Compiling a puzzle
Researchers study water cycle in the tropics to help refine climate model
By Kathy Wythe
Global climate models could be compared to giant jigsaw puzzles. It takes hundreds of pieces — or components — to make up the whole puzzle or global model.
A group of Texas A&M University scientists has just completed refining one small piece of the puzzle for a particular model, the Community Land Model (CLM), which feeds into a model of the global climate, the Community Earth System Model (CESM).
Through three years of research funded by the U.S. Department of Energy, Drs. Gretchen Miller and Tony Cahill, associate professors in Texas A&M’s Zachry Department of Civil Engineering, and Dr. Georgianne Moore, associate professor in Texas A&M’s Department of Ecosystem Science and Management, and their graduate students studied evapotranspiration processes in the tropical forests near Texas A&M’s Soltis Center for Research and Education in Costa Rica.
Particularly, Moore said, they wanted to improve the modeling of fluxes or flow of water vapor and carbon dioxide to and from tropical forests with a particular focus on transitions between wet and dry canopy conditions.
Having a better understanding of the climate processes in the tropics is important, Miller said, because weather and other large-scale global processes in the models are heavily influenced by what goes on in the tropics.
“The Department of Energy knew there was a disconnect in the models in what was happening on the land in the tropics and what was happening in the atmosphere,” Miller said. “For our project, we were looking at one of the disconnects: evapotranspiration.”
Understanding the disconnect
Moore said the inadequate representation of evapotranspiration may partly explain why global climate models do not match observed precipitation patterns.
“The questions we had centered on how we could better inform the processes within these models,” Miller said. “For example, rainfall patterns in the tropics can really influence how wet the forest stays.”
The area around the Soltis Center receives a lot of rain, averaging 4,200 millimeters, or 165 inches, a year, about four times more than College Station averages.
“The Department of Energy knew there was a disconnect in the models in what was happening on the land in the tropics and what was happening in the atmosphere. For our project, we were looking at one of the disconnects: evapotranspiration.”
“We wanted to know how the water cycle changed because of all that rain,” Miller said, specifically how much water was flowing to the atmosphere through plant transpiration versus falling to the ground and then evaporating into the atmosphere.
Miller said one of the reasons they were interested in evapotranspiration is to better understand carbon uptake within the tropics.
“We think of these tropical forests as huge stores of carbon,” she said. “But we weren’t sure what trees were doing when they had all this water on their leaves,” she said, whether the leaves were able to take up the carbon at normal rates or if they were suppressed when wet.
“We looked more closely at these processes inside the forest so we can better resolve them to feed into the bigger picture.”
Measuring water, carbon and more
Over the course of the project, the researchers collected a whole suite of measurements to determine the different ways water moved throughout the canopy.
For example, Moore and her graduate students monitored how fast water was moving through the trees using sap flow sensors, made from scratch in Moore’s laboratory. She said the sensors were inserted into the xylem of the tree trunks where water flows from the roots to the canopy.
“The sensors can detect changes in the velocity of water inside the trees, which is then used to calculate transpiration for individual trees based on the cross-section of wood that carries water,” Moore said. “We have a total of 64 sensors running in trees of all sizes throughout the forest, which allows us to estimate the total amount of transpiration every 30 minutes of every day for the entire forest stand.”
Another experiment looked at the fluxes of carbon and water vapor into and out of the forest. Using a technique called eddy covariance, researchers can measure how much carbon enters the forest from the atmosphere and, conversely, how much water is released to the atmosphere from the forest.
“Some researchers refer to this as the ‘breathing’ of the biosphere,” Miller said, “and tracking these fluxes is important for understanding both forest health and changes in the forest over time.”
One of Moore’s doctorate students looked at different leaves in the tropical forest to determine how quickly they were moving water off their surfaces and how this affected carbon uptake. Moore said water droplets on the leaf surface block the uptake of carbon dioxide for photosynthesis, and in the tropics, leaves in some parts of the canopy remain wet more than 50 percent of the time.
“Not only did we want to try to document a decline in photosynthesis in wet leaves, which is really difficult to measure with the instruments currently available, but we also wanted to learn about how some species are better adapted to shed water than others,” she said.
They found that some leaves had particular adaptations to channel water quickly off their surfaces while others would hold water longer on their surfaces.
“It turns out that there is an amazing amount of diversity in leaf adaptations, from a shaggy carpet of trichomes that soak up water like a sponge, to thick plastic-like waxy layers with channels that divert away water immediately,” Moore said.
To understand variations throughout the canopy, the researchers established a profile system that measured humidity, carbon concentration, temperature, light and leaf wetness up and down the height of the canopy to show how those components changed over time and how the different environments within the canopy affected the plants’ behavior.
“We have a total of 64 sensors running in trees of all sizes throughout the forest, which allows us to estimate the total amount of transpiration every 30 minutes of every day for the entire forest stand.”
For instance, Miller said, in the lower canopy the combination of the low light, wetness of the leaves and the humidity alters how the smaller trees behave and how long it takes those trees to dry out after rain versus the larger, taller trees higher in the canopy with more sunlight.
The researchers found that the wet and dry processes are actually occurring faster than the Earth system models are currently running.
The timing of the rainfall, or when the leaves are wet, is extremely important for distinguishing between transpiration and evaporation processes, Miller said, because the movement of water was suppressed when the trees had wet leaves. However, the lower portion of the canopy could be very wet while the upper portions had dried, making for large contrasts in their behavior.
Compiling the pieces
Miller said they are now working on putting some of the concepts they developed into the CLM, the land surface model. One of Miller’s graduate students is currently developing equations to better reflect the timings of precipitation and how the model partitions what is wet and what is dry.
She said the CLM, which shows what the plants are doing, and the community atmosphere model or CAM “talk” to each other.
“If the atmosphere model is saying that it is raining at some low rate all the time in the forests — which is not true — then that is going to be giving incorrect information to the land model,” she said. “But if we have not been getting the wetness right, not getting when it rains, then that essentially means we get lots of other things wrong. We get temperature of the trees in the canopy wrong, then we get their carbon uptake and water use wrong. There is a whole cascade of things if we don’t get the precipitation right in the model.”
Miller said the Department of Energy is developing a new iteration of the Earth system model, and the Texas A&M group is hoping to contribute its findings to improve estimates of evapotranspiration and rainfall in the tropics.
"There is a whole cascade of things if we don’t get the precipitation right in the model.”
“We have one small part in helping to get these overall models right,” Miller said. “Our project might just be one small part, but the small parts add up.”