Photo credit: www.sciencedaily.com

Advancements in Earth Systems Modeling: Enhancing Understanding of Plant Water Dynamics

Earth systems models (ESMs) serve as crucial instruments for analyzing intricate processes that take place throughout the planet, specifically within the interactions between the atmosphere and the biosphere. These models enable researchers and decision-makers to grasp better the complexities surrounding climate change. An effective way to boost the accuracy of such simulations is by integrating more extensive data sets; nevertheless, this endeavor often involves the substantial challenge of compiling millions of individual data points.

In a significant study, researchers from various institutions, including Assistant Professor James Knighton from the University of Connecticut’s Department of Natural Resources and the Environment, Pablo Sanchez-Martinez from the University of Edinburgh, and Leander Anderegg from the University of California, Santa Barbara, have devised a novel approach to circumvent the need for extensive data collection from over 55,000 tree species. This innovation aims to enhance the understanding of how plants influence the global movement of water. Their results were published in Nature Scientific Data.

Plants are indispensable to the functioning of Earth’s ecosystems, playing a vital role in carbon capture and oxygen production for organisms, including humans. According to Knighton, around 60% of rainfall is returned to the atmosphere through the process of transpiration, highlighting the substantial role plants play in the hydrological cycle. Currently, ESMs simplify this intricate process by categorizing all plants in a region as a single unit, known as a Plant Functional Type (PFT).

“We use Plant Functional Types because detailed information about individual species is often lacking,” Knighton explains, emphasizing the challenges involved in mapping vegetation across large geographic areas accurately. “Employing a generic PFT simplifies the modeling process.”

However, reliance on PFTs introduces limitations, as different species exhibit significant variations in their hydrologic properties, such as water movement and absorption. This oversimplification can hinder the models’ effectiveness in forecasting future ecological changes. While initiatives like the TRY Plant Trait Database aim to address this by cataloging plant traits, Knighton notes that these efforts have only successfully documented the characteristics of about 5,000 to 15,000 species over centuries of research.

“With approximately 60,000 to 70,000 tree species globally, we have only scratched the surface of understanding plant traits,” Knighton remarks. “If we depend entirely on traditional methods, it could take us another 2,000 years to gather the necessary data, which may be too late in the context of climate change. Relying solely on field studies is inadequate for achieving timely progress.”

To accelerate this process, Knighton and his team focused on leveraging existing data regarding tree traits, such as growth height, root depth, and internal water flow. They conducted phylogenetic analyses to compare these traits across related species.

“We assessed how trait values are shared among closely related species, positing that evolution would preserve these critical traits,” Knighton explains. “For instance, if deep roots were essential for a plant’s survival, its descendants and relatives would likely share this characteristic.”

The researchers conducted this analysis across various traits and discovered notable conservation patterns throughout the phylogenetic tree, indicating that closely related species generally exceed in similar trait values.

“We mapped the phylogeny of all plant species, allowing us to visualize their relationships and similarities,” Knighton adds.

Utilizing this information, the team can impute trait data for species lacking sufficient measurements, effectively bypassing the need for extensive fieldwork.

“Employing advanced numerical machine-learning techniques, we compiled a database of critical tree traits for 55,000 species,” Knighton states. “This allows for more detailed global modeling while moving beyond the overly simplistic one-size-fits-all approach.” Through this newfound granularity, researchers can explore various plant species in ways previously thought impractical.

While Knighton regards this work as a preliminary approximation, he acknowledges its significance as a foundational step. With time, as researchers gather more precise data, the interpolated information can be refined to enhance accuracy.

This developing research is part of a larger initiative, preceded by a localized proof-of-concept study that established the efficacy of imputing hydrologic traits. The next phase involves comparing the imputed data against observational data collected from various sites, including UConn Forest and other locations across the United States.

Knighton elaborates that data collection is ongoing across ten U.S. sites, serving as case studies. Master’s student Caroline Stanton ’26 is building ecosystem models for each location and calibrating high-resolution models to estimate traits, which will later be evaluated against data compiled over the last two decades. Researchers intend to weigh the accuracy of plant trait estimates against real-world observations from these sites.

Ultimately, the researchers aspire to expand their methodology to forested areas worldwide, striving to discern what drives variability in traits among different plant species. Greater insight into the factors influencing trait diversity has the potential to elevate the precision of ecological models, while also providing critical information about the traits themselves.

Knighton emphasizes the broader implications of their findings, expressing hope that climate modelers will find their data useful. He also aims to enhance overall comprehension of Earth’s systems and the vital ecological functions that plants perform. “Plants exert tremendous control over our environment,” he asserts.

Source
www.sciencedaily.com