A spectral analysis of biosphere–atmosphere trace gas flux densities and meteorological variables across hour to multi-year time scales

Abstract

The advent of long-term studies on CO 2 and water vapor exchange provides us with new information on how the atmosphere and biosphere interact. Conventional time series analysis suggests that temporal fluctuations of weather variables and mass and energy flux densities occur on numerous time scales. The time scales of variance that are associated with annual time series of meteorological variables, scalar flux densities and their covariance with one another, however, remain unquantified. We applied Fourier analysis to time series (4 years in duration) of photon flux density, air temperature, wind speed, pressure and the flux densities of CO 2 and water vapor. At the daily time scale, strong spectral peaks occurred in the meteorological and flux density records at periods of 12 and 24 h, due to the daily rising and setting of the sun. At the synoptic time scale (3–7 days) the periodic passage of weather fronts alter available sunlight and temperature. In turn, variations in these state variables affect carbon assimilation, respiration and transpiration. At the seasonal and semi-annual time scales, a broad spectral peak occurs due to seasonal changes in weather and plant functionality and phenology. In general, 21% of the variance of CO 2 exchange is associated with the annual cycle, 43% of the variance is associated with the diurnal cycle and 9% is associated with the semi-annual time scale. A pronounced spectral gap was associated with periods 3–4 weeks long. Interactions between CO 2 flux density ( F c) and sunlight, air temperature and latent heat flux density were quantified using co-spectral, coherence and phase angle analyzes. Coherent and in-phase spectral peaks occur between CO 2 exchange rates and water vapor exchange on annual, seasonal and daily time scales. A 180° phase shift occurs between F c and photon flux density ( Q p ) on seasonal and daily time scales because the temporal course of sunlight corresponds with the withdrawal of CO 2 from the atmosphere, a flux that possesses a negative sign. Covariations between F c and T air experience a 180° phase shift with one another at the seasonal time scale because rising temperatures are associated with more carbon uptake. At daily time scales the phase angle between F c and T air is on the order of 130°. This phase lag can be explained by the strong dependence of canopy photosynthesis on available light and the 2–3 h lags, which occur between the daily course of sunlight and air temperature. Spectral analysis was used to investigate the performance of a biophysical model (CANOAK) across a spectrum of time scales. By varying meteorology, leaf area index and photosynthetic capacity, the model was able to replicate most of the spectral gaps and peaks that were associated with CO 2 exchange, when soil moisture was ample.