| dc.description.abstract | This study uses high resolution time series (10Hz) of wind components (u, v, w), sonic temperature (Ts), carbon dioxide (CO2), and vapor(H2O) density data to compute surface fluxes with application of Eddy Covariance method (EC) and Relaxed Eddy Accumulation method (REA). Eddy Covariance method is the most accurate method for measuring surface fluxes at the atmospheric boundary layer. However, the cost of fast response equipment required for EC measurement is very expensive for some trace gases. Recently, REA measurement, without the need of fast response equipment, is widely applied to measure greenhouse gas flux. Nevertheless, this method requires setting parameters before estimating, and has not yet been commercialized. This study takes flux es estimated by the EC method to discuss characteristics of empirical constant (b) and deadband (wd) required in REA techniques.
Observation data in 2012 was used to help determine the adequate deadband and empirical constant setting of REA with referenced fluxes estimated by EC. Observation data in 2013 was used to verify obtained REA parameter setting. Standard deviation of b is smaller with the deadband set as 0.6 to 0.8 times standard deviation of vertical wind speed. For vapor and carbon dioxide fluxes, using b value of sensible heat and eliminating data with vapor flux differences ranging between -20 and 100 m-mole/m3 or with carbon dioxide flux differences ranging between 0.05 and -0.2 m-mole/m3 can obtain good result. Comparing the surface flux calculated by Eddy Covariance method and Relaxed Eddy Accumulation method, the R-squared values are 0.98, 0.88, and 0.90 in sensible heat, vapor, and carbon dioxide fluxes, respectively. This study provides analysis of deadband and b value of great benefit to actual assembly of REA system in the future. | en_US |