A 2D reactive transport model of the Dixie Valley geothermal area in Nevada, USA was developed to assess fluid flow pathways and fluid rock interaction processes. The model includes two major normal faults and the incorporation of a dual continuum domain to simulate the presence of a small-scale thermal spring being fed by a highly permeable but narrow fracture zone. Simulations were performed incorporating fluid flow, heat conduction and advection, and kinetic mineral-water reactions. Various solute geothermometry methods were applied to simulated spring compositions, to compare estimated reservoir temperatures with "true" modeled reservoir temperatures, for a fluid ascending the simulated fracture and cooling on its way to the surface. Under the modeled conditions (cooling but no mixing or boiling), the classical Na-K(-Ca) geothermometers performed best because these are least affected by mineral precipitation upon cooling. Geothermometry based on computed mineral saturation indices and the quartz geothermometer were more sensitive to re-equilibration upon cooling, but showed good results for fluid velocities above ca. 0.1 m/d and a reactive fracture surface area 1-2 orders of magnitude lower than the corresponding geometric surface area. This suggests that such upflow rates and relatively low reactive fracture surface areas are likely present in many geothermal fields. The simulations also suggest that the presence of small-scale fracture systems having an elevated permeability of 10-12 to 10(-10) m(2) can significantly alter the shallow fluid flow regime of geothermal systems. For the Dixie Valley case, the model implies that such elevated permeabilities lead to a shallow (less than 1 km) convection cell where superficial water infiltrates along the range front normal fault and connects the small-scale geothermal spring through basin filling sediments. Furthermore, we conclude that a fracture permeability on the order of 10(-12) m(2) may lead to near surface temperature >100 degrees C whereas a permeability of 10-10 m2 is not realistic because this permeability led to extreme upflow velocities and to a short-circuit of the regional fault zone. (c) 2013 Elsevier Ltd. All rights reserved.