A modification of the moving point test method for nighttime eddy flux filtering on hilly and complex terrain

Nighttime correction of CO2 flux is one of the most important and challenging tasks in eddy covariance measurements over a complex mountainous terrain. There are two types of widely used methods for nighttime correction: u* filtering method and advection based filtering method. The most popular method is the u* filtering method that estimates the parameters of the ecosystem respiration function using the observed nighttime CO2 flux when u* is higher than a threshold, above which the dependency of the nighttime CO2 flux on u* fades away. The u* filtering method cannot be used at sites where the u* threshold cannot be identified and/or the drainage flow is developed at night, resulting in an underestimation of the CO2 flux. For overcoming that, the advection based method was developed for hilly terrain sites affected by drainage flow using the observed CO2 flux data from near sunset, when nighttime advection effect has not yet manifested. Gwangneung deciduous and coniferous sites in Korea are typical sites situated in hilly and complex terrain where the aforementioned methods are difficult to apply correctly because the filters' assumptions are violated. In this study, we propose a modified moving point test method (i.e., an automated statistical method for determining the u* threshold based on an iterative approach using a moving window for u*) using moving windows for `time' as well as u*, which can determine the `timing when CO2 drainage is generated' as well as the u* threshold for application of nighttime correction to the sites. Acknowledgment. This work was funded by the Weather Information Service Engine Program of the Korea Meteorological Administration under Grant KMIPA-2012-0001.