<p>Biological soil crusts (biocrusts) have been reported to play a considerable role in the global carbon budget through CO<sub>2</sub> uptake by photosynthesis. However, it is still unclear if ecosystems dominated by biocrusts are net carbon sinks. That is mainly because so far, most research have focused on characterizing photosynthesis <em>ex-situ</em>, neglecting the underlying soil component, and particularly the <em>in-situ</em> spatio-temporal variability of soil CO<sub>2</sub> fluxes, which can be substantial. Moreover, it is still unknown how those CO<sub>2</sub> fluxes evolve during the ecological succession of biocrusts and which are the biophysical and geochemical factors that control them. Therefore, this research aimed to (1) identify those factors and (2) describe and explain the evolution of annual cumulative soil CO<sub>2</sub> fluxes over ecological succession in a dryland.</p><p>To this end, we conducted continuous measurements over 2 years of the topsoil CO<sub>2</sub> molar fraction (<em>&#967;</em><sub>s</sub>) in association with below- and aboveground microclimatic variables in 21 locations representative of the ecological succession of biocrusts, characterized by 5 stages: (1) physical depositional crust; (2) incipient cyanobacteria; (3) mature cyanobacteria; (4) lichen community dominated by <em>Squamarina lentigera</em> and <em>Diploschistes diacapsis</em> and (5) lichen community of <em>Lepraria isidiata</em>. Those measurements were also conducted under plants (<em>Macrochloa tenacissima</em>, <em>Salsola genistoides</em>, and <em>Lygeum spartum</em>). Using spatio-temporal statistics, an explanatory model of <em>&#967;</em><sub>s </sub>dynamics was calibrated on the first year of data and cross-validated to test prediction on the second year. An explanatory model of annual cumulative fluxes was also developed.</p><p>The biocrust type, soil water content (<em>&#977;</em>) and temperature (<em>T</em><sub>s</sub>) and interactions between those variables explained and predicted efficiently the <em>&#967;</em><sub>s </sub>dynamics. Among those factors, the effect of <em>&#977;</em> was preponderant and dependent on <em>T</em><sub>s</sub> and antecedent soil moisture conditions. The magnitude of the <em>&#977;</em> effect tended to increase in late successional stages, producing greater CO<sub>2</sub> emissions, most likely as a result of progressive soil organic carbon accumulation resulting in greater substrate availability for microbial respiration, and higher porosity enhancing CO<sub>2</sub> diffusion. The calcite content (and potentially indirectly the pH through a buffering effect of CaCO<sub>3</sub>) also played a role in explaining annual cumulative CO<sub>2 </sub>fluxes. Those fluxes were particularly mitigated where CaCO<sub>3</sub> was abundant, apparently due to a substantial nocturnal uptake of atmospheric CO<sub>2 </sub>by soil (influx) throughout the study. The cumulative annual influx represented up to 115% of the cumulative annual efflux, generating a net annual carbon uptake by soil in some locations. Influxes have been increasingly reported recently from drylands soils, which are now regarded as potential carbon sinks. Those influxes have been attributed to different abiotic processes which are still debated. In this ecosystem, in the light of our observations, we assume that a geochemical process of CO<sub>2</sub> dissolution in soil water followed by CaCO<sub>3 </sub>dissolution that consumes CO<sub>2 </sub>might be involved. If this assumption could be verified, this geochemical process consuming CO<sub>2</sub> would need to be separated from biocrust photosynthesis and respiration, when measuring soil surface CO<sub>2</sub> fluxes, to not overestimate and underestimate respectively the biotic contribution to the global carbon budget.</p>