Extreme soil moisture conditions is more likely to cause deviations between soil respiration and ecosystem...

Ecosystem respiration (R_eco) is the main source of terrestrial carbon flux, and soil respiration (R_s) is its important component. Numbers of studies have investigated the deviation between R_eco and R_s and their interpolation methods. However, the causes of the deviation and how to establish an appropriate interpolation method in ecosystems under extreme hydrology conditions remain poorly understood. Two field sites (a mobile dune and a meadow) in the Horqin Sandy Land in a semiarid region of China were selected to observe R_s and R_eco during the growing season in 2015 and 2016. The results found that the effect of temperature (soil and air temperature (T_s and T_air)) and M_s on the R_s and R_eco varied in different growth periods. We determined that the threshold of the soil moisture (M_s) influencing R_s and R_eco in the meadow could be depicted by the field and capillary water holding capacity, respectively. Herein, A method to piecewise interpolate missing R_eco data in ecosystems under extreme soil moisture conditions according to the M_s and T_air at different phenological phases was proposed. Moreover, R_s/R_eco of the mobile dunes was underestimated because the influences of night condensate and vegetation in low-lying areas were ignored. R_s/R_eco of the meadows was overestimated because the surface M_s was regulated by groundwater. Therefore, non-uniform underlying surfaces and patch vegetation distribution in ecosystems stressed by M_s, and the slight changes in ground elevation in ecosystems inhibited by M_s are the main reasons for the deviation between R_s and R_eco. Our findings provide a method for interpolating missing R_eco data and identifying the sources of discrepancy between these two fluxes.