Spatio-temporal distribution of soil respiration in dune-meadow cascade ecosystems in the Horqin Sandy Land, China

Purpose: This study investigated soil respiration (R-s) in adjacent ecosystems with a cascade distribution in the semi-arid region of the Horqin Sandy Land (HSL) and determined the spatio-temporal distribution of R-s as well as its relationship with the influencing factors. Methods: Four ecosystems with a cascade distribution (i.e., cascade ecosystems) were selected from the dune meadow area of the HSL, including sandy bare ground (SBG), a sandy Caragana microphylla community (SCC), transitional artificial Populus forest (TPF), and a meadow Phragmites communis community (MPC). In 2014 and 2015, R-s, soil temperature (T-s) and soil moisture content (M-s) were measured every 10-15 days during the growing season, and 24-h continuous observations were carried out during the early, middle and late stages of growth. In addition, landforms, soil properties, vegetation characteristics and groundwater table depths of the cascade ecosystems were surveyed. Results: The R-s values of the cascade ecosystems showed significantly different seasonal variations (P < 0.05). During the growing season, the mean, coefficient of variation, and temperature sensitivity of R-s were in order of SBG < SCC < TPF < MPC, which was consistent with landforms, soil properties and vegetation characteristics. Due to the differences in landforms, soil properties, vegetation characteristics and groundwater table depth, the R-s values of SBG and SCC were significantly positively correlated with M-s. but not with T-s. TPF had the optimum M-s. for R-s, while the R-s value of MPC was significantly negatively correlated with M-s. Further, the R-s values of TPF and MPC were significantly positively correlated with T-s. T-s and M-s. together explained 49.9%, 74.8%, 79.5%, 92.1% of R-s in the cascade ecosystems(SBG, SCC, TPF, MPC). At the diurnal scale, the R-s values of SBG, TPF and MPC had different relationships with T-s and M-s. during different stages of the growing season. The R-s values of SBG and TPF also had different diurnal patterns during these stages. Conclusion: The critical factors that led to the spatial variation of R-s in the cascade ecosystems in the study area were root biomass, vegetation cover, soil organic matter and M-s. These factors also resulted in the complex relationships of R-s with its influencing factors. The critical factors that led to the temporal variation of R-s in the cascade ecosystems were T-s, M-s. and plant growth. Cascade ecosystems should be considered when estimating the terrestrial carbon balance, and caution must be taken to select the appropriate relation between R-s and its influencing factors. Results of this study will contribute to the accurate estimation of terrestrial carbon emissions and our understanding of the carbon cycle process.