Observations of CH₄ flux from inland waters based on the relaxed eddy accumulation method

Inland water bodies are important CH₄ sources. The accurate observation of CH₄ fluxes is key to quantitatively assessing emissions. The relaxed eddy accumulation (REA) method is a technique for calculating material fluxes using the differences of gas concentration in upward and downward air movements over a period, the standard deviation of vertical wind speed (σ_w), and an empirical coefficient b. We utilized observation data obtained with the eddy covariance (EC) method at the East Taihu Lake site (DTH site) of the mesoscale flux observation network in the large water body of Lake Taihu and the Guandu small water body aquaculture pond site (GD site) in Anhui. By leveraging the similarity between physical quantities in the REA method and adopting the "proxy variable method", three types of five methods, including synchronous b-value, fixed b-value (including mean, median, and fitted slope), and corrected b-value, were used to determine the empirical coefficient b in REA on the basis of determining the optimal proxy variable, and ultimately obtain CH₄ flux. We assessed the applicability of the REA method for observing CH₄ fluxes in large and small inland water bodies and clarified the optimal of the key coefficient b. The results showed that among the b-values for all flux results at both sites, the b-value for water vapor flux exhibited the smallest dispersion, making it suitable as the optimal proxy variable for the REA method. The interquartile range (the difference between the third and first quartiles) of the water vapor flux b-value showed a trend of first decreasing and then increasing with the increase of the vertical wind speed threshold (w_d). In this study, w_d was set to 0.4 times σ_w. Compared with CH₄ fluxes observed by the EC method, at the DTH site, CH₄ fluxes calculated using the fixed b-value mean and fixed b-value median performed best. At the GD site, CH₄ fluxes calculated using the fixed b-value fitted slope performed best. The optimal b-values for the DTH and GD sites were 0.443 and 0.500, respectively. CH₄ fluxes obtained by the REA method at both sites showed good consistency with those obtained by the EC method, indicating that the REA method is applicable for observing CH₄ fluxes over inland water underlying surfaces. Generally, the b-value of large water bodies is smaller than that of small water bodies.