This study analyzes data collected from repeated aircraft runs 30 m ov
er alternating regions of irrigated and dry nonirrigated surfaces, eac
h region on the order of 10 km across, during the California Ozone Dep
osition Experiment (CODE). After studying the scale dependence of the
flow, the variables and their fluxes are decomposed into means for sub
legs defined in terms of irrigated and nonirrigated regions and deviat
ions from such subleg means. Since the repeated runs were flown over t
he same track, compositing the eight flight legs for each of the two d
ays allows partial isolation of the influences of surface heterogeneit
y and transient mesoscale motions. A variance analysis is carried out
to quantify the relative importance of surface heterogeneity and trans
ient mesoscale motions on the variability of the turbulence fluxes. Th
e momentum and ozone fluxes are more influenced by transient mesoscale
motions while fluxes of heat, moisture and carbon dioxide are more in
fluenced by surface heterogeneity. The momentum field is also influenc
ed by a quasi-stationary mesoscale front and larger scale velocity gra
dients. For the present case, the mesoscale modulation of the turbulen
t flux is numerically more important than the direct mesoscale flux. T
his spatial modulation of the turbulent fluxes leads to extra Reynolds
terms which act to reduce the area-averaged turbulent momentum flux a
nd enhance the area-averaged turbulent heat flux.