Monitoring gas and heat emissions at Norris Geyser Basin, Yellowstone National Park, USA based on a combined eddy covariance and Multi-GAS approach

We quantified gas and heat emissions in an acid-sulfate, vapor-dominated area (0.04-km²) of Norris Geyser Basin, located just north of the 0.63 Ma Yellowstone Caldera and near an area of anomalous uplift. From 14 May to 3 October 2016, an eddy covariance system measured half-hourly CO₂, H₂O and sensible (H) and latent (LE) heat fluxes and a Multi-GAS instrument measured (1 Hz frequency) atmospheric H₂O, CO₂ and H₂S volumetric mixing ratios. We also measured soil CO₂ fluxes using the accumulation chamber method and temperature profiles on a grid and collected fumarole gas samples for geochemical analysis. Eddy covariance CO₂ fluxes ranged from - 56 to 885 g m^{- 2} d^{- 1}. Using wavelet analysis, average daily eddy covariance CO₂ fluxes were locally correlated with average daily environmental parameters on several-day to monthly time scales. Estimates of CO₂ emission rate from the study area ranged from 8.6 t d^{- 1} based on eddy covariance measurements to 9.8 t d^{- 1} based on accumulation chamber measurements. Eddy covariance water vapor fluxes ranged from 1178 to 24,600 g m^{- 2} d^{- 1}. Nighttime H and LE were considered representative of hydrothermal heat fluxes and ranged from 4 to 183 and 38 to 504 W m^{- 2}, respectively. The total hydrothermal heat emission rate (H + LE + radiant) estimated for the study area was 11.6 MW and LE contributed 69% of the output. The mean ± standard deviation of H₂O, CO₂ and H₂S mixing ratios measured by the Multi-GAS system were 9.3 ± 3.1 parts per thousand, 467 ± 61 ppmv, and 0.5 ± 0.6 ppmv, respectively, and variations in the gas compositions were strongly correlated with diurnal variations in environmental parameters (wind speed and direction, atmospheric temperature). After removing ambient H₂O and CO₂, the observed variations in the Multi-GAS data could be explained by the mixing of relatively H₂O-CO₂-H₂S-rich fumarole gases with CO₂-rich and H₂O-H₂S-poor soil gases. The fumarole H₂O/CO₂ and CO₂/H₂S end member ratios (101.7 and 27.1, respectively, on average) were invariant during the measurement period and fell within the range of values measured in direct fumarole gas samples. The soil gas H₂O/CO₂ end member ratios ( 15-30) were variable and low relative to the fumarole end member, likely resulting from water vapor loss during cooling and condensation in the shallow subsurface, whereas the CO₂/H₂S end member ratio was high ( 160), presumably related to transport of CO₂-dominated soil gas emissions mixed with trace fumarolic emissions to the Multi-GAS station. Nighttime eddy covariance ratios of H₂O to CO₂ flux were typically between the soil gas and fumarole end member H₂O/CO₂ ratios defined by Multi-GAS measurements. Overall, the combined eddy covariance and Multi-GAS approach provides a powerful tool for quasi-continuous measurements of gas and heat emissions for improved volcano-hydrothermal monitoring.