Moridis and Reagan [131] showed that depressurization-induced dissociation appears to be the most promising gas production strategy in Class 2 deposits. They proposed new well configurations to maximize production and alleviate a persistent problem of substantial secondary hydrate (or ice) formation in a narrow zone (r < 10 m) around the well. Using the properties and conditions representative of the Tigershark formation and producing at an initial constant mass rate of QM = 19.2 kg/s (=10,000 BPD), Moridis and Reagan [131] showed (Fig. 13) that (a) QM cannot be

Fig. 11 Gas production from a class 1 hydrate deposit. Left: evolution of (a) the rate of CH4 release from hydrate dissociation, (b) the rate of CH4 production at the well, and (c) the corresponding rate replenishment ratio over the 30-year production period. Right: evolution of (a) the cumulative CH4 volume released from hydrate dissociation, (b) the produced CH4 volume at the well, and (c) the corresponding volume replenishment ratio over the 30-year production period [129]

maintained constant during the production period (but has to decline), (b) the gas production rate is highly variable, (c) it is encumbered by a long initial lead time during which little gas is produced, but (d) it can reach levels as high as QP=4.8 x 105 m3/day (=17 MMSCFD), with an average gas production Qav over the 5,660-day period of simulation is about 2.2 x 105 m3/day (=7.8 MMSCFD). This study showed very high recovery from hydrate deposits, although economic and geomechanical considerations may limit total recovery. Similar results were obtained from the study of an oceanic Class 2 deposit in the Ulleung Basin of the Korean East Sea [136] and (b) a permafrost-associated deposit in the North slope [133, 134], leading to the observation that QP on the order of several MMSCFD is attainable in Class 2 deposits despite significant differences in reservoir tempera­ture, HBS thickness, and salinity.

The use of horizontal wells can substantially improve gas production from such deposits and reduce the initial period of low QP [137]. Conversely, Moridis and Kowalsky [128] determined that QP was too low to justify considering such accumula­tions as viable targets in the presence of permeable boundaries and/or with a deep WZ.