The Effect of Temperature and Increased Rainfall on Carbon Dioxide Exchange in a High Arctic Ecosystem: Improving Models and Testing Linearity of Response

Ecosystem carbon dioxide exchange determines the terrestrial flux of carbon dioxide to the atmosphere through the two component processes of photosynthesis and respiration. Temperature and water availability are dominant factors that regulate carbon dioxide exchange and ecosystem productivity across the globe. Yet, in many ecosystems, the complex interaction of temperature and water availability and their individual and combined effects on photosynthesis and respiration make it difficult to predict how climate change will affect carbon dioxide exchange. For example, climate warming can increase carbon dioxide uptake in wetter Arctic ecosystems, but leads to the loss of carbon dioxide to the atmosphere in drier Arctic ecosystems. Characterizing how temperature and water availability affect ecosystem carbon exchange in the Arctic is essential to determine whether the rate of climate warming could accelerate due to carbon dioxide losses from Arctic ecosystems. We conducted a multi-level warming experiment that included control plots and two- levels of warming in a widespread High Arctic ecosystem. Infrared lamps were used to warm the tundra during the growing season and rainfall was increased by 50 percent in control plots and the higher level warming treatment. Carbon dioxide exchange was measured using chamber techniques over several 24-hour periods during the growing season for three years and was resolved into the component fluxes. Climate and biophysical variables that affect carbon dioxide exchange rates were measured in coordination with these flux measurements. We chose to analyze the data from this experiment by fitting the data to light and temperature response functions for gross ecosystem photosynthesis and ecosystem respiration, respectively. Based on our sample size of 30 experimental plots (5 treatments x 6 replicates), we selected relatively simple models of carbon dioxide exchange to minimize overfitting, but considered linear and nonlinear models, and static and temperature dependent values for key parameters, such as the apparent quantum yield of the ecosystem. We fit the models to the data and estimated values and confidence intervals for the parameters by maximizing the likelihood. We will discuss the results of this experiment highlighting the unexpected effects of temperature on the parameterization of the carbon exchange models, the linearity of the response, and the merits of our approach to data analysis. Our analysis integrated remotely sensed estimates of the leaf area index with the ecosystem process models of carbon dioxide exchange. In the High Arctic, improving this integration and the assimilation of experimental data into these models is essential to predict the system's response to environmental change over the spatially heterogeneous landscape.