Summary The five liquid-dominated geothermal fields under production in Japan (Otake, Onuma, Onikobe, Hatchobaru, and Kakkonda) represent more than half the liquiddominated fields now in production worldwide. All five have total liquid reinjection programs. and thus collectively represent a major fraction of all full-scale water reinjection operations in the world. The implications of the Japanese reinjection experience are therefore of extreme importance to future liquid-dominated geothermal development. The major impact of the experience in these fields is that, with the exception of Otake, the reinjection wells show a rapid interference with the production wells. In most cases, the reinjected water travels to the production zone within a very short time, and there is consequently a discernible thermal drawdown. Tracer tests in most of the fields indicate flow paths of very high permeability, probably along fractures. For this reason, resource estimates based on assessed total reservoir volume are likely to be gross overestimates of the producible energy. The Japanese industry has made considerable advances in the determination of this kind of behavior, and their results suggest that the current geothermal reservoir concepts need some rethinking. Introduction Reinjection of water into geothermal reservoirs during utilization is intended to serve the dual purposes of waste water disposal and improved resource recovery. To assess the relative importance of these two purposes accurately, note that reservoir maintenance by reinjection is a controversial subject, and, in actual field cases to date, water has been reinjected solely for disposal purposes. Currently reinjection is a method of water disposal, and consideration of reservoir maintenance by reinjection is restricted largely to avoiding detrimental effects. Commercial-scale reinjection in geothermal power plants has been practiced at The Geysers, CA; Ahuachapan, El Salvador; Mak Ban, the Philippines; and in five Japanese fields-Otake, Onuma, Onikobe, Hatchobaru. and Kakkonda. With the exception of The Geysers, all these fields are liquid-dominated, with producing steam/water ratios between 1:2 and 1:6. Liquiddominated geothermal utilization represents the most difficult configuration for reinjection since the quantity of hot water to be disposed of is greater than those of steam-dominated geothermal or conventional thermal generation systems. Since almost all future geothermal utilizations will be of the liquid-dominated type, it is important to evaluate current reinjection experience in this type of system. For example, power stations under construction or planning with definite commitments to reinject waste water include East Mesa, Heber, and Brawley, CA; Raft River, ID, Tongonan, the Philippines; Broadlands, New Zealand; and Nigorikawa, Japan. In addition, reinjection has been suggested for existing power stations at Cerro Preto, Mexico; Tiwi, the Philippines; and Wairakei, New Zealand. In terms of its principal role as a waste disposal mechanism, reinjection is clearly a greater problem when the quantity of water to be reinjected (relative to the useful steam produced) is larger. However, there are additional constraints on the physical characteristics of the injected water. Because of dissolved solid deposition effects, the temperature of the injected water is a critical parameter. Thus the evaluation of a reinjection scheme needs to consider the quality, temperature, and chemical nature of the water to be injected. The supposed benefits of reinjection to the reservoir itself usually are attributed to the maintenance of reservoir pressure and mass of fluid in place. In theory, maintaining high reservoir pressures and mass in place should reduce the effects of loss of deliverability, and also those of subsidence or formation collapse. JPT P. 495^