Validating turbulence models for cooling supercritical carbon dioxide (sCO2) in a horizontal pipe is challenging due to the lack of experimental data with spatially resolved local temperature measurements. Although many variables may be present to cause discrepancies between numerical and experimental data, this study focuses on how the choice of reference temperatures (both wall reference temperature and fluid bulk reference temperature) when calculating the heat transfer coefficient influences turbulence-model validation results. While it may seem straightforward to simply use the same parameters as the experimental setup, this has not been observed in practice. In this work, numerical simulations are performed for cooling sCO2 in a horizontal pipe for p = 8 MPa, d = 6 mm, G = 200, and 400 kg/(m2s), and qw = 12, 24, and 33 kW/m2. Local and average heat transfer coefficients with different reference temperatures, found to be frequently used in the literature, are presented and compared with commonly used experimental data. It was found that the choice of reference temperatures has a significant influence on the results of the numerical validation. Historically, the higher heat flux cases have been more difficult to validate, theorized due to using reference temperatures differing from the experiment; however, good agreement was found here using the reference temperatures that most closely matched the experiment. This not only highlights the need for careful selection of reference temperatures in simulations, but also the importance of clearly defining the reference temperature employed when reporting experimental results.