Exploitation of the Pauzhetsky geothermal field started in 1966 with a 5 MW, power plant. A hydrogeological model of the Pauzhetsky field has been developed based on an integrated analysis of data on lithological units, temperature, pressure, production zones and natural discharge distributions. A one-layer "well by well" model with specified vertical heat and mass exchange conditions has been used to represent the main features of the production reservoir. Numerical model development was based on the TOUGH2 code [Pruess, 1991. TOUGH2-A General Purpose Numerical Simulator for Multiphase Fluid and Heat Flow, Lawrence Berkeley National Laboratory Report, Berkeley, CA; Pruess et al., 1999. TOUGH2 User's Guide, Version 2.0, Report LBNL-43134, Lawrence Berkeley National Laboratory, Berkeley, CA] coupled with tables generated by the HOLA wellbore simulator [Aunzo et al., 1991. Wellbore Models GWELL, GWNACL, and HOLA, Users Guide, Draft, 81 pp.]. Lahey Fortran-90 compiler and computer graphical packages (Didger-3, Surfer-8, Grapher-3) were also used to model the development process. The modeling study of the natural-state conditions was targeted on a temperature distribution match to estimate the natural high-temperature upflow parameters: the mass flow-rate was estimated at 220kg/s with enthalpy of 830-920 kJ/kg. The modeling study for the 1964-2000 exploitation period of the Pauzhetsky geothermal field was targeted at matching the transient reservoir pressure and flowing enthalpies of the production wells. The modeling study of exploitation confirmed that "double porosity" in the reservoir, with a 10-20% active volume of "fractures," and a thermo-mechanical response to reinjection (including changes in porosity due to compressibility and expansivity), were the key parameters of the model. The calibrated model of the Pauzhetsky geothermal field was used to forecast reservoir behavior under different exploitation scenarios for the central part of the field. The basic scenario assumes that the wellhead pressures of the eight exploitation wells and the injection rates of the three reinjection wells are maintained at the same conditions as of December 2000. In the base case, the model predicts a 12% decline in steam production rate (at 2.7 bar) during the next 30 years, even as the steam supply for the 5 MWe power plant is maintained. The modeling study confirmed that 30-60 kg/s is an optimal reinjection rate. An increase in the exploitation load has no significant effect on steam production from the central section of the Pauzhetsky field during the 30-year exploitation period; load doubling (eight additional exploitation wells) leads to a mere 16-27% increase in steam production. (C) 2004 CNR. Published by Elsevier Ltd. All rights reserved.