Directional analysis procedures were extended to study the CO2 daily cycle in the low atmosphere at a rural
site. Weighted average, asymmetric circular functions, and conditional probability curves were applied to
concentrations recorded at three levels near surface. Wind speed and potential temperature gradient were
the meteorological variables used to establish a link to CO2. The highest weighted concentration was obtained
for wind speed below 4 m s−1 and for potential temperature gradient above 0.05 °C m−1. Asymmetric circular
functions were fitted to selected percentiles, the generalized von Mises function providing the lowest
RMSE. Slight differences among levels were observed in the transition between day and night, and higher
gradients were observed during the night. Four intervals were proposed when concentrations were binned
by intervals of meteorological variables. Wind speed was only relevant to separate intermediate concentrations.
Finally, conditional probability curves provided differences between the highest and lowest levels in
up to 20% of cases at 20–22 GMT.
