e log data from 62 wells near the center of the field were studied to characterize the porosity,
permeability, and water saturation of the Lance reservoirs. The logs were environmentally corrected and
normalized, shale volume and porosities were calculated, water saturations were determined by the dual
water model, and net pay was calculated using field-specific pay criteria. Ultimate gas recovery per well was
estimated by decline curve analysis of monthly production data.
Within the upper 2500 ft (760 m) of the Lance Formation, which includes the entire productive interval in
nearly all wells, the average well has 1000 ft (30 m) of net sandstone, having an average porosity of 6.4%. The
average permeability of all sandstones, estimated from core data-derived equations, is an astonishingly low
6 Ad. The average water saturation of all sandstones is 45%.
Net pay criteria were determined from cumulative storage-capacity and cumulative flow-capacity plots.
Although the average sandstone may have only 6% porosity, the low-porosity sandstones contribute an
insignificant fraction of the reservoir flow capacity. We estimate that more than 95% of the flow capacity
is from sandstones with greater than 6% porosity. A small percentage of high-porosity (>10%) and high-
permeability rocks dominate the flow behavior of the reservoir and are probably critical to economic pro-
duction. Using 6% porosity as an absolute net pay cutoff, the average net pay thickness at Jonah is 440 ft
(130 m), with 9.3% porosity and 33% water saturation. The estimated average permeability of net pay is
25 Ad. Estimated ultimate recovery per well is approximately 4 bcf gas on current 40-ac (0.16-km
2
) well
spacing.
INTRODUCTION
Jonah field, located in Sublette County, Wyoming,
was one of the largest onshore natural gas discoveries of
the 1990s in the United States. Although many aspects of
this field challenge our perceptions of what is required
to form a giant gas accumulation, one of the most re-
markable aspects are the petrophysics of the productive
sandstones. The reservoir sandstones at Jonah have aver-
age porosities and permeabilities that would be consid-
ered the sealing facies in many gas fields, which imme-
diately raises questions of how pay is defined and what
constitutes valid criteria for differentiating pay from non-
pay intervals.
The purpose of this chapter is to document the petro-
physical properties of the Lance sandstones at Jonah
Jonah Field: Case Study of a Giant Tight-Gas Fluvial Reservoir, J. W. Robinson and K. W. Shanley, eds.: AAPG Studies in Geology 52 and Rocky
Mountain Association of Geologists 2004 Guidebook
215 field, evaluate criteria for defining net pay, and propose
a methodology to determine net pay cutoffs.
METHODS
Data Sources
Wire-line log data from 62 wells near the center of
Jonah field were selected for study out of a total of
nearly 400 wells drilled by the end of 2002 (Figure 1).
Although not every well in the field was included in the
summations used for this chapter, we have run compa-
rable calculations on many other wells across the field
area and consider the wells in the central field area to be
a representative sample for the productive Lance For-
mation. The data set includes several wells in the upper-
most and the lowermost quartiles of the reserve size
distribution of the field.
The basic analysis procedure we used involves the
following steps, each of which is described in the fol-
lowing sections:
(1)
Acquire data from either vendor field tapes or high-
accuracy digitization of paper log prints;
(2)
Merge log runs and depth shift curves between log-
ging passes, if required;
(3)
Import and depth shift core data;
(4)
Apply environmental corrections and normalize po-
rosity and gamma-ray logs;
(5)
Compute shale volume from the gamma ray;
(6)
Compute total porosity and shale-corrected (effec-
tive) porosity from the density, neutron, and sonic
logs;
Figure 1. Structure contour map on the base of the Tertiary Fort Union Formation. Solid circles denote the location of wells used in
the study. The study area is shown by the outline. Contour interval = 100 ft (30 m).
Suzanne G. Cluff and Robert M. Cluff
216
AAPG Studies in Geology 52 and Rocky Mountain Association of Geologists 2004 Guidebook (7)
Compute water saturation by dual water model;
(8)
Calculate net pay using field-specific net pay cutoffs.
For this study, most of the logs were digitized from
paper copies. Nearly all wells have been logged with
an array induction resistivity tool and a compensated
neutron-compensated density-gamma ray tool. Only a
few sonic logs have been run in the field.
All logging run data were merged and depth shifted
into alignment with the resistivity logs, which were as-
sumed to be the reference depth in all cases. Because a
large portion of the wells at Jonah were logged with a
Schlumberger Platform Express tool, which includes the
resistivity and porosity measurements all on a single tool
string, relatively few wells require significant depth shifting.
Core data were available for four wells in the field,
including porosity and permeability at net overburden
stress conditions. The core data were imported as ASCII
files, merged with the open-hole logs, then depth shifted
into alignment with the logs by correlation with the gamma-
ray and porosity curves. A total of 225 core sample points
were used.
Environmental Corrections and
Log Normalization
Most petrophysical field studies begin by normalizing
the log data to minimize differences caused by random
logging errors and other noise in the data set. This noise
results from a variety of causes. Logging tools built by
different vendors can have different responses to iden-
tical formation conditions. The study wells were drilled
between 1995 and 2002 and logged by several different
wire-line companies. Different generations of logging
tools and different processing algorithms were used, de-
pending on the vendor and when the well was logged.
Logging tool response is also strongly affected by bore-
hole and mud conditions. Environmental correction rou-
tines are used to compensate for most of these environ-
mental effects, but many of the key variables are poorly
known or are not measured, and all environmental cor-
rection algorithms assume ideal, commonly unrealistic
conditions (e.g., a perfectly centered tool in a cylindrical
borehole). The environmental corrections applied to this
data included (1) borehole size and mud weight correc-
tion to the gamma-ray logs; (2) matrix corrections to the
neutron porosity; and (3) invasion (or tornado chart)
corrections to the three-curve induction resistivity logs.
Array induction logs were not environmentally corrected,
and the deepest array curve was used as an approxima-
tion of true formation resistivity. Other common environ-
mental corrections applied by the logging vendor were
determined to be negligible for the borehole sizes and
logging conditions at Jonah field or are compensated for
by the normalization procedure described below.
Normalization is a process used to reduce the resid-
ual errors by comparing the log response in a zone of
known and constant properties with the expected re-
sponse in that lithology. For example, a clean tight lime-
stone or an anhydrite bed might have consistent density,
neutron porosity, and sonic transit time over a broad area
and would be expected to have the same log response
in every well. In practice, the log responses differ, be-
cause the wells were logged with different tools, bore-
hole conditions varied, the environmental corrections
did not exactly match the actual conditions during logg-
ing, a tool might have been miscalibrated, or the logging
operator simply made an error. The normalization pro-
cedure to reduce these errors involves either applying
an offset (bulk shift) to the log curve or applying a gain
correction between two known end points. In some
cases, both adjustments will be made to the observed
response to achieve the expected response in the con-
stant property intervals. In all cases, we assume the log
response is linear over the range of interest and can be
approximated by a simple slope-intercept method.
Unfortunately, normalization in the Green River basin is
problematic because there are no zones with uniform li-
thology over a large geographic area. Many log analysts
use statistical techniques to reduce the variation between
wells toward a field or area average, but most of these
methods run the risk of eliminating actual geologic varia-
tion (signal) as opposed to random noise between logging
jobs. As an alternative, we make the assumption that the
minimum porosity over a field size area will be constant.
We believe that this assumption is reasonable, because
burial depth and time-temperature history are the major
controls on porosity loss, whereas mineralogical varia-
tions are secondary and tend to be minor over the area of
a field. Porosity variations are minor in the Lance Forma-
tion in the immediate vicinity of Jonah field. Because the
average porosity of an interval varies with the sand to shale
(or net/gross) ratio, we made no attempt to normalize the
average properties of the logs; only the minimum ob-
served porosity in clean sandstones and siltstones is used.
Neutron-porosity and bulk-density logs were normal-
ized to match a suite of eight type logs. Composite his-
tograms for each measurement were constructed from
the data over the top 1000 ft (300 m) of the Lance Formation
in these wells (Figures 2, 3). The extr