Conditional sampling, bursting, and the intermittent structure of sensible heat flux

Sensible heat flux measurements were carried out to examine the characteristics related to frequency and duration of extreme flux events above a desert surface. First, the flux bursts were identified by a modified hyperbolic hole “quadrant analysis” in which the threshold criterion depends on the mean and standard deviation of the flux time statistics. Second, the frequency distribution of the extreme flux burst length was examined. It was found that the frequency distribution of the burst length is well approximated by two‐parameter power laws which account for two distinct types of bursting phenomena: one is organized, coherent, and has a relatively long duration, while the other is short lived, less organized, and has a relatively short duration. Linear regression analysis was used to determine the two power exponents of the burst length probability density function. Both exponents compare favorably with other reported measurements on concentration bursting in dispersing plumes from an elevated source. The cutoff burst length from one power law to the next also matches other studies. A strong correlation between bursting frequency and flux integral time scales was observed under a variety of atmospheric stability and surface wetness conditions. This strong correlation suggested a relationship between atmospheric stability and the mean time interval between bursts. It was found that the variation in the interval between bursts for the dynamic and dynamic convective sublayers is strongly related to the variation in stability. This was not the case for near‐convective and stable atmospheric conditions.