Carbon dioxide (CO2) is likely to reach the bottom of injection wells at a colder temperature than that of the storage formation, causing cooling of the rock. This cooling, together with overpressure, tends to open up fractures, which may enhance injectivity. We investigate cooling effects on injectivity enhancement by modeling the In Salah CO2 storage site and a theoretical, long-term injection case. We use stress-dependent permeability functions that predict an increase in permeability as the effective stress acting normal to fractures decreases. Normal effective stress can decrease either due to overpressure or cooling. We calibrate our In Salah model, which includes a fracture zone perpendicular to the well, obtaining a good fitting with the injection pressure measured at KB-502 and the rapid CO2 breakthrough that occurred at the observation well KB-5 located 2 km away from the injection well. CO2 preferentially advances through the fracture zone, which becomes two orders of magnitude more permeable than the rest of the reservoir. Nevertheless, the effect of cooling on the long-term injectivity enhancement is limited in pressure dominated storage sites, like at In Salah, because most of the permeability enhancement is due to overpressure. However, thermal effects enhance injectivity in cooling dominated storage sites, which may decrease the injection pressure by 20%, saving a significant amount of compression energy all over the duration of storage projects. Overall, our simulation results show that cooling has the potential to enhance injectivity in fractured reservoirs. © 2017 Elsevier Ltd This work was supported by the Assistant Secretary for Fossil Energy, Office of Natural Gas and Petroleum Technology, through the National Energy Technology Laboratory, under the U.S. Department of Energy Contract No. DE-AC02-05CH11231. V. Vilarrasa acknowledges financial support from the “TRUST" project (European Community's Seventh Framework Programme FP7/2007-2013 under grant agreement n. 309607) and from “FracRisk" project (European Community's Horizon 2020 Framework Programme H2020-EU.3.3.2.3 under grant agreement n. 640979). A.P. Rinaldi is currently funded by Swiss National Science Foundation (SNSF) Ambizione Energy grant (PZENP2_160555). The authors would like to thank the In Salah JIP and their partners BP, Statoil, and Sonatrach for providing field data and technical input over the past 10 years as well as for financial support during LBNL's participation in the In Salah JIP, 2011–2013 Peer reviewed