Turbulence intermittency driven by submeso motions limits the progress of turbulence theory. Field observations from the Horqin Atmospheric Boundary-Layer and Environment Experimental Station (relatively flat underlying surface) and the Semi-Arid Climate and Environment Observatory of Lanzhou University (typically complex underlying surface of the Loess Plateau), China were used to investigate turbulence intermittency. At first, datasets over relatively flat Horqin station were used to explore the quantitative characterization and basic features of turbulent intermittency. An automated algorithm to Separate and reconstruct Submeso and Turbulent motions (SST) was improved for more accurately extraction and quantitative characterization of submeso motions. The existing intermittency intensity indices, the local intermittency strength of turbulence (LIST) and intermittency strength (IS), which are based on kinetic energy only, are revised by considering the potential energy of submeso and turbulent motions to quantify intermittency intensity more comprehensively. The analysis of eight cases revealed that turbulent intermittency events are characterized by quiescent (pulsation, material, and energy transportation are weak) and burst (pulsation, material, and energy transportation fluctuate violently) periods. The conversion of both the kinetic and potential energy of submeso to turbulent motion contributes to the transition from quiescent to burst periods. The transition always occurs after  <0 (the Total Energy difference between the submeso motion and turbulence), followed by a significant increase in . Atmospheric stability decreases during the transition from quiescent to burst periods in most cases. In a totally intermittent night, the burst periods take up most of the material and energy transport, and the amount transported is not smaller than that during a totally turbulent night. The weaker the intermittency at night, the greater the capacity of turbulent transport. A comparison of five types of turbulence intermittency intensity indices highlights the consistency and advantages between LIST (IS) and indices in the literature. Then, field observations of the Loess Plateau were used to investigate turbulence intermittency over complex underlying surface. The findings revealed that basic features of turbulence intermittency events over complex underlying surfaces were consistent with that over relatively flat underlying surface, including distinct performance of quiescent and burst periods, energy conversion from submeso motion to turbulent motion, and the change in atmosphere stability. Finally, the influence of turbulence intermittency on the classical energy non-closure issue was explored. The results show that different periods of turbulence intermittency events make different contribution to energy non-closure over the Loess Plateau. We found that the presence of submeso motions contributed to energy closure, with an energy closure of 86% during daytime and 36% during nighttime over Loess Plateau. The energy closure during the burst period was approximately 98% during daytime and 68% during nighttime, approaching closure and far exceeding the overall closure rate at night, whereas the energy closure during the quiescent period was significantly low, at only 70% during daytime and only 17% during nighttime. This suggests that turbulent intermittency is a very important factor causing energy non-closure over complex underlying surfaces. The results are highly significant for a better comprehension of turbulence intermittency and surface-atmosphere interactions.

Turbulence intermittency,submeso motion,stable boundary layer,surface energy imbalance