Reply to the comment on Vickers et al. (2009): Self-correlation between assimilation and respiration resulting from flux partitioning of eddy-covariance CO_2 fluxes

摘要

We thank Lasslop et al. (2010) for their remarks and interest, and appreciate the opportunity to clarify our manuscript. Hopefully this attention will increase awareness of the issue of self-correlation in the CO_2 flux community. There is currently arevival of the long-known concerns about self-correlation in the turbulence and boundary-layer community. Anderson (2009) suggests that because the degree to which self-correlation affects regression relationships is not simple to calculate in all cases, techniques that avoid the problem are preferred. Self-correlation (Hicks, 1978; Klipp and Mahrt, 2004; Baas et al, 2006; Toba et al, 2008) has also been referred to as spurious correlation (Pearson, 1897; Kenney, 1982; Kenney, 1991; Jackson and Somers,1991; Aldrich, 1995; Brett, 2004), spurious inference or faulty inference (Prairie and Bird, 1989) and as the shared variable problem in the statistics literature (see additional references in Vickers et al., 2009). It arises when one group of variablesis plotted against another, and the two groups have one or more variables in common. For example, x+y and x are self-correlated because they share the common variable x. The degree of self-correlation is proportional to the ratio of the variances of theshared to the unshared variable. The relationship between the orientation of the self-correlation and of the true physical correlation is important.