High precision GPS measurements have been used to measure surface deformation associated with CO2 injection at an enhanced oil recovery (EOR) field in South Texas. We describe a filtering procedure to reduce noise associated with seasonal hydrologic effects, achieving post-filter precisions of better than 2 mm and 3 mm in horizontal and vertical components respectively. A model assuming uniform pressurization of a thin horizontal disc-shaped pressure source in an elastic half-space fits the surface deformation data quite well. The model predicts a location of the pressurized source consistent with injection locations, and suggests minimal horizontal migration of the CO2 fluid during the test period. Our results suggest that a sparse network of dual frequency GPS receivers can be used to augment sub-surface data for monitoring, verification and accounting (MVA) activities associated with carbon capture, utilization and storage, deriving independent constraints on pressure changes in the reservoir at depth as well as CO2 plume migration.