Choosing the averaging interval for comparison of observed and modeled fluxes along aircraft transects over a heterogeneous surface

Data from the April-May 1997 Cooperative Atmosphere Surface Exchange Study (CASES-97) are used to illustrate a holistic way to select an averaging interval for comparing horizontal variations in sensible heat (H) and latent heat (LE) fluxes from low-level aircraft flights to those from land surface models (LSMs). The ideal filter can be defined by considering the degree to which filtered aircraft fluxes 1) replicate the observed pattern followed by H and LE at the surface, 2) are statically robust, and 3) retain the heterogeneity to be modeled. Spatial variability and temporal variability are computed for different filtering wavelengths to assess spatial variability sacrificed by filtering and how much temporal variability can be eliminated; ideally, spatial variability should approach or exceed temporal variability. The surface pattern to be replicated is a negative slope when H is plotted against LE for a given time. This is required for surface energy balance if H or LE vary horizontally more than their sum, R_n - G, the difference between the net radiation and heat flux into the ground. Statistical confidence is estimated using conventional techniques. The same factors can be used to examine comparisons already done, or to estimate the number of flight legs needed to measure heterogeneity at a given scale in future field programs.