My name is Atticus Stovall and I’m a Postdoctoral Fellow at NASA Goddard Space Flight Center in the Biospheric Sciences Laboratory where I use 3D data to study the ecology of trees and forests from around the world. Trees are one of our most valuable resources for reducing the global consequences of climate change. I dedicate my time at NASA to understanding forests in 3D to improve our predictions of how they will respond to and affect climate change in the future.
My introduction to 3D forests
I’ve been working with forests in 3D since 2012—the beginning of my graduate schooling at the University of Virginia Department of Environmental Sciences with Dr. Hank Shugart. It was here where I started working with a new ground-based LiDAR (Light Detection And Ranging) instrument—a terrestrial laser scanner (or TLS)—to improve estimates of carbon in trees, setting the course for much of what I work on now. TLS collects millions to billions of 3D measurements, capturing the structure of forests with millimeter detail. To clarify, this means our group visits different forests around the world—hiking through dense leaves and difficult terrain—to collect these invaluable 3D data. From these data, we create truly unprecedented 3D models of trees that are revolutionizing our understanding of forests at a global scale.
Why do we measure forests in 3D?
Fine-scale TLS measurements like these are dramatically impacting how much carbon we think forests are holding and how much we believe will be gained/lost in the future. Global forest carbon maps are built from millions of on-the-ground tree measurements—trunk diameter and height allow us to estimate how much carbon a single tree holds. I say “estimate” because carbon is virtually never measured—the process of accurately measuring carbon requires us to cut, dry, and weigh trees (about half of the dry weight of trees is carbon). Instead of weighing every tree in the forest, we relate a small number of observations of carbon to tree diameter (and height), allowing us to predict tree carbon from this set of simple measurements and estimate total forest carbon. Since we so rarely measure carbon, we believe our predictions may not be correct in all forests, so we need a better way to get at these numbers. TLS allows us to measure the entire structure of a tree in 3D—from the trunk to the leaves—enabling extremely precise carbon estimates that are improving global-scale maps of forest carbon.
Bringing ground-based 3D data to the entire Earth
Now, NASA and the University of Maryland have a new LiDAR mission aboard the International Space Station—the Global Ecosystem Dynamics Investigation (GEDI; pronounced Jedi)—measuring forests, for the first time, in 3D at a global scale. GEDI is now capturing a “first-of-its kind” high resolution 3D picture of forests and carbon storage. The global mapping from GEDI relies on the same millions of tree measurements on the ground, but with TLS we are precisely calibrating NASA’s LiDAR instrument for a state-of-the-art carbon map.
To summarize, we visit forests around the world to collect high-resolution 3D LiDAR data on the ground with TLS, create detailed models of trees and leaves in 3D, and ultimately relate these measurements to GEDI.
This process is not without its challenges. Along the way we are often visiting isolated forests, hiking through tough terrain with heavy equipment. Analysis of these data is still experimental and under iterative improvements, so we are constantly finding better ways to capture certain elements of the forest. Now that GEDI data is available globally we are actively researching the best way to connect our ground-based 3D data to those collected from space.
Visualizing GEDI in 3D
To create this visualization, I first downloaded all available GEDI data (~2 months of data, at the time), where I started with more than 1 billion measurements globally or ~500 Gb of GEDI files. After filtering out many of the lower quality data points, I was left with ~600 million height measurements. I incorporated full color into the global model by sampling RGB values from the NASA Blue Marble data. To create the visual with amplified height values, I used some simple trigonometry to convert the latitude and longitude values into a cartesian coordinate space, with the assumption that the Earth has a radius of 1000 m (it doesn’t!) for visual effect. The resulting GEDI data is in a simple ASCII file format and can be loaded into open 3D software, like CloudCompare.
I submitted my entry on the very last day of the contest and was honored to win! I decided Sketchfab was the best platform to share this 3D model in an interactive way:
Soon after, the model was downloaded from Sketchfab and used for the August cover of Trends in Ecology and Evolution (exciting!).
One of the greatest advantages of using Sketchfab is that it provides an interactive way of showcasing datasets such as those collected with TLS and GEDI. For instance, most people curious about GEDI will not download the entire catalog of data, but only those data relevant to their select area of interest. With Sketchfab people can freely explore these data in 3D without downloading anything, making exploration much more accessible. One of my favorite features is the ability to set up a virtual tour within the 3D data. This approach to data exploration is really changing the way we democratize learning for greater technological understanding, especially in 3D remote sensing fields.
What’s next for 3D forests?
Really, we are only beginning to fully understand what we can learn from 3D data in forests. For the past 20 years we have been approaching 3D data with a 2D mindset. We are developing more advanced means of processing and interrogating 3D ecological data. 3D in ecology is providing a more detailed understanding of forest processes—the movement of water through trees or the spread of fire through the canopy or the loss and gain of forest carbon over different timescales. Capturing these processes in 3D will, for the first time, provide a realistic picture of these ecosystems that are so critical for the future wellbeing of the planet.
With this more complete picture of our Earth, we can make the most informed decisions to protect, conserve, and understand the impact of our actions on the state of forest ecosystems moving forward.