Toward spectrally quantifying biodiversity’s effects on ecosystem function across planet Earth

Motivated by the goal of remotely assessing biodiversity and ecosystem function across planet Earth, our study takes an important initial step: showing how remotely-sensed spectral reflectance can quantify the effect of biodiversity on aboveground biomass in a young experimental forest.
Published in Ecology & Evolution
Toward spectrally quantifying biodiversity’s effects on ecosystem function across planet Earth

Share this post

Choose a social network to share with, or copy the shortened URL to share elsewhere

This is a representation of how your post may appear on social media. The actual post will vary between social networks

While scientists have shown through hundreds of experiments that biodiversity enhances ecosystem biomass and productivity, scaling up to large spatial extents and naturally assembled communities—to really understand the importance of biodiversity for ecosystem function across the globe—is the next frontier.

Remotely sensed spectral reflectance represents an important tool within the suite of possible Earth Observations to help manage our planet. The idea is that sensors mounted on planes (such as NASA’s AVIRIS-NG used in the US National Ecological Observatory Network, NEON) or satellites (such as the forthcoming NASA Surface Biology and Geology mission or ESA’s CHIME mission) capture in detail the photons that are reflected off Earth’s surfaces. These sensors have the potential to enable us to see the state of our planet—such as the biodiversity and functioning of ecosystems—in real time and across the entire globe, including in places that are otherwise hard to access. 

If we can integrate remotely sensed information with a careful understanding of biodiversity on the ground to accurately interpret the signals they provide, we have the keys to monitoring biodiversity and its consequences. This capability would enhance the toolkit available to target, evaluate and adapt our actions to restore and maintain a habitable planet for humanity. 

As ecologists begin to embrace and explore the potential of emerging remote sensing technologies—and as these data become more accessible—it feels like we are at the dawn of a new era. But, along with the possibilities of these technologies comes a major caveat: remotely sensed signals are only as useful as our ability to interpret them

Our study grew from a NASA-NSF Dimensions of Biodiversity grant led by Jeannine Cavender-Bares, one of my PhD advisors. The objective of the grant was to address just that: to investigate how remotely sensed optical diversity might be used to predict ecosystem processes.

Moving from black and white to full colour vision by adding just three colour bands—red, green and blue—markedly increases the information we can extract from a picture. Imaging spectroscopy captures the reflectance of light across hundreds bands, including the section of the electromagnetic spectrum visible to humans (400-700 nm) and far beyond—in our case, sensors captured spectral reflectance in 432 bands from 400-2500 nm. (Idea from Jose Eduardo Meireles)

This grant brought together a formidable team and set of resources. At the time, I was in the middle of my thesis work, narrowing in on how plants interact with each other to understand why mixing species often enhances productivity. This work focused on a network of tree diversity experiments, and this kind of experiment seemed like a fitting place to test whether we could spectrally detect diversity–ecosystem function relationships. The IDENT-Cloquet experiment was reaching an appropriate developmental stage and was added to the set of study sites along with experimental grasslands and natural areas

This study was conducted at an IDENT site in Cloquet, Minnesota. Experimental systems like these are a useful starting point. In this experiment, biodiversity was manipulated while potentially confounding environmental variation was minimised—this allows us to assert causality—and the effects of biodiversity on ecosystem functions can be accurately quantified on the ground. 

Initially, I helped acquaint others with the experiment during a big team fieldwork campaign. A couple of years later, I took the lead in piecing together this study as part of my postdoc work. This was a close collaboration throughout. For instance, Peter Reich, Christian Messier and Artur Stefanski designed, implemented and continue to maintain the tree diversity experiment and core field-based data collections, while Zhihui Wang, John Couture and Phil Townsend shared many of the spectral analysis techniques. One of the most exciting and intellectually engaging parts of the work was working through ideas on a regular basis with Jeannine Cavender-Bares. In these conversations, I would often come with questions or problems I had encountered, we would brainstorm ideas and Jeannine would work through possible solutions on paper—all to help us figure out what analyses could be done to test whether spectra from airborne imagery could detect overyielding.

Two NASA AVIRIS-NG flights over the study site acquired the spectral images used in this study. Here is an aerial image (from Google Maps) showing part of the study site (left) alongside a false-coloured spectroscopic image (AVIRIS-NG) (right) showing just three of the 432 spectral bands. Imaging spectroscopy captures reflectance from the forest canopy. This reflectance signal is affected by the architecture and chemistry of the canopy—including the arrangement of branches and leaves, tissue nitrogen and water content, and species identity.

With remote spectroscopic imaging, we found that we could accurately detect variation among our study plots in tree biomass and detect how much more (or less) trees grew when mixed with other species rather than in monoculture. These findings form the core of the paper. But, building on my thesis work, we were also interested in whether we could extract information from spectra about how biodiversity affects ecosystem function. We drew upon the ability to spectrally detect nitrogen as well as partition spectral differences using simulations. An important conclusion from our work is that, as biological interactions shape the structure and composition of the canopy, spectra capture the outcomes of these biological interactions. 

Our study demonstrates some ecological insights that may be extracted from spectral reflectance. But, there is so much more to explore and discover. Much of the learning and team building from this project fed in to the development of a new NSF Biology Integration Institute, especially Themes 3 and 4 of the Institute. The synergy of this new research group was instrumental in finalising our study and developing some future research directions. 

Right now, the post-2020 Global Biodiversity Framework is being developed and, recently, the 2020 UN Summit on Biodiversity again highlighted the need to recognise the state of planet Earth. The stakes could not be higher: human existence depends on the natural world. The biodiversity and climate crises facing our world need to be addressed in a coordinated fashion. Remote sensing technologies such as imaging spectroscopy—which could allow rapid and accurate assessments of the state of the planet—are shaping up to serve as a crucial tool to help.

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in

Follow the Topic

Life Sciences > Biological Sciences > Ecology