The role of noninductive graupel‐ice charge separation in the early electrification of the July 31, 1984, New Mexico mountain thunderstorm is assessed with a three‐dimensional kinematic cloud model along with multiple Doppler radar and in situ measurements. Observations of the early electrification rate and the electric field distribution are consistent with modeled values that result when the noninductive mechanism works under the influence of convective motions and precipitation growth. An increase in ice particle concentrations and sizes, arising from vigorous precipitation growth, accelerates graupel‐ice collision rates and hence the noninductive charging rate. Growing graupel particles experience increasing fall speed as they rise toward the top of the updraft. The resulting vertical flux convergence of graupel containing charge from previous noninductive collisions is a significant factor in the growth of the main negative charge density. This implies that a combination of air motion, precipitation interaction, and sedimentation contributes to the rapid intensification of storm electric fields. The linear electrification phase, which begins with the cessation of convective growth, is caused by a roughly constant noninductive charging rate and by the separation of negatively charged graupel and positively charged smaller ice particles by differential sedimentation, downdrafts, and horizontal advection in vertically sheared flow. When the sign reversal temperature for noninductive charging is assumed to be −10°C, the model results are characterized by a main negative charge in middle levels and provide the best overall agreement with the in situ field measurements in the July 31 storm. For a sign reversal temperature of −21°C the model results are characterized by a main positive charge center in middle levels, and the electric field polarity is opposite to the polarity measured at low and middle levels of the storm. The model and observational data, combined with findings of some laboratory studies, support the hypothesis that the actual reversal temperature in the July 31 storm is around −10°C. When the model includes the inductive graupel‐droplet charging mechanism in addition to the noninductive mechanism, the effect of inductive charging is secondary to that of noninductive charging. The net effect of adding induction is dissipative. For example, the maximum field strength at the location of the aircraft measurements is slightly less than in the case where the noninductive mechanism acts alone. Weak charge screening layers were found to develop on the boundary of the modeled cloud.