QUALITY CONTROL OF HIGH FREQUENCY DATA IN THE ATTO PROJECT

Control quality in turbulence data is an essential step before carrying on any analysis. However, there are not standard tests to follow, and the evaluation criteria are left to the researcher. In this context, the main goal of this study is perform a quality control of high frequency data (10 Hz) measured at two levels above an Amazonian forest (39.4 and 81.6 m, where the average forest height is approximately 40 m). First of all, spikes were removed; in the following, changes in standard deviation in each variable were evaluated with the intention to detect sensor malfunction or low levels of turbulence. Finally, to detect trends and unsteady cases the Reverse Arrangement Test was applied, followed by an analysis of the temporal variation of each variable (difference between maximum and minimum values). After all control quality procedures, 15.8% of the data measured at 81.6 m and 40.7% of the data measured at 39.4 m remained. These results show the high quantity of inconsistent runs, highlighting the importance of the control quality in order to avoid that they contaminate the ensuing analysis.