An interview with remote sensing expert Pierre Markuse.
May 3rd 2020. Melissa Sterry
Our eyes in the sky, satellites play a vital role in illuminating the workings of wildfires. The data they capture both complex and copious, but for the likes of Pierre Markuse, valuable insights could easily become lost in translation. Here he sheds some light on the field, and his work within it…
Q. Describe yourself...
A. I’m a freelance remote sensing expert - I use that term very loosely, as the field is so big today that nobody can really be an expert in every aspect of it - and a blogger with an interest in optical remote sensing. While I like to look at everything Earth has to offer, I have a special interest in wildfires and how satellite images can be used to visualize them, be it for scientists or the public. How to make satellite images more approachable to the public in general, the scientific background of the field, and the usage of satellite images in journalism are additional fields of interest.
Q. Working in the fast-developing field of satellite imaging and its applications to, amongst other fields, wildfire research, how did you enter this niche, but now expanding profession?
A. That’s a slightly convoluted story. My first experience of fires and disaster relief came in my youth. First, when I trained to become a paramedic at a fire-station, then when I volunteered and later worked for the German Red Cross. Though, in the years that followed, I ended up working in IT, my interests in fires and disaster relief endured. Around 2008/2009 the Landsat archives became freely available, and that’s when I started processing satellite images, initially as a hobby. But, learning as I went, I started to publish the images, mostly for science communication, creating featured collections on the now defunct Google+, reaching almost half a million people with my posts. More recently, I’ve been active on Flickr and Twitter. Grega Milcinski, the CEO of Sinergise, the company behind the Sentinel Hub EO Browser, noticed my work and asked me if I would be interested in an account, and to maybe use the EO Browser to process images and write scripts. I was. It is an impressive tool that allowed me to automatize some of the methods involved in processing images into scripts. I can’t begin to tell you how much time this has saved me since then. Not to mention that even more people started to look at my work. I really like that my scripts and images are used by researchers like you, and I also like to see them used in science communication more generally, and by media for reporting.
Q. One of the most visually striking variances of satellite images of this past few decades is their resolution – with imaging capacity becoming ever more detailed over time. How has this impacted on your work, and its applications?
A. While I’m a big fan of high-resolution data, it always depends on what you want to do. For some, a big archive of data going into the past, like the Landsat archives, might be more useful than hi-res data. Hi-res satellites lack the multitude of frequency bands that satellite systems like Landsat or Sentinel-2 offer. In my own work, high-res data isn’t that important. Sure, sometimes a high-resolution image, on its own or additionally, is useful, especially for wildfire delineation maps. Then again, when used for science communication, hi-res images can be very eye-catching. If a new satellite system with frequency bands like the Sentinel-2 satellites, but with 1m or even 50cm spatial resolution is be developed, it could be interesting to work with. That, together with a daily image-stream - and a lot of remote-sensing people, me included, would be very happy!
Q. There now several ways in which satellite images can illuminate various aspects of wildfire behaviour over space and time, which do you typically work with, and why?
A. Depends on the audience. I like natural-colour images with overlaid IR data to show hot spots. These images tend be very good at getting public interest, and they are easy to understand, meaning they work really well for science communications, and particularly when using social media, where they can link to articles that explain the images’ contents in more detail. Whereas, images that incorporate more NIR/SWIR data, which allows users to see burns scars, the available fuel in a fire’s path and surroundings, as well as hot spots, are typically more useful scientifically. These images can sometimes be further enhanced by making them 3D using a digital elevation model. This can be particularly helpful when predicting fire-behaviour and possible future spread rate and course.
Q. They say that pictures say a thousand words, and that’s certainly true of the many satellite images you’ve processed. Describe some of the organisations and/or researchers that work with your images, and what they do with them.
A. Often published under CC-BY license, which grants use without notice, I don’t always know when someone’s used my images. But, I do know that they have made their way into a few newspapers, journal papers, and theses. Every now and then, I’ll hear from fire-fighters who, having fought fires that have featured in images that I’ve processed, thank me for sharing them, which I find most rewarding. But, I’ve found my images used in lots of different types of print and online media, and they even made their way into a Canadian television series about fire-fighters, together with countless televised news and weather stories, and two or three documentaries that are in the production works!
Q. As someone that works with your images, one of the things that strikes me is how extraordinarily beautiful they are – and in turn, how beautiful our planet is. Do you feel that same sense of awe and wonder from your digitally-enabled beyond-birds-eye-view of the world?
A. Oh yes, I still do, again and again. And, while most of my images might feature wildfires, the most beautiful ones are usually showing the colder regions of Earth. Glaciers can look very majestic in real life, but equally so from space for example. But, there are other places that I like to look at, particularly mountains, rivers, and lakes... so many beautiful places! If I had to pick a few favourite regions, I’d probably say the Pacific Northwest, the Arctic regions and Antarctica, and maybe surprisingly New York City!
Q. Sadly, it’s not just the beauty of our world that’s captured by satellite imagery, but crimes against the environment and humanity. Which particular problems do you think satellite- imaging technology and its applications are well-placed to tackle?
A. Oil spills come to mind, be it at sea or from refineries or pipelines on land, but with more and better atmospheric sensors it will also be easier to measure emissions of industrial areas and make sure treaties to reduce emissions are verifiable. Satellite images can also show human and animal migrations, such as people fleeing from areas during war or famines. Hi-res images have been witnesses of crimes against humanity, showing the likes of illegal prison camps, military attacks, and of course, they are and always have been a method of military reconnaissance. Declassified data from military reconnaissance satellites can also be interesting to take a look at for scientific reasons, such as the impact of military actions over time on ecological systems and towns and cities, or to get images from a time, when remote sensing satellites were not routinely monitoring Earth.
Q. What inspires you to share your work and insights through social media and blogging?
A. I think science communication is vital for educating the public and politicians alike, helping them to make the right, or at least well-informed choices and decisions. The last few years have seen a lot of science scepticism, and in some instances, outright science denialism, especially online. It is important to work against this, using social media to communicate scientific facts to as many people as possible. Because, if used responsibility, social media can do lots of good and push back the tide of mis- information and half-truths!
Q. A pro-active science communicator, you regularly share insight, information, and tools to help others understand the potentialities of satellite imaging, and with that, to learn hands-on visual processing skills. If selecting just 3 of these, which would they be and why?
“Revisit” your work, which could happen during or right after processing an image, or even months after. Frequently, I start processing an image and then something instinctively looks wrong. Maybe it’s the colours, the contrast, or the style of visualization that aren’t working. Sometimes you can correct it. Other times you delete the file and start over with the raw data from the satellite. I lose hours of work because of this, again and again. But, I believe it is necessary to achieve the best results.
Develop an eye for what you want to show, how to best show it, and to whom. Many decisions depend on the intended audience. For example, a visualization that works really well for scientists might not be visually striking enough for the aesthetic tastes of press and media. Vice versa, many images that grab public attention, hence work well for press and media, do so at the expense of the details needed to be useful to scientists. In brief, know your audience.
Keep alert to the work of your peers and wider community. Often, I process a scene to then see that same scene has been processed by somebody else, and I think to myself “that’s cool, I wish I had thought of that!”. On the same note, experiment with your visualizations. When new techniques and approaches work, expand upon them. When they don’t, try something else.
Q. As satellite-imaging has advanced, so too have global communications technologies. As someone who has harnessed the potentialities of both these developments, how have you found the latter influenced your work in the former?
A. Global communication technologies, especially social media, have played a big role in how my work with satellite images has developed over the years. The now defunct Google+ network allowed me to routinely reach a many times larger audience with my images and other science communications than would otherwise have been the case. Now, it’s Flickr and particularly Twitter, that provide me with a platform to both reach a large audience, and to focus their attention on the images and the events they relate to. Twitter also has a very friendly and active remote sensing community, which helps when seeking answers to questions and queries, and knocking ideas around. The feedback from the general public is also helpful, for example as pointing out things that should be addressed in some way to make images easier to be understood.
Q. The year is 2030. Satellite-imaging and related technologies have continued to develop apace. What can you do then that you can’t do now and why?
A. I think that advancements in scripting software will mean that many of the tasks that, today, I still have to perform manually will become automated, in part because I think the software will become increasingly powerful in its performance. A bigger archive of available satellite data and increased processing speeds are likely to make time-series analysis much easier and faster. I think machine learning will help greatly with image data processing, and with that assist in understanding and evaluating it. I also hope that, by then, hi-res data is at a price-point low enough to enable broader use of these new technologies. More and more countries are launching remote sensing satellites and I hope to see the data that many of those satellites gather becoming readily available to both the image processing and research community, and the general public, ideally as open-access data.
All images by Pierre Markuse. Find more examples of his work at his site and on Twitter and Flickr.
