Microwaving Trees? My PhD Plan.

Covid-19, for me as a PhD student, has meant a few things going out the window, but it has given me the time to write up my plan for measuring trees from space: I'm going to use microwaves.

Out the window went two weeks of field assistance for laser scanning in a Swiss Alpine Forest (bet there would have been less flies there). Out of the window went plans for an interactive virtual reality exhibit at Edinburgh Science Festival. Out of the window went plans for FODEX to get back to Gabon and Peru this summer to collect more field data.

Also out the window, however, went my gaze as I began to think hard about what my PhD was really going to be about. Handily enough, my "home office" window looks out at some trees from the third floor, making it somewhat like being up in the canopy of a small forest. I could look out and try to imagine how light of different kinds would dance around between the trunks, leaves and branches, and how this elaborate interaction might tell us about changes in forest biomass.

Above ground biomass (i.e. the mass of vegetation contained in plants and trees) is the crucial thing we want to measure changes in. Why? That biomass contains carbon. If biomass is lost, then some of that carbon has moved elsewhere - most likely it will eventually end up in the atmosphere as carbon dioxide. Equally, if biomass is gained, the carbon drawn in by the trees will have come from the atmosphere. So, by measuring biomass change, we can help to understand the flow of carbon between the world's forests and the atmosphere. In the end, that will help to improve predictions of climate change.

But we can't exactly put forests on a set of scales and weigh them to find out biomass. So, my approach is instead to focus mostly on measuring the volume of wood, and to multiply that by an estimate of wood density to get biomass. Mass is density times volume. I knew that physics degree would come in handy one day.

Okay then, you might ask how exactly I'm going to measure the volume of wood in a forest from space. Is there some way of seeing the structure of every branch from 500 km up? Um, no, not yet. The only way to truly do that is use laser scanning on the ground, to get super accurate 3D models like this one:

Terrestrial Laser Scanning (TLS) data of two trees in Gabon- acquired and processed by Andy Burt at UCL.

So here is my idea. If we can use this laser scanning data to measure wood volume, it will also help to derive relationships between wood volume and forest volume i.e. estimate what proportion of the space between the top of the canopy and the ground is made up of wood. Of course this will vary a lot between different forests, but seeing as we have data from two different continents, it should be possible to account for some of this variation. 

And a change in forest volume is something we might be able to detect from space. This is where the microwaves come in.

A microwave oven heats up your soup for lunch using radiation with a wavelength of around 12 cm (it works by precisely choosing a wavelength that is absorbed by water). It might be hard to imagine, but it is perfectly possible to ping a pulse of similar microwave radiation at the Earth from space, and to detect its return when it bounces back. That's basically all radar is: a powerful microwave generator that "listens" carefully for echoes.

In particular, I'm interested in a technique called interferometric radar. The idea here is to use two radar satellites to obtain 3D information about where the echoes are coming from. This is kind of for the same reason we have two ears - by comparing the signals arriving at each ear, we get a much better idea of where a sound is coming from.

Using this technique, and by using a wavelength of microwaves (around 3 cm) that is reflected by the forest canopy, it is therefore possible to make a surface map of the forest canopy. Now, what happens if, say, a tree is cut down? If we come back and remap the forest canopy surface, there will be a difference. There will be a "space" between the two canopy surfaces where the felled tree used to be. And that "space" is, of course, the change in forest volume! 

So, by using interferometric radar to estimate forest volume changes, and by relating these changes in forest volume to changes in wood volume through the FODEX field data, and by multiplying this by regional wood density... maybe (just maybe) it will be possible to map biomass change.

That's my PhD plan.