Effects of climate variability and functional changes on carbon cycling in a temperate deciduous forest

Abstract

Temperate forests are globally important carbon (C) stocks and sinks. A decadal (1997-2009) trend of increasing C uptake has been observed in an intensively studied temperate deciduous forest, Sorø (Zealand, Denmark). This gave the impetus to investigate the factors controlling the C cycling and the fundamental processes at work in this type of ecosystem. The major objectives of this study were to (1) evaluate to what extent and at what temporal scales, direct climatic variability and functional changes (e.g. changes in the structure or physiological properties) regulate the interannual variability (IAV) in the ecosystem C balance; (2) provide a synthesis of the ecosystem C budget at this site and (3) investigate whether terrestrial ecosystem models can dynamically simulate the trend of increasing C uptake. Data driven analysis, semi-empirical and process-based modelling experiments were performed in a series of studies in order to provide a complete assessment of the carbon storage and allocation within the ecosystem and clarify the mechanisms responsible for the observed variability and trend in the ecosystem C fluxes. Combining all independently estimated ecosystem carbon budget (ECB) datasets and other calculated ECB components based on mass balance equations, a synthesis of the carbon cycling was performed. The results showed that this temperature deciduous forest was moderately productive with both high rates of gross primary production and ecosystem respiration. Approximately 62% of the gross assimilated carbon was respired by the living plants, while 21% was contributed to the soil as litter production, the latter balancing the total heterotrophic respiration. The remaining 17% was either stored in the plants (mainly as aboveground biomass) or removed from the system as wood production. In general, the ECB component datasets were consistent after the cross-checking. This, together with their characterized uncertainties, can be used in model data fusion studies. The sensitivity of the C fluxes to climatic variability was significantly higher at shorter than at longer time scales and changed seasonally. At the annual time scale, the IAV in net ecosystem exchange of CO2 (NEE) was mostly determined by changes in the ecosystem functional properties. This indicated that the processes controlling the function change need to be incorporated into the process-based ecosystem models. The process-based model (CoupModel) applied in this study was able to simulate the phenology and observed carbon fluxes well at short (i.e. diurnal or seasonal) time scales but did not reproduce the decadal trends in NEE when global parameter estimates were used. Annual based parameter estimates were able to reproduce the trends; changes in the yearly fitted parameters (e.g. the light use efficiency) indicated the importance of functional change, as also shown in the analysis using the semi-empirical models. A role for nitrogen demand during mast years was also demonstrated in the estimated parameters. Nitrogen cycling and dynamics were identified as possible internal factors that need further investigation.