New insights on the variability of ecosystem functioning across time scales

C. Pappas, M.D. Mahecha, D.C. Frank, and D. Koutsoyiannis, New insights on the variability of ecosystem functioning across time scales, AGU 2015 Fall Meeting, San Francisco, USA, doi:10.13140/RG.2.2.24568.65280, American Geophysical Union, 2015.



Ecosystem functioning is monitored worldwide over several decades. However, a comparative in-depth characterization of the temporal variability of essential ecosystem processes, such as for example carbon assimilation and respiration is still lacking. The intra-annual (sub-diurnal, diurnal, and seasonal) variability of these processes can be well described by basic mechanisms such as the plant response to light. In contrast, the inter-annual variability and its origins and magnitude, remain highly uncertain. To date, there have only been a few attempts to investigate these issues across sites, ecosystems variables, and time scales, yet a general and comprehensive overview is outstanding. Here, we present a synthesis of a wide range of observations over Europe, namely: (i) eddy covariance measurements of carbon, energy, and water fluxes, (ii) satellite data of leaf area index and photosynthetically active radiation absorbed by plants, (iii) tree-ring widths, and (iv) dendrometer measurements of tree stem radius changes, and we analyze their variability from the half-hourly to the decadal time scale. Our analysis shows that all ecosystems can be characterized by three distinct regimes of variability (sub-daily, daily-seasonal, and seasonal-annual) confined within the ranges of the available resources, i.e., water (precipitation) and energy (radiation and temperature). We find a convergence of the range of variability of hydrometeorological drivers. Surprisingly, such convergence is not reflected in the variability of the ecosystem responses across sites. Although the magnitude of variability of ecosystem functioning varies across sites, the temporal dependences present the same characteristics over time scales spanning five orders of magnitude. We show that this behaviour can be well simulated by combining simple stochastic models with deterministic harmonics (diurnal and annual cycles). This allows us to statistically characterize the short- and long-term variability of ecosystem functioning and to disentangle the relative contribution of major climatic drivers across time scales, offering a parsimonious representation of ecosystem dynamics.

PDF Full text (2091 KB)

See also: