Relative seismic velocity changes (dv/v) are being increasingly used to monitor changes in groundwater. However, it remains challenging to verify its implementation in watersheds without direct groundwater well measurements. In this study, we conduct a twelve year dv/v observation in a watershed of the Yellowstone National Park. We find that the seasonal fluctuations and long-term trend of the measured dv/v are highly correlated with the estimated baseflow, which serves as a constraint for groundwater changes. We integrate the estimated baseflow into a poroelastic mechanism and conduct two dv/v simulations based on pressure diffusion. These two simulations closely match with our observed dv/v variations. In addition, our analysis suggest that the measured dv/v is primarily influenced by hydrologic pressure diffusion rather than surface air temperature. We conclude that the baseflow analysis can further enhance the seismic monitoring of groundwater changes.