Urban green space plays a significant role on carbon cycling in urban areas. Accurate estimation of CO₂ fluxes above urban green space using the long-term field observation method provides critical information about climate change reactions and regional carbon sequestration evaluations. Continuous eddy-covariance measurements of CO₂ fluxes were conducted to analyze the carbon exchange status in 2009 and 2010. The analyses included daily variations in the CO₂ fluxes during different seasons, annual variations in light response parameters, including apparent light use efficiency (α) and maximum photosynthetic productivity (P_max). Annual variations in the net ecosystem CO₂ exchange (NEE) and the ecosystem respiration (R_eco) were also studied.
A tower located in the botanical garden of Dongguan with 4 levels and a total height of 20 m was used for the measurements. The CO₂ flux at a height of 20 m was analyzed to study the CO₂ exchange between an urban green space and the atmosphere interface. According to the eddy-covariance method and the detection limits of the instrument, abnormal data were removed to ensure accuracy. The daytime NEE and the photosynthetic active radiation (PAR) were used to formulate the Michaelis-Menten model as the gap-filling model during the daytime, while the night respiration (R_eco) and soil temperature at 5 cm depth (T_s) were used to calculate the exponential equation as the gap-filling model at night.
The major results were: (1) the NEE was -91.9 gC·m^-2·a^-1in 2009 and -116.5 gC·m^-2·a^-1 in 2010, indicating that urban green space was a net carbon sink. The NEE varied seasonally with environmental factors. Urban green space was a carbon source from December-March and a carbon sink from April - November each year. (2) The daily total CO₂ flux (F_c) was negative during the winter, but positive during the other seasons. CO₂ storage (F_s) occupied a larger proportion during winter than during the other seasons. Daily CO₂ concentrations (C_c) were highest before sunrise and lowest after sunset. C_c was highest in the winter and lowest in the summer. (3) The annual average α was (0.00134±0.00035) mgCO₂·μmol^-1Photons and P_max was (1.006±0.283) mgCO₂·m^-2·s^-1. (4) The annual R_eco was 1370.4 gC·m^-2·a^-1in 2009 and 1384.8 gC·m^-2·a^-1 in 2010. R_eco varied with changes in the soil temperature during the year. The simulated R_eco during the night was less than the daytime result produced by the Michaelis-Menten model at the same temperature. (5) Because the urban green space was a mixture of forest and grassland, the NEE of the urban green space was smaller than that of a forest ecosystem, but larger than a grassland ecosystem. Compared with other urban green spaces, the NEE in Dongguan was lower than northern urban forests due to the high R_eco caused by high temperatures. (6) Correlation analyses between the NEE and meteorological factors like the PAR, air temperature (T_a) and saturation vapor pressure differential (VPD) were performed. The partial correlation coefficient of the PAR was larger than the others, indicating that the PAR was the most important factor to the NEE. The partial correlation coefficient of T_a was negative during the winter and positive during the summer. The partial correlation coefficient of VPD was positive year-round and was higher under high VPD conditions.