The effects of different calibration and frequency response correction methods on eddy covariance ozone flux measured with a dry chemiluminescence analyzer

• O3 flux over a wheat field in China was measured with eddy covariance method.
• We evaluated the effects of 3 calibrations and 3 spectral correction methods on O3 flux.
• The Allan–Werle variance was used to determine optimal averaging time for calibration.
• The fast-response O3 analyzer showed noise and damping in the high frequency range.
• There were differences among O3 fluxes corrected with 3 spectral correction methods.

Ozone (O3) flux (Fo) was measured over a wheat field in Yucheng, China, using the eddy covariance technique. A dry chemiluminescence fast-response O3 analyzer and a UV-absorption based slow-response O3 analyzer were used for measuring fast O3 concentration fluctuations and absolute concentrations, respectively. The main objective of this study was to determine a group of suitable methods for calculating Fo based on the performance of the two O3 analyzers in the field. We evaluated the effects of three calibration methods on Fo. These calibration methods include the ratio method (RM, it assumes that the fast-response analyzer’s signal-output is directly proportional to absolute ambient O3 concentration within a 30-min interval), the ratio offset method (ROM, it is based on the ratio method with analyzer’s offset modification), and the ratio variation method (RVM, it is based on the ratio method with the fast-response analyzer’s gain factor variation within a 30-min interval). Three frequency response corrections (analytic, revised analytic, and in-situ) for the estimation of Fo were also evaluated. Our results show that: (1) Compared to the O3 flux with the commonly-used method (RM), on average, the fluxes with ROM and RVM were decreased by about 9% and 7%, respectively. (2) Based on the spectral analysis, the fast-response O3 analyzer’s noise could cause random errors, but did not result in a systematic error in Fo. Tubing attenuation in O3 concentration fluctuations can cause up to 20% loss in Fo. (3) Frequency response corrections for Fo with the original analytic method, the in-situ method, and the revised analytic method were 34.6%, 23.8% and 25.4% on average, respectively.