Influence of gross primary production in energy partitioning in three different forest ecosystems based on Eddy Covariance time series analysis

Ecosystems plays a key role on the interaction between the land surface and the atmospheric processes being responsible for strong feedback processes that affect the climate by modifying the relative contribution of the latent and sensible heat to the total energy of the atmospheric air, i.e., the energy partitioning processes. The mechanisms and consequences of this feedback are uncertain and must be studied to evaluate their influence on global climate change.

In this study, our overall objective was to assess the Gross Primary Production (GPP) dynamics and the energy partitioning patterns in three different European forest ecosystems through time series analysis of eddy covariance data. The three contrasted forest types in terms of functioning and climate were an Evergreen Needleleaf Forest in Finland (ENF_FI), a Deciduous Broadleaf Forest in Denmark (DBF_DK), and a Mediterranean Savanna Forest in Spain (SAV_SP). In each site there was and eddy covariance flux tower from which meteorological data, carbon and energy fluxes were analyzed. Firstly, a univariable time series analysis of all variables was made by means of the Buys-Ballot tables, i.e., average year, to study the intra-annual dynamics and then, through the use of the autocorrelation function the interannual dynamics were assessed.  Finally, causality of GPP and energy fluxes was studied with Granger causality tests.

Results show that temperature and solar radiation were the main limiting factors in the Northern ecosystems while water availability was determinant for growth in the Mediterranean ecosystem. The autocorrelation function showed that GPP and the meteorological variables in the SAV_SP were more irregular and show lower memory at the long term than at the short one. In addition, this ecosystem presented higher radiation and a larger amount of H+LE, showing the highest Bowen ratio and a lower primary production efficiency in terms of total energy (GPP/(H+LE)). On the contrary, both northern ecosystems showed similar production efficiencies in terms of total energy. However, the DBF_DK showed lower Bowen ratio related to a larger amount of latent heat in relation to sensible heat in associated to the larger plant activity in this forest. Finally, the Granger causality tests showed that the vegetation feedback to the atmosphere was more noticeable in the ENF_FI and the DBF_DK at the short term, influencing latent and sensible heat fluxes.

In conclusion, the impact of the vegetation on the atmosphere influences the energy partitioning in a different way depending on the vegetation type, which makes essential the study of the vegetation dynamics at the local scale to parameterize with more detail these processes and build improved global models.