New computer models, based on photosynthesis and respiration data from dogwood trees, can improve the accuracy of climate change predictions. This is the conclusion that a team at in Oak Ridge National Laboratory’s Environmental Sciences Division (ORNL) reached in attempt to determine the quantity of carbon dioxide removed from the atmosphere by these deciduous trees.
The group of eight scientists at ORNL, led by Jeff Warren, conducted a number of experiments that use a traceable form of carbon dioxide that contains stable carbon isotope. This way, they were able to mark trees individually and track the carbon dioxide as it moves through the tree, by analyzing samples from the leaves, fruits, stem tissue, roots and soil.
Each tree was isolated using a handmade PVC structure, equipped with air conditioning and environmental monitoring systems. These act as greenhouses that trap the gases, allowing the scientists to introduce the traceable carbon in controlled conditions.
Warren is certain that these tests will provide the missing information about the fate of carbon dioxide when it enters the plant. The trees were exposed for two hours. After that, two of the test trees were covered with shade-cloths in order to limit the photosynthesis process.
The experiment was repeated at three different seasons in order to determine the influence of lighting conditions and temperature. According to Warren, the seasonal tests allowed the team to establish that existing models disregard a substantial amount of carbon released in the atmosphere.
Considering that deciduous forests are a big percentage of the Earth’s forest cover, Colleen Iversen, another member of the ORNL team, points out that small errors caused by the model, can influence greatly the output. In order to make accurate predictions of future climates, and be able to react and prepare adequately, the models should be as representative as possible.
The aim of the team is to eventually make their findings publicly available by incorporating them in the Community Land Model.