking part in the course :)

  • Sentinel 2b launch date is currently march next year, Sentinel 3b is planned for next year as well, but later in the year. You can see latest on Sentinel 2b here, getting packed up to head off for launch prep: http://www.esa.int/Our_Activities/Observing_the_Earth/Copernicus/Sentinel-2/Farewell_to_Sentinel-2B

  • I believe the paper has recently been submitted, but here's a poster from my colleague (TJ - who presents the ocean colour and models, and citizen science work on this course!) http://www.eposters.net/pdfs/utilising-optical-water-classification-for-pixel-by-pixel-assignment-of-uncertainties-in-a-merged.pdf

  • You might find the STEP forum useful, if a little technical - http://forum.step.esa.int/

    There are also specific groups related to software/sensors , and alumni from some courses (IOCCG has one on facebook). But I find the best way to keep in touch is through twitter - lots of individual scientists as well as organisation on there.

  • That's very kind Richard, would love to know what you think of it - the paper was a labour of love that came out of my PhD :)

  • Hi Siham,

    It's important to consider the angular components when making radiometric measurement. You'll probably find this paper I shared below quite interesting: http://www.jeos.org/index.php/jeos_rp/article/viewFile/13058/1064. It covers some discussion of this.


  • Hi Siham,

    That's one of the simpler and more long standing ways yes, but there are other ways that have been developed too. The OC-CCI project uses 14 water classes, based on the spectra for example, these are used for error characterisation. You could also put ranges on the absorption/scattering/chl etc - I've used broad categories like those to look at...

  • Hi Nick,

    You can alter the 'viewing geometry' of the instrument but turning it in relation to the sun. This serves several purposes as there are angular components that need to be considered when making radiometric measurements (by satellite or in situ).


  • Hi Ann,

    Apparently there are few to none farmed in Knysna now. We visited an area where there are several farms in Saldanha. Luckily no blooms at the moment, but it was quite rough when we went out.


  • Hi Mick,

    Yes there are various booms which launch instruments (including the CTD as you'll see in the later videos with Gavin) - so the hardhats, steel toe caps, and life jackets are required :)


  • Hi Phil,

    Thanks for your insight, I'll admit it's not my area of expertise, particularly in the UK context as much of research has taken place in South Africa. I didn't realise the area for no take was so low!


  • Hi Richard,

    Interesting you mention that - I did a lot of work on optimisation and the ocean colour problem during my PhD. Local vs global minima is indeed an issue, and the ocean colour problem can be very ambiguous, depending on the parameters you're trying to estimate, and in what watertype. We were looking particularly at cell size, but you may find the...

  • Hi Edwin,

    Indeed the measurements are relatively sparse, though they are getting better. One of the things we do is try and characterise regions by their optical properties so that we can use highly sampled regions to inform our understanding of other regions that have similar characteristics.


  • Hi Balasaheb,

    Just one example, but I visited an aquaculture farm in South Africa a few days ago, that is developing their own wave energy platform to make their operations both more sustainable and more affordable. I agree the upfront costs can be expensive, but hopefully this new technology can spill over and be available for further use.


  • Hi Christopher,

    You might be interested in the community seagrass initiative (http://www.csi-seagrass.co.uk/about/), or perhaps taking a secchi disk along with you, next time you do a boat dive (http://www.secchidisk.org/).


  • Hi Paul,

    Here is the recent paper that Tom and colleagues published on the surfing project (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0127706). The project was exploratory and funded internally at the lab, but the team are currently applying to various sources for external funding to increase the scope of the work.


  • Hi Hally,

    Good question! This is part of several methods to avoid uncertainties in our observations caused by glint from the sun, shadows from the ship, and other factors. There's a detailed paper on this here if you'd like to read further (http://www.jeos.org/index.php/jeos_rp/article/viewFile/13058/1064).