In the later part of the 1990s, a large die-off of desert shrubs occurred over an approximately 1 km2 area in the northwestern section of the Dixie Valley (DV) geothermal field. This paper reports results from accumulation-chamber measurements of soil CO2 flux from locations in the dead zone and stable isotope and chemical data on fluids from fumaroles, shallow wells, and geothermal production wells within and adjacent to the dead zone. A cumulative probability plot shows three types of flux sites within the dead zone: Locations with a normal background CO2 flux (7 g m-2 day-1); moderate flux sites displaying "excess" geothermal flux; and high flux sites near young vents and fumaroles. A maximum CO2 flux of 570 g m-2 day-1 was measured at a location adjacent to a fumarole. Using statistical methods appropriate for lognormally distributed populations of data, estimates of the geothermal flux range from 7.5 t day-1 from a 0.14-km2 site near the Stillwater Fault to 0.1 t day-1 from a 0.01 -km2 location of steaming ground on the valley floor. Anomalous CO2 flux is positively correlated with shallow temperature anomalies. The anomalous flux associated with the entire dead zone area declined about 35% over a 6-month period. The decline was most notable at a hot zone located on an alluvial fan and in the SG located on the valley floor. Gas geochemistry indicates that older established fumaroles along the Stillwater Fault and a 2-year-old vent in the lower section of the dead zone discharge a mixture of geothermal gases and air or gases from air-saturated meteoric water (ASMW). Stable isotope data indicate that steam from the smaller fumaroles is produced by ??? 100??C boiling of these mixed fluids and reservoir fluid. Steam from the Senator fumarole (SF) and from shallow wells penetrating the dead zone are probably derived by 140??C to 160??C boiling of reservoir fluid. Carbon-13 isotope data suggest that the reservoir CO2 is produced mainly by thermal decarbonation of hydrothermal calcite in veins that cut reservoir rocks. Formation of the dead zone is linked to the reservoir pressure decline caused by continuous reservoir drawdown from 1986 to present. These reservoir changes have restricted flow and induced boiling in a subsurface hydrothermal outflow plume extending from the Stillwater Fault southeast toward the DV floor. We estimate that maximum CO2 flux in the upflow zone along the Stillwater Fault in 1998 was roughly seven to eight times greater than the pre-production flux in 1986. The eventual decline in CO2 flux reflects the drying out of the outflow plume. Published by Elsevier Science B.V.