Diurnal and Seasonal Patterns in Ecosystem CO2 Fluxes and Their Controls in a Temperate Grassland

Abstract

There is considerable interest in understanding processes of carbon dioxide (CO2) uptake and release in grasslands and the factors that control them. Many studies have investigated how CO2 fluxes vary over time (monthly, seasonally, annually). However, with the exception of net ecosystem CO2 exchange (NEE) and ecosystem respiration (Reco), little information is available on diurnal flux patterns, despite their importance in determining total ecosystem CO2 gains and losses. To better understand these variations, we measured CO2 fluxes (NEE, Reco, soil respiration [Rsoil], canopy respiration [Rcanopy], plant assimilation [assimilation]) with a climate-controlled closed-chamber system over 24 h once a month from May to September during the 2005 growing season in a mesic grassland in Yellowstone National Park. We also assessed how environmental factors (photosynthetic active radiation [PAR], air temperature, soil temperature, soil moisture) were associated with these diurnal and seasonal flux patterns to identify the main drivers of the fluctuations in CO2. Measurements were conducted simultaneously on two plots: one irrigated, the other unirrigated. Absolute values of all fluxes were greatest in midsummer (June–July), and lowest in spring and fall (May, September) at both plots. Variation in soil moisture as a result of irrigation did not lead to pronounced differences in seasonal CO2 fluxes and did not influence the diurnal patterns of CO2 uptake and release. Instead, the diurnal and seasonal variations of our ecosystem fluxes were related to PAR and temperature (air/soil) and soil moisture and temperature (air/ soil), respectively, at both plots. Thus, continued anthropogenic increases in greenhouse gas emissions that are expected to change the intensity of radiation, temperature, and precipitation may strongly affect the diurnal and seasonal patterns in CO2 uptake and release. Such chamber-based information combined with the measurement of environmental variables could be important for modeling CO2 budgets when no continuous measurements are available or affordable.