Abstract. Atmospheric gravity waves (GWs) are essential for the dynamics of the middle atmosphere. Recent studies have shown that these waves are also important for the thermosphere/ionosphere (T/I) system. Via vertical coupling, GWs can significantly influence the mean state of the T/I system. However, the penetration of GWs into the T/I system is not fully understood in modeling as well as observations. In the current study, we analyze the correlation between GW momentum fluxes observed in the middle atmosphere (30-90 km) and GW-induced perturbations in the T/I. In the middle atmosphere, GW momentum fluxes are derived from temperature observations of the Sounding of the Atmosphere using Broadband Emission Radiometry (SABER) satellite instrument. In the T/I, GW-induced perturbations are derived from neutral density measured by instruments on the Gravity field and Ocean Circulation Explorer (GOCE) and CHAllenging Minisatellite Payload (CHAMP) satellites. We find generally positive correlations between horizontal distributions at low altitudes (i.e., below 90 km) and horizontal distributions of GW-induced density fluctuations in the T/I (at 200 km and above). Two coupling mechanisms are likely responsible for these positive correlations: (1) fast GWs generated in the troposphere and lower stratosphere can propagate directly to the T/I and (2) primary GWs with their origins in the lower atmosphere dissipate while propagating upwards and generate secondary GWs, which then penetrate up to the T/I and maintain the spatial patterns of GW distributions in the lower atmosphere. The mountain-wave related hotspot over the Andes and Antarctic Peninsula is found clearly in observations of all instruments used in our analysis. Latitude-longitude variations in the summer midlatitudes are also found in observations of all instruments. These variations and strong positive correlations in the summer midlatitudes suggest that GWs with origins related to convection also propagate up to the T/I. Different processes which likely influence the vertical coupling are GW dissipation, possible generation of secondary GWs, and horizontal propagation of GWs. Limitations of the observations as well as of our research approach are discussed.