Mapping the flux field around eddy-covariance measurements

Tower-based eddy-covariance (EC) measurements are suitable for continuously monitoring the surface-atmosphere exchange of heat, water vapor, CO₂ and other trace gases at selected sites. However, such sites are often located in somewhat heterogeneous terrain. Depending on atmospheric variability, the resulting observations represent only a subset of source areas surrounding the EC measurement (“location bias”). It is hence desirable to improve the spatial representativeness and temporal consistency of EC measurements, in particular for synthesis with remote sensing and numerical modelling applications. The objective of this study is to provide consistent flux time-series for a target region, rather than for a spatio-temporally variable patch of surface close to the measurement location.

We present a procedure that determines from a single EC tower the spatio-temporally explicit flux field of its surrounding, based on the extraction and projection of environmental response functions (ERF, Metzger et al., 2013). The underlying principle is to extract the relationship between biophysical drivers and ecological responses from measurements under varying environmental conditions. Provided sufficiently good calibration, the resulting ERF can be used for projecting the surface-atmosphere exchange to the larger surrounding of the EC measurement. Among others, the resulting flux grids can then be summarized as probability density functions of the surface-atmosphere exchange for individual land covers. This allows informing mechanistic models capable of representing sub-grid land cover heterogeneity with the expected flux value and its spatial variability over a target domain. In such way linearity assumptions can be mitigated when scaling between finer (order 0.01 km²) and coarser (order 100 km²) spatial resolutions of in-situ observations and model representations, respectively.

Here, the procedure is applied to EC data from July and August 2011 at the AmeriFlux Park Falls tower, Wisconsin, U.S.A. For heat fluxes, the residual standard error between EC measurement and calibrated ERF is on the order of 2 W m^-2. When the ERF is used for hourly projections onto a 0.1 km grid, the spatial coverage is >95%, >80% and >70%, for target areas around the tower of 25 km², 100 km² and 400 km², respectively. During noontime periods, the average heat fluxes from the measurement source areas and from the target domain differ up to 100 W m^-2 or 65%. Moreover, the spatial variation of the surface-atmosphere exchange can be equal to or larger than the magnitude of the measured EC flux. Lastly, we discuss the potential of this procedure for consistently informing mechanistic models on in-situ fluxes.

Reference

Metzger, S., Junkermann, W., Mauder, M., Butterbach-Bahl, K., Trancón y Widemann, B., Neidl, F., Schäfer, K., Wieneke, S., Zheng, X. H., Schmid, H. P., and Foken, T.: Spatially explicit regionalization of airborne flux measurements using environmental response functions, Biogeosciences, 10, 2193-2217, doi:10.5194/bg-10-2193-2013, 2013.