Energy balance closure of eddy-covariance data: A multisite analysis for European FLUXNET stations

摘要

a- Analysis of EB deficit in multi-site analysis with up to 26 sites. a- Relative EB deficit is larger for very unstable conditions than for less unstable conditions, because of reduced mechanically-induced turbulence for very unstable conditions, and not related with increased thermally-induced turbulence. a- Average absolute EB deficits are largest when relative EB deficits are smallest. a- Relative EB deficits are smallest for strong thermally-induced turbulence (TT) or strongly suppressed TT, but largest for intermediate conditions. a- For non-neutral atmospheric conditions the relative EB deficit is not given by the slope of the regression line of the absolute EB deficit as a function of net radiation. a- Storage terms are important for reducing relative EB deficit during nighttime and stable conditions, but otherwise not so important. This paper presents a multi-site (>20) analysis of the relative and absolute energy balance (EB) closure at European FLUXNET sites, as a function of the stability parameter I34, the friction velocity u *, thermally-induced turbulence, and the time of the day. A focus of the analysis is the magnitude of EB deficits for very unstable conditions. A univariate analysis of the relative EB deficit as function of I34 alone (both for individual sites and a synthesis for all sites), reveals that the relative EB deficit is larger for very unstable conditions (I34 degree 1.0) than for less unstable conditions (a degree 0.02> I34 a[control][yena degree 1.0). A bivariate analysis of the relative EB deficit as function of both I34 and u *, however, indicates that for situations with comparable u * the closure is better for very unstable conditions than for less unstable conditions. Our results suggest that the poorer closure for very unstable conditions identified from the univariate analysis is due to reduced u * under these conditions. In addition, we identify that the conditions characterized by smallest relative EB deficits (elevated overall turbulence, mostly during day time) correspond to cases with the largest absolute EB deficits. Thus, the total EB deficit at the sites is induced mostly under these conditions, which is particularly relevant for evapotranspiration estimates. Further, situations with the largest relative EB deficits are generally characterized by small absolute EB deficits. We also find that the relative EB deficit does generally not correspond to the regression line of absolute EB deficit with the net radiation because there is a (positive or negative) offset. This can be understood from theoretical considerations. Finally, we find that storage effects explain a considerable fraction of the large relative (but small absolute) nocturnal EB deficits, and only a limited fraction of the overall relative and absolute EB deficits.