A multivariate conditional sampling method introduced in a companion paper is extended to investigate length scale contribution to turbulent fluxes. The method is briefly introduced. A model of convective processes thought to be active over the continental boundary layer during the First International Land Surface Climatology Project Field Experiment (FIFE) is used to interpret the conditional sampling results. The conditional sampling scheme is used to investigate moisture flux data obtained near the surface and near the inversion during the morning transition (a period of rapid mixed layer growth) and the afternoon's fully developed mixed layer (a period of slow mixed layer growth). Length scale distributions associated with the primary conditional sampling pair, vertical wind speed, and virtual potential temperature (a surrogate for density) were highly skewed with a predominance of shorter lengths relative to the longer ones. There was little difference in the length scale distributions for near-surface and near-inversion legs, nor were there obvious differences between morning and afternoon.

The contribution of length scales to the moisture flux showed that the bulk of the flux was associated with the relatively fewer longer length scale events and not with the predominance of shorter lengths. In general, the length scale decomposition showed the bulk of the moisture flux to be associated with longer lengths (greater than 300 m) near the inversion compared to the length scales of the near-surface flux. While there was little difference in the moisture flux distribution between morning and afternoon in the near-surface layer, there were substantial differences in the near-inversion distribution. Although these observations are consistent with cospectral decomposition of turbulent fluxes, the magnitudes of cospectral wavelengths were much longer compared to this analysis. Importantly, the near-surface flux has a substantial contribution from downward moving coherent structures originating in the inversion layer. Length scales greater than 200 m are a significant portion of this inversion contribution to the near-surface flux, a fact that must be taken into account in the determination of the surface flux ''footprint''. ¿ American Geophysical Union 1995