Partitioning of eddy covariance-measured net ecosystem exchange of CO2 in tropical lowland paddy

Abstract

Net ecosystem exchange of CO2 (NEE) measurement was carried out in tropical lowland paddy at ICAR-National Rice Research Institute, Cuttack, Odisha, India, in 2015 using eddy covariance technique with the objective to assess the variation of NEE of CO2 in lowland paddy and to find out the most suitable model for better partitioning of net ecosystem exchange of CO2 in tropical lowland paddy. Paddy is grown twice (dry and wet season) a year in this region in the lowland, and the field is kept fallow during the remainder of the year. Two different flux partitioning models (FPMs)—the rectangular hyperbola (RH) and the Q10, were evaluated to assess NEE of CO2, and its partitioning components—gross primary production (GPP) and ecosystem respiration (RE), and the resulting flux estimates were compared. The RH method assessed the effects of photosynthetically active radiation on the NEE, whereas the Q10 method utilized the relationship between ecosystem respiration and temperature in lowland paddy. The average NEE during the dry season and wet season was − 1.62 and − 1.83 g C m−2 d−1, respectively, whereas it varied from − 5.71 to 2.29 g C m−2 d−1 during the observation period covering both the cropping seasons and the fallow period. The mean difference between modeled GPP and RE from two FPMs was found significant in both the seasons. The maximum correlation for GPP estimation was found between two FPMs at the panicle initiation stage during both the dry season (R2 = 0.767) and wet season (R2 = 0.321). It was evident from the study that the Q10 method reliably produced the most realistic carbon flux estimates over the RH method, for the lowland paddy. The Q10 model which used nighttime flux and temperature data to estimate RE produced estimates that had lower prediction error (RMSE) as compared to the RH model. It can be concluded that in lowland paddy, the Q10 predicted better estimates of RE and GPP values than the RH method, suggesting that the Q10 model can be used for partitioning of NEE in tropical lowland paddy.