Alternate conceptual models to explain near-surface heat-flow observations in the central Oregon Cascade Range involve (1) an extensive mid-crustal magmatic heat source underlying both the Quaternary arc and adjacent older rocks or (2) a narrower deep heat source which is flanked by a relatively shallow conductive heat-flow anomaly caused by regional ground-water flow (the lateral-flow model). Relative to the mid-crustal heat source model, the lateral-flow model suggests a more limited geothermal resource base, but a better-defined exploration target. We simulated ground-water flow and heat transport through two cross sections trending west from the Cascade range crest in order to explore the implications of the two models. The thermal input for the alternate conceptual models was simulated by varying the width and intensity of a basal heat-flow anomaly and, in some cases, by introducing shallower heat sources beneath the Quaternary arc. Near-surface observations in the Breitenbush Hot Springs area are most readily explained in terms of lateral heat transport by regional ground-water flow; however, the deep thermal structure still cannot be uniquely inferred. The sparser thermal data set from the McKenzie River area can be explained either in terms of deep regional ground-water flow or in terms of a conduction-dominated system, with ground-water flow essentially confined to Quaternary rocks and fault zones.