CONTINUOUS-TYPE ACCUMULATIONS

                            
COAL-BED GAS


     The unit of assessment of potential additions to reserves of
coal-bed gas was the play. Coal-bed gas plays were defined as areas within
widespread, commonly basin-wide, accumulations that have similar
conditions of generation, accumulation, and production of gas (fig. 8).
The factors that define the plays include coal-bed thickness,
heterogeneity, depth, and composition; seals; gas content; gas
composition; permeability; pressure regime; structural setting; and
hydrology; as well as conventional trapping mechanisms. It is
postulated that recoverable coal-bed gas reserves are generally
restricted to present-day depths of burial of 500 to 6,000 ft because
of gas content and formation permeability.

     The assessment of potential additions to reserves of coal-bed gas
was based on the estimation of the number and estimated ultimate recoveries
(EUR's) of untested cells within each assessed play. The procedure is
similar, in part, to that used in assessment of continuous-type
accumulations in sandstones, shales, and chalks that is described above
and by Schmoker in the supporting CD-ROM (Gautier and others, 1995).
However, the coal-bed gas assessment relied heavily on production
forecasting using a reservoir simulator. A range of EUR's and
production rates of both gas and water were projected on a "per-well"
and "per-foot-of-coal" basis for each play. The reservoir simulator was
used because: (1) coal-bed gas accumulations are in early stages of
development, and long-term production histories are generally not
available, and (2) other methods, such as decline curve analysis and
material balance are not adequate for expressing the complex movement
of gas and water in coal.

     Input parameters for modeling in this study were based on actual
data, analog information, and judgments of geologists and engineers. To
resolve some of the data uncertainty, particularly for key reservoir
parameters such as gas content and permeability, well production was
compared to that predicted by the simulator for selected wells. This
process is known as "history matching" because the initial data
estimates commonly are adjusted to obtain simulated results that are
characteristic of actual well performance.

     For most plays, long-term production from vertical wells, with 
a variety of completion techniques, was forecasted. In mining areas,
production from wells was modeled. The EUR's predicted by reservoir
simulation were used in conjunction with coal thicknesses to establish
an EUR probability distribution for potentially productive, untested
cells in each play. Seven fractiles (100th, 95th, 75th, 50th, 25th,
5th, and 0th) were provided for the computational model, and the
distribution was assumed to be lognormal. For plays in which no
reservoir simulation was performed, EUR's on a per-foot-of-coal basis
from analog plays were scaled, and a similar procedure was used. The
assessment of coal-bed gas is based on existing technology.