Direct and remote measurements of volcanic gas composition, SO2 flux, and eruptive SO2 mass from Bezymianny Volcano were acquired between July 2007 and July 2010. Chemical composition of fumarolic gases, plume SO2 flux from ground and air-based ultraviolet remote sensing (FLYSPEC), and eruptive SO2 mass from Ozone Monitoring Instrument (OMI) satellite observations were used along with eruption timing to elucidate magma processes and subsurface conditions, and to constrain total volatile flux. Bezymianny Volcano had five explosive magmatic eruptions between May 2007 and June 2010. The most complete volcanic gas datasets were acquired for the October 2007, December 2009, and May 2010 eruptions. Gas measurements collected prior to the October 2007 eruption have a relatively high ratio of H2O/CO2 (81.2), a moderate ratio of CO2/S (5.47), and a low ratio of S/HCl (0.338), along with moderate SO2 and CO2 fluxes of 280 and 980 t/d, respectively, and high H2O and HCl fluxes of ~ 45,000 and ~ 440 t/d, respectively. These results suggest degassing of shallow magma (consistent with observations of lava extrusion) along with potential minor degassing of a deeper magma source. Gas measurements collected prior to the December 2009 eruption are characterized by relatively low H2O/CO2 (4.13), moderate CO2/S (6.84), and high S/HCl (18.7) ratios, along with moderate SO2 and CO2 fluxes of ~ 220 and ~ 1000 t/d, respectively, and low H2O and HCl fluxes of ~ 1700 and ~ 7 t/d, respectively. These trends are consistent with degassing of a deeper magma source. Fumarole samples collected ~ 1.5 months following the May 2010 eruption are characterized by high H2O/CO2 (63.0), low CO2/S (0.986), and moderate S/HCl (6.09) ratios. These data are consistent with degassing of a shallow, volatile-rich magma source, likely related to the May eruption. Passive and eruptive SO2 measurements are used to calculate a total annual SO2 mass of 109 kt emitted in 2007, with passive emissions comprising ~ 87–95% of the total. Total annual volatile masses for the study period are estimated to range from 1.1 × 106 to 18 × 106 t/year. Annual CO2 masses are ~ 8 to 40 times larger than can be explained by degassing of dissolved CO2 within eruptive magma, suggesting that the eruptive magma contained a significant quantity of exsolved volatiles sourced either from the eruptive melt or unerupted magma at depth. Variable total volatile fluxes ranging from ~ 3000 t/d in 2009 to ~ 49,000 t/d in 2007 are attributed to variations in the depth of gas exsolution and separation from the melt under open-system degassing conditions. We propose that exsolved volatiles are quickly transported to the surface from ascending magma via permeable flow through a bubble and/or fracture network within the conduit and thus retain their equilibrium composition at the time of segregation from melt. The composition of surface CO2 and H2O emissions from 2007 to 2009 are compared with modeled exsolved fluid compositions for a magma body ascending from entrapment depths to estimate depth of fluid exsolution and separation from the melt. We find that at the time of sample collection magma had already begun ascent from the mid-crustal storage region and was located at maximum depths of ~ 3.7 km in August 2007, approximately 2 months prior to the next magmatic eruption, and ~ 4.6 km in July of 2009 approximately five months prior to the next magmatic eruption. These findings suggest that the exsolved gas composition at Bezymianny Volcano may be used to detect magma ascent prior to eruption.