Temporal complexity of terrestrial ecosystem functioning and its drivers.

The development of non-linear dynamics theory showed that simple processes can lead to high complexity in the functioning of nature, with ecological studies showing that non-linear dynamics are common across populations of different taxa. However, whether the energy and matter fluxes of entire ecosystems follow non-linear dynamics, and how complex these dynamics are, is still unknown. We investigate the drivers of- and trends in the temporal complexity of ecosystem functioning by calculating the correlation dimension of gross primary production (GPP), ecosystem respiration, and net ecosystem production. We use long-term, eddy-covariance C fluxes from 57 terrestrial ecosystems, including boreal, temperate, and Mediterranean biomes. Generally, ecosystems located under more temporally complex weather also show more complex C fluxes. Causal analyses indicate that larger C fluxes generally cause higher temporal complexity, and larger and temporally complex C fluxes reduce interannual variability, suggesting higher resistance to perturbations. We report a positive trend in GPP complexity over time, which correlates with increasing GPP. This result may indicate that ecosystems are increasingly responsive to endogenous or exogenous stimuli, but the biology underlying these trends is not yet understood. We show that the short-term temporal complexity of ecosystem functioning can elucidate ecosystem properties otherwise missed by longer timescales.