Multi-component tracer tests were conducted at a 10 x 10 m well field located in the Altona Flat Rocks of northern New York. Temperature advancement between two wells separated by 14 m was monitored throughout the well field during progressive heating of the reservoir over 6 d. Multiple approaches to predicting heat transport were applied to field data and compared to temperature rise recorded during reservoir heat-up. Tracer analysis incorporated both an analytical one-dimensional model and a two-dimensional numerical model for non-uniform fractures experiencing “flow-channeling.” Modeling efforts demonstrated that estimating heat transfer surface area using a combined inert/adsorbing tracer (cesium-iodide) could provide accurate forecasting of premature thermal breakthrough. In addition, thermally degrading tracer tests were used to monitor inter-well temperature during progressive reservoir heating. Inert tracers alone were, in general, inadequate in forecasting thermal performance. In fact, moment analysis shows that, mathematically, thermal breakthrough is independent of parameters that primarily influence inert tracers. The most accurate prediction of thermal breakthrough using inert tracer alone was produced by treating hydrodynamic dispersion as a truly Fickian process with known and accurate mathematical models. Under this assumption, inert tracer data was matched by solving an inverse problem for non-uniform fracture aperture. Early arrival of the thermal front was predicted at the production, but was less accurate than using a combined inert/adsorbing tracer test. The spatial distribution of fluid flow paths in the plane of the fracture were identified using computational models, Fiber-Optic Distributed Temperature Sensing (FO-DTS), and Ground Penetrating Radar (GPR) imaging of saline tracer flow paths in the target fracture. Without exception, fluid flow was found to be concentrated in a roughly 1 m wide flow channel directly between the two wells. The following key findings summarize the results of this study: (1) flow channeling resulted in rapid thermal breakthrough; (2) advection-dispersion models for inert tracers cannot calibrate forward-models of thermal performance; (3) adsorption tracers accurately estimated heat transfer surface area; (4) thermally degrading tracer tests identified the migration of a thermal front; and (5) Principal Component Analysis can be used to simplify fracture aperture fields and subsequently aid in solving an inverse problem for non-uniform aperture. In addition, two significant fluid/rock interactions were found: (1) adsorption was rate-limited over the timespan of tracer tests at Altona; and (2) the target fracture catalyzed thermal degradation leading to substantial over-estimates of inter-well temperature.