Diploma Thesis

Multilevel investigation of subcanopy respiration flux by REA conditional sampling above and within a spruce forest

Tobias Gerken (02/2008-12/2008)

Support: Thomas Foken, Andrei Serafimovich

This work investigates the subcanopy respiration flux (Re), which reflects mainly CO2 soil efflux, at the coniferous Weidenbrunnen site, located in the Fichtelgebirge Mountains (Northern Bavaria) and is applying a recently developed Relaxed Eddy Accumulation (REA) conditional sampling approach to high frequency eddy covariance data. Eddies transporting respiration flux information from the ground to the air above the canopy are considered to possess a unique and detectable CO2 and water vapor signature. The method used combines REA with hyperbolic dead bands and quadrant analysis, extracting the respiration events from the overall dataset. Data was collected during the EGER IOP1 experiment from Sep. 20 to Sep. 24 at five sampling heights. The overall REA formulation and its statistics were checked. The REA method is likely to overestimate fluxes due to overestimation of the β- coefficient. Re- events were identified and extracted from the data with hyperbolic thresholds of H = 0.25 and 0.5. For daytime conditions and above canopy systems the results were in the same order of magnitude to expected respiration fluxes, with H = 0.5 yielding best results. The estimates at the below canopy systems and during nighttime conditions were too large. The time fraction of daytime respiration events was < 10% and correlation coefficients (rc,q) approached -1, resulting in periods with no respiration signal extractable (22% of daytime data). The Re- signal above the canopy was correlated to net fluxes near ground, but the explained variance was small (~5%). Canopy storage of CO2 seemed negligible, but partial reassimilation by understory vegetation might occur, leading to underestimation of respiration. Re- events seemed rather dependent on turbulence (u*) than on canopy exchange regime, but sample size was too small for sound statistical testing. Most events are very short, but longer events contribute significantly to overall flux. Encountered event durations are within the timescales of coherent structure ejection phase. Connection between Re and coherent motions are supported by temporal distribution of events and phase space localization of Re- events similar to coherent structures. For future applications the REA formulation (β ; H ) should be improved and a longer data set is needed for final evaluation of the model‟s performance with subcanopy eddy covariance data and direct respiration measurements.

There are additional file downloads belonging to this publication


last modified 2009-03-14