Abstract The objective of this study is to evaluate the sensitivity of Distributed Temperature Sensing (DTS) data to detect CO2 leakage along an injection well casing. This paper describes the relationship between the CO2 leakage rate and temperature response at DTS locations, and the method and numerical model used for understanding such a relationship. The uncertainties in the parameters are propagated to the interpretation of DTS measurements, which may lead to false positive and false negative identifications of CO2 leakage. We propose to identify CO2 leakage by analyzing both temperature history plots at selected vertical locations as well as the vertical temperature profiles at different times. The analysis should be combined with numerical simulations for the estimation of leakage rates. In addition, leakage needs to be confirmed by data from a warm-up test (temperature recovery after injection of cold CO2 is stopped) to minimize the probability of false identifications.