Operators including Aera Energy LLC are evaluating distributed temperature sensing (DTS) systems as an alternative to production logging tools (PLT) and radioactive tracer surveys (RTS) for the determination of zonal allocation factors. If DTS data can be interpreted to provide wellbore velocities, zonal allocation can be determined without well entry and practically in real-time. In comparison to PLT and RTS, DTS could provide zonal allocation at a higher measurement frequency, lower cost, and lower risk to personnel, environment, and the well asset. In wells where casing deformation precludes wireline operations, DTS may be the only viable alternative. The purpose of this work is to investigate whether water injection zonal allocation can be derived from DTS data acquired in thermal tracer injection tests within an uncertainty comparable to RTS. Three waterflood injection wells with RTS and DTS data in Aera Energy"s Belridge field in California are analyzed. The zonal allocation factors derived from the DTS are shown to agree within uncertainty with the RTS-based results. The DTS for two of the wells is interpreted through history matching with an OLGA thermal-fluid model. The DTS for a third well is interpreted using the Temperature Inflection Point (TIP) thermal tracer method proposed in this work. The TIP method is a simpler alternative to history matching and was shown to be theoretically valid for wells with negligible heat transfer at the walls over the duration of the injection test. In this regard the TIP method is less restrictive than the Slopes method, which is currently used in industry but theoretically requires negligible heat transfer at the walls and within the fluid. The TIP method is also shown to be more consistent with the field data of this study. The paper concludes with guidelines on the design and implementation of thermal injection tests and TIP analysis. The most significant practical limitation of the TIP method is its sensitivity to noise in the data, motivating DTS acquisition at the maximum practical resolution, data smoothing, and averaging over multiple test realizations.