The Power of Scientific Modeling for Monitoring Ecosystems

Crowther Lab – A Resource Watch Partner Profile

Climate change and biodiversity loss are among the two greatest threats facing society – and they are intrinsically linked. Preserving terrestrial ecosystems is of the utmost importance if we want to protect our environment and our communities. Understanding those ecosystems is a key step to informing how to protect and restore them.
At the Crowther Lab (ETH Zurich), we create scientific models – often in the form of spatially explicit global maps – which help give us a new or better understanding of Earth’s ecosystems. These maps not only inform us about current ecological diversity, but we are now also starting to generate an understanding of ecological potential, e.g. how much vegetative carbon could be supported on the planet, at both global and regional scales. Our lab was founded on the belief that it is only with a global perspective of all ecological ecosystems, be they above or below ground, can society define effective targets and strategies for nature-based solutions.
But, every good model and map starts with data. Lots of data. We work with global networks of scientific collaborators (including the Global Forest Biodiversity Initiative and the Global Soil Biodiversity Initiative) to build and analyze large databases of ground truthed observations. These databases currently consist of more than 30 million measurements of biodiversity, where scientists have painstakingly identified tree species and measured them with diameter tapes and height sticks, or collected soil samples using soil cores, collection bags and so on. We also look at databases with information about the physical environment such as precipitation and temperature, and sometimes even satellite imagery. After identifying relationships and patterns between the ecological and environmental variables, we then use machine learning models to extend those patterns across the globe to generate spatially explicit information (e.g. data maps about trees and soil) with a higher level of predictive accuracy.

Global map of soil nematode density at the 30 arcsec (approximately 1 km2) pixel scale, created by the Crowther Lab.

Global map of soil nematode density at the 30 arcsec (approximately 1 km2) pixel scale.
van den Hoogen et al., 2019
We hope these maps will provide the scientific basis to guide restoration and conservation organizations, providing insights as to which parts of the world we could focus on if we are going to address one of the greatest threats to modern society. 
The process of creating data, providing insight, and effecting change can be a complex and lengthy one. An example is an ongoing collaboration with Resource Watch to support Gouritz Cluster Biosphere Reserve (GCBR), a nonprofit dedicated to protecting the reserve’s nature and biodiversity and a partner of ours through founding partner, DOB Ecology. This collaboration, while still in very early stages, is an exciting case study for how our models can be trained, using ground truthed data, to function at a much finer scale. Through this collaboration with GCBR and Resource Watch, we can take global satellite data to the ground in a way that wasn’t possible before.
If we are able to apply scientific models with this level of predictive accuracy, at such a fine scale, and high resolution, it could transform how we plan, monitor and improve restoration projects around the world. 
If you are interested in learning more about Crowther Lab’s work and how we are uniting global models and satellite data with ground truthing, please visit our website and join the network of restoration practitioners we are working with here
Cover photo taken from Flicker: Forest, shared under Creative Commons Attribution-Share Alike 2.0 Generic license and overlaid with the Crowther Lab logo.
Share on facebook
Share on twitter
Share on linkedin