CO2 and heat fluxes were measured over a six-week period (09/08/2006 to 10/24/2006) by the eddy covariance (EC) technique at the Horseshoe Lake tree kill (HLTK), Mammoth Mountain, CA, a site with complex terrain and high, spatially heterogeneous CO2 emission rates. EC CO2 fluxes ranged from 218 to 3500 g m-2 d-1 (mean=1346 g m-2 d-1). Using footprint modeling, EC CO2 fluxes were compared to CO2 fluxes measured by the chamber method on a grid repeatedly over a 10-day period. Half-hour EC CO2 fluxes were moderately correlated (R2=0.42) with chamber fluxes, whereas average daily EC CO2 fluxes were well correlated (R2=0.70) with chamber measurements. Average daily EC CO2 fluxes were correlated with both average daily wind speed and atmospheric pressure; relationships were similar to those observed between chamber CO2 fluxes and the atmospheric parameters over a comparable time period. Energy balance closure was assessed by statistical regression of EC energy fluxes(sensible and latent heat) against available energy (net radiation, less soil heat flux). While incomplete (R2=0.77 for 1:1 line), the degree of energy balance closure fell within the range observed in many investigations conducted in contrasting ecosystems and climates. Results indicate that despite complexities presented by the HLTK, EC can be reliably used to monitor background variations in volcanic CO2 fluxes associated with meteorological forcing, and presumably changes related to deeply derived processes such as volcanic activity.