As an important component of ecosystem carbon (C) budgets, soil carbon dioxide (CO2) flux is determined by a combination of a series of biotic and abiotic processes. Although there is evidence showing that the abiotic component can be important in total soil CO2 flux (R-total), its relative importance has never been systematically assessed. In this study, after comparative measurements of CO2 fluxes on sterilized and natural soils, the R-total was partitioned into biotic flux (R-biotic) and abiotic flux (R-abiotic) across a broad range of land-cover types (including eight sampling sites: cotton field, hops field, halophyte garden, alkaline land, reservoir edge, native saline desert, dune crest and interdune lowland) in Gurbantunggut Desert, Xinjiang, China. The relative contribution of R-abiotic to R-total, as well as the temperature dependency and predominant factors for R-total, R-biotic and R-abiotic, were analyzed. Results showed that R-abiotic always contributed to R-total for all of the eight sampling sites, but the degree or magnitude of contribution varied greatly. Specifically, the ratio of R-abiotic to R-total was very low in cotton field and hops field and very high in alkaline land and dune crest. Statistically, the ratio of R-abiotic to R-total logarithmically increased with decreasing R-biotic, suggesting that R-abiotic strongly affected R-total when R-biotic was low. This pattern confirms that soil CO2 flux is predominated by biotic processes in most soils, but abiotic processes can also be dominant when biotic processes are weak. On a diurnal basis, R-abiotic cannot result in net gain or net loss of CO2, but its effect on transient CO2 flux was significant. Temperature dependency of R-total varied among the eight sampling sites and was determined by the predominant processes (abiotic or biotic) of CO2 flux. Specifically, R-biotic was driven by soil temperature while R-abiotic was regulated by the change in soil temperature (Delta T). Namely, declining temperature (Delta T<0) resulted in negative R-abiotic (i.e., CO2 went into soil) while rising temperature (Delta T>0) resulted in positive R-abiotic (i.e., CO2 released from soil). Without recognition of R-abiotic, R-biotic would be overestimated for the daytime and underestimated for the nighttime. Although R-abiotic may not change the sum or the net value of daily soil CO2 exchange and may not directly constitute a C sink, it can significantly alter the transient apparent soil CO2 flux, either in magnitude or in temperature dependency. Thus, recognizing the fact that abiotic component in R-total exists widely in soils has widespread consequences for the understanding of C cycling.