The "Smile Curve" of evergreen trees

Our journey began with a simple conversation about the "smile curve," a notable pattern that describes the distribution of evergreen trees. When you examine their global presence, these trees are primarily clustered near the poles and the equator. This distribution creates an arc along latitude, reminiscent of a smile across the Earth. Naturally, we were intrigued and eager to explore this phenomenon further.

Unravelling the secrets of forests

Equipped with an extensive collection of forest inventory datasets, I was keen to explore the various types of leaves and their impact on our environment. We began with a simple concept and started mapping the global distribution of two major tree types: evergreens, which retain their leaves throughout the year, and deciduous trees, which shed their leaves during less favorable seasons. The project was both fun and illuminating.

Yet, as I delved deeper into the environmental factors influencing leaf habit distribution, I encountered substantial uncertainties. Evergreen trees, in particular, presented a puzzle, with broadleaved and needle-leaved trees exhibiting distinct adaptations and evolutionary histories shaped by past climate variations. This challenge drove me to further differentiate leaf types by introducing an additional leaf trait – leaf form. This added dimension allowed me to distinguish evergreen trees thriving in warm and humid environments from their needle-leaved counterparts in colder regions.

Navigating Through Complexity

However, a new obstacle emerged. When dealing with just two types of leaves, modelling was straightforward. But with a total of four leaf types that can co-occur within a forest, classical modeling frameworks no longer sufficed. After being stuck for a while, Dr. Dan Maynard joined the project. With his expertise in statistics, we came up with a new analysis, a four-element array that encapsulated proportions of broadleaved evergreen, broadleaved deciduous, needle-leaved evergreen, and needle-leaved deciduous treess. This new approach let us use machine learning to classify tree type proportions around the world. This did not only allow us to characterize tree types; it helped us understand why certain trees grow where they do.

Thanks to a previous study¹ by my supervisor, Tom Crowther, we were even able to estimate how many trees of each type there are worldwide. The results were surprising, showing that evergreen needleleaf trees are the most abundant trees on Earth. Of the three trillion trees existing worldwide, 38% are needle-leaved evergreen, 29% are broadleaved evergreen, 27% are broadleaved deciduous, and only 5% are needle-leaved deciduous.

But another interesting detail was missing: By combing our results with biomass data, we could show that even though there aren't as many broadleaved evergreen trees, they make up more than half of the tree biomass worldwide. This concentration of biomass within broadleaved evergreen trees underscores their pivotal role, particularly in the context of pan-tropical forest restorations. These regions face vulnerabilities posed by climate change and human-induced land use changes. Recognizing the significance of broadleaved evergreen trees in these landscapes is pivotal for effective conservation and restoration efforts.

The road ahead presented new challenges, particularly when we aimed to predict future forest leaf types at each pixel globally. We soon realized that relying solely on future climate data was insufficient. Factors such as atmospheric CO2 levels and species' demographic dynamics played pivotal roles in accelerating or decelerating the process of leaf type change. Despite these complexities, we could identify regions where future climates would no longer sustain present leaf types. Leveraging my machine learning model outputs, I could identify potential hotspots of change. If these climate projections are realized, trees in these areas will be at risk. They will either have to adapt to more demanding environmental conditions or shift their geographic distributions, causing changes in forest composition.

Imagine that trees in Zurich would experience a climate like Milan in Italy in the near future². The consequences would be profound, with a surge in tree mortality due to heightened drought stress and intensified transpiration. Trees in Zurich, accustomed to a different climate, would struggle to retain water, leading to a significant ecological upheaval. In the grander scheme, these insights have the potential to guide informed decisions that shape global initiatives for the preservation, rejuvenation and sustainable management of forest ecosystems, which is essential for the wellbeing of all life forms on our planet.

My dive into the world of plant ecology

My journey shows how combining a simple idea with modern technology and a wealth of data can give us important new insights. Rather than just relying on satellite images, we took a hands-on approach, collecting real data from the ground. This approach matters because the kind of leaf a tree has can reveal a lot – from how it consumes water to its role in our ecosystem. By getting a clearer picture of our trees, we're better equipped to make informed decisions about conserving our forests. Being in this field now is thrilling, with endless possibilities and a deeper understanding of our natural world just around the corner.

References

1. Crowther, T. W. et al. Mapping tree density at a global scale. Nature 525, 201–205 (2015).
2. Bastin, J.-F. et al. Understanding climate change from a global analysis of city analogues. PLoS One 14, e0217592 (2019).