Estimating Spatially Heterogeneous Contributions to Ecosystem Scale Fluxes Directly From Eddy Covariance Measurements: A Case Study in Siberian wet Polygonal Tundra on Samoylov Island, Lena River Delta
Abstract
The eddy covariance method is used widely to measure the turbulent exchange fluxes of climate relevant gases such as carbon dioxide and methane. One important assumption in eddy covariance theory is homogeneity of the surface over which measurements are conducted. However, in reality the method is often applied in very heterogeneous areas and the effect of that heterogeneity on the measurement time series is not fully agreed on in the scientific community. Since the eddy covariance method relies on the time-for- space substitution concept (Taylor's frozen turbulence field assumption), spatial disturbance of the assumed homogeneity should leave similar traces in the time series data that could be detected with state-of-the-art times series statistics approaches, primarily frequency analysis. This should be easiest to detect where the observed heterogeneity is characterized by the steepest possible small-scale spatial contrast of fluxes while also exhibiting certain regularity. Thus, polygonal tundra with its regular micro-relief of very wet polygon depressions with high rates of photosynthesis and methane emission on the one hand and relatively "dry" elevated polygon rims with lower rates of photosynthesis, higher rates of respiration, and extremely low methane emissions on the other hand is considered well-suited to explore the performance of this technique of flux footprint separation. We present a case study using eddy covariance data of water vapour, carbon dioxide and methane from the Lena River Delta, Siberia, and high-resolution aerial photography to demonstrate that spatial heterogeneity correlates with deviations in eddy covariance co-spectra from the idealized co-spectra. This new method - if successful beyond a single case study - could become widely used wherever fluxes are measured over spatially heterogeneous surfaces. It would be especially helpful to move towards more accurate upscaling in areas where emission rates and processes vary greatly on small spatial scales, such as the vast and hard to access high latitude tundra ecosystems.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2008
- Bibcode:
- 2008AGUFM.B24A..08S
- Keywords:
-
- 0426 Biosphere/atmosphere interactions (0315);
- 0428 Carbon cycling (4806);
- 0430 Computational methods and data processing;
- 0475 Permafrost;
- cryosphere;
- and high-latitude processes (0702;
- 0716);
- 0490 Trace gases