Practical applicability of high frequency correction theories to CO2 flux measured by a closed-path system