Marcellus Frac Design Parameters

Appalachian Basin Marcellus Shale: hydraulic fracture design inputs and pump schedule for a single stage of a multi-stage horizontal well

Aegis Solutions Group, LLC | FastFrac | fastfrac.net

Compiled 20 October 2026

What this is. The frac design parameters for the Marcellus example that ships inside FastFrac, entry by entry, with the published range for each and the source behind it. Open the software, click Select design, and this is the design you are looking at.

Where a value sits outside current published practice, or rests on a screening estimate rather than a measurement, the entry says so. Section 9 sets the whole design against what Marcellus operators are pumping today.

What it does not give you is the answer. The fracture geometry, the treating pressures, the horsepower and the placement checks come from running the design. Open the example in FastFrac and press Compute.

1 The Marcellus play, and why this fracture is confined

The barriers above and below the Marcellus, and the height they leave.

The Marcellus is a Devonian organic-rich black shale across the Appalachian Basin, spanning Pennsylvania, West Virginia, Ohio and New York. It is the largest natural gas field in the United States by production.

This design models the fracture as confined to 350 ft. FastFrac takes the containment height as the total vertical extent the fracture is held to, and solves a PKN fracture over that height. A containment height of zero would instead give an unconfined penny growing outward from the perforations in every direction.

The Marcellus is bounded above and below, and both bounds are limestone. The Tully Limestone above is the upper barrier, and in south-western Pennsylvania it sits about 280 ft above the Marcellus. Containment there is a stress contrast: the Tully is tight and has a high Poisson's ratio, which gives it a high minimum horizontal stress. The Onondaga Limestone below is the lower barrier, and a microseismic-monitored horizontal found no fracture propagating downward through it.

The measured heights agree with that. Microseismic from central Pennsylvania puts fracture growth about 200 ft above the treatment zone, terminating at the Tully. Downward growth into the Onondaga is short. Starting from a Marcellus interval of 100 to 150 ft, that puts total fracture height in the 250 to 400 ft band, and 350 ft sits inside it.

Containment is not absolute, and the exception is worth knowing. A DOE and NETL study of six horizontal wells in south-western Pennsylvania found roughly 40 per cent of microseismic events above the Tully, although all of them remained at least 2,000 ft below the overlying producing zones. A separate statistical analysis of microseismic data puts the maximum Marcellus fracture height near 1,900 ft. Those are the tail of the distribution rather than a design basis.

Two cautions on the number itself. Microseismic marks where rock is failing, not where sand is placed, and the propped height is a fraction of it: the Permian HFTS-1 slant core measured propped height at roughly 5 per cent of microseismic height. And a DFIT or microseismic on your own acreage beats any published figure, including this one.

The Marcellus is a Devonian organic-rich black shale, and it is a gas reservoir rather than an oil play. That changes several inputs away from anything a Permian design would use, which Section 4 covers.

2 Design identification

Free text, no effect on the calculation. Name the stage as well as the well, because a single-stage design separated from its context cannot be traced back to anything.

3 Well and completion data for a Marcellus horizontal

Well type - Measured depth - True vertical depth

Horizontal. MD 9,000 ft. TVD 7,000 ft.

The measured depth is to the stage being treated. At 9,000 ft measured against 7,000 ft vertical, this is a stage near the heel of the lateral rather than out toward the toe.

This bears on the surface pressure. Pipe friction scales with the length of pipe the fluid travels, so a toe stage on a long Marcellus lateral sees considerably more surface pressure than this one does. Published Marcellus laterals now run 11,000 to 12,700 ft, so a toe stage could sit near 19,000 ft of measured depth.

Source: published Marcellus TVD runs 5,000 to 9,000 ft, shallower on the western edge and deeper in south-western and north-eastern Pennsylvania. At the Marcellus Shale Energy and Environment Laboratory in West Virginia the formation was cored at 7,445 to 7,557 ft. Median lateral length rose from about 11,100 ft in 2022 to about 12,700 ft in 2024.

Casing - Liner ID

5.5 in. OD, 23 lb/ft, P-110. ID and liner ID 4.670 in.

FastFrac computes the burst rating from OD, weight and grade and returns 14,658 psi for this pipe, the heavier of the two common walls. Published Marcellus practice is 4.5 in. or 5.5 in. production casing in P-110 or Q-125, at 11.6 to 13.5 lb/ft for the 4.5 in. and 20.0 to 23.0 lb/ft for the 5.5 in.

The heavier wall was chosen for margin. Field practice puts the maximum allowable treating pressure at 80 to 85 per cent of the burst rating, which on this string is 11,726 to 12,459 psi. The 20 lb/ft wall rates lower, and on a deeper or faster job that difference decides whether the stage fits a 10,000 psi surface spread or needs a 15,000 psi one.

Source: published Marcellus casing practice. ID and burst follow from API dimensional and 5C3 data once OD, weight and grade are set.

Number of parallel fractures (clusters per stage)

15. Published figures range from 5 to 15.

The rate is divided equally between clusters, so this design puts 5.3 bbl/min into each fracture, which is modest by current standards.

Fifteen is at the top of every published figure found, and the sources conflict. One gives 10 to 15 clusters at 12 to 20 ft spacing, which would make this a 180 to 300 ft stage. Another gives about 5 clusters across 200 ft, and a published reference design uses 8. No Marcellus-specific field study of cluster counts was found.

Source: median Marcellus stage length is about 200 ft, confirmed by two independent studies. The cluster count itself is not independently sourced.

Fracture complexity adjustment

0%. No equivalent-fracture adjustment applied.

The effective count is 15, equal to the primary cluster count. No nonzero percentage has been established for this example. Effective count = primary count x (1 + adjustment / 100). A nonzero value is an assumed sensitivity that changes the fracture calculations while retaining the physical perforation count. It does not simulate a branching network.

Perforations per cluster - Perforation diameter

5 shots per cluster at 0.40 in. Published range 3 to 5 shots, 0.375 to 0.42 in.

Enter shots per cluster; FastFrac multiplies by the cluster count itself, so this design opens 75 holes. Small holes are chosen deliberately in a limited entry design, to generate perforation friction and force the fluid to divide evenly among clusters rather than taking the easiest path.

With 75 holes the perforation friction is modest, which goes with the low rate per cluster. A tighter limited entry design uses fewer, smaller holes to raise the pressure drop and even out the distribution.

Expect the holes to grow during the stage. Pumping several hundred thousand pounds of sand erodes the entrance holes outward by roughly 0.05 to 0.10 in. by the end of the job. FastFrac holds the diameter constant and does not model erosion, so read its perforation drop as the start-of-stage value.

Source: published Marcellus limited-entry practice for entrance hole diameter and erosion. The shot count per cluster is not independently sourced.

4 Marcellus reservoir properties for fracture design

A gas reservoir, so the fluid properties differ from a Permian oil design.

Net pay thickness

350 ft. Published target range 50 to 250 ft.

In FastFrac this is the permeable thickness that leaks, which caps the leakoff area. It is not the fracture height.

This sits above the published target range. Published work puts the organic-rich interval operators land in at 50 to 250 ft, typically around 100 ft, with 50 ft treated as the economic threshold. The gross Marcellus section, though, runs from 50 ft to over 900 ft, and this design describes a thick section where the whole contacted interval takes fluid.

Net pay and containment height are set to the same 350 ft here, and that is deliberate. Net pay sets the leaking share of the fracture face, so when the two are equal the whole face stands in permeable rock and every square foot of it leaks. Set the net pay below the containment height and the fracture grows out of zone, the part outside the pay stops leaking, and the treatment reaches further on the same volume.

Source: published Marcellus thickness and net pay studies, EIA and USGS derived.

Average permeability

0.0004 md, 400 nD. Published range 0.0001 to 0.0008 md.

Matrix permeability of the unstimulated shale, reported at 100 to 800 nanodarcy and averaging 250 to 400 nanodarcy under reservoir stress. This design sits at the top of that average band.

That is roughly a millionth of a conventional sandstone, which is why the well has to be drilled horizontally and fractured in stages. It feeds both leakoff terms, so it sets the fluid efficiency and with it the pad the frac treatment design needs.

Source: published Marcellus matrix permeability measurements, West Virginia University research repository and allied studies.

Average porosity

0.08. Published range 0.06 to 0.10.

Reservoir porosity, not the proppant pack porosity entered later. Published Marcellus porosity is 6 to 10 per cent averaging about 8.5, so this is mid-range. Most of that pore space is micro- and nanopores inside the organic matter rather than between grains.

Source: published Marcellus petrophysical summaries.

Reservoir fluid viscosity - Total compressibility

0.0238 cp and 116.46 x10^-6 psi^-1.

These are gas values, and an oil design's values do not transfer. Methane at reservoir conditions is hundredths of a centipoise, not tenths, and a gas system is far more compressible than an oil one, so it offers much less resistance to fluid leaking into it. Both feed the reservoir-controlled leakoff term.

The FastFrac Toolbox computes both from gas gravity, reservoir pressure and temperature, which is where these came from.

Source: the FastFrac Toolbox, from standard gas property correlations at the stated pressure and temperature.

Reservoir pressure gradient

0.60 psi/ft. Published range 0.45 to 0.65 psi/ft.

The Marcellus is overpressured across most of its productive core, because gas generation raised the pore pressure above the 0.433 psi/ft a normal water column would give. This design sits near the top of the published band, where a core-area well sits.

With the fracture gradient this fixes the differential driving leakoff, so it has to stay below the fracture gradient or the model refuses the run. Here the two are 0.60 and 0.80, giving a 1,400 psi differential.

Source: published Appalachian Basin pore pressure studies.

Reservoir temperature

116 degF. Published basin range roughly 130 to 200 degF.

FastFrac does not read this in the fracture calculation. It seeds the Toolbox and prints on the report, and changing it will not move a single result.

It is low against published values and worth correcting if you adapt this design. Published Appalachian geothermal gradients of 0.011 to 0.022 degF per foot put a 7,000 ft well nearer 130 to 150 degF. The figure matters for fluid selection and breaker scheduling even though this model does not use it.

Source: published Appalachian Basin geothermal gradients.

Drainage area - Hole diameter - Skin

160 acres, 8.75 in., skin 0.

All three feed only the folds of increase, which FastFrac will not report for a multi-cluster stage, so on this design they are inert. Fill them for the record.

5 Marcellus rock mechanical properties

Two values taken from a screening table rather than from a measurement.

The Young's modulus and Poisson's ratio in this design came from the FastFrac Toolbox rock property tables, which are screening estimates by rock type rather than measurements on the Marcellus. The Toolbox says exactly that in the panel beneath those tables. They are a reasonable way to get a design moving and they are not a substitute for logs or core.

Young's modulus

6.25 x10^6 psi. Published Marcellus range 2.5 to 6.5.

This is the Toolbox average for hard, dense sandstone, which spans 5 to 7.5 million psi. It lands at the top of the published Marcellus range of 2.5 to 6.5 million psi, equivalently 18 to 45 GPa.

A stiff rock is the right direction for the Marcellus. Its high silica content makes it stiff and brittle: it resists stretching, breaks cleanly, and yields narrow fractures. A high-clay rock would deform softly and give wide, inefficient fractures that heal.

One caution. The published range almost certainly refers to dynamic moduli from logs, and FastFrac wants static, which runs lower. Taking a dynamic figure straight off a sonic log will overstate the stiffness.

Source: the FastFrac Toolbox Young's modulus table, hard dense sandstone row. Published Marcellus elastic property studies give the comparison range.

Poisson's ratio

0.15. Published Marcellus range 0.18 to 0.28.

This is the Toolbox value for hard sandstone. It sits below the published Marcellus range.

The direction is supported even though the value is outside the band. The MSEEL logs describe the Marcellus pay zone as having an abnormally low Poisson's ratio, under 0.2. A low value means the rock does not distribute stress sideways into the formation, it snaps, and it translates directly into lower minimum horizontal stress. That is a large part of why the Marcellus fractures as well as it does.

Poisson's ratio enters only through the plane strain modulus, and across its whole plausible range it moves the answer far less than Young's modulus does, which in turn barely moves this design at all.

Source: the FastFrac Toolbox Poisson's ratio table, hard sandstone row. The MSEEL log description corroborates the direction; published Marcellus studies give the comparison range.

Fracture gradient

0.80 psi/ft. Published range 0.80 to 1.15 psi/ft.

The closure stress gradient, which the model also uses as the treating pressure in the fracture. This design sits at the bottom of the published band for the core producing areas of Pennsylvania and West Virginia, where shallow underpressured fringes can drop as low as 0.50 to 0.60.

At the MSEEL site the minimum horizontal stress is reported near 6,500 psi against an overburden near 8,800 psi, with the Marcellus cored at 7,445 to 7,557 ft. That works out near 0.87 psi/ft, so a core-area well would commonly sit above this design.

At 0.80 psi/ft this is 5,600 psi of closure at 7,000 ft, which with the 4,200 psi of pore pressure leaves 1,400 psi of net stress on the pack before any drawdown. Raise the gradient and both the closure and the treating pressure rise with it.

Sources: published Marcellus fracture gradient range; MSEEL reported stress magnitudes and cored interval depth, the latter cited second-hand in a conference paper rather than read from the primary data.

Containment height

350 ft. Defensible band 300 to 400 ft.

The total vertical extent the fracture is held to, between the Tully Limestone above and the Onondaga below. Enter the full height, not half of it. See Section 1 for the basis.

Entering 0 instead switches the model to a radial fracture, an unconfined penny, and the length column then reports a radius rather than a half length. That is the right choice only where no barrier is reached, which published Marcellus work does not generally describe.

It is one of the strongest levers in the design. Taller fracture, narrower fracture, since the same slurry is spread over more face, and a narrower fracture is harder for coarse sand to enter. Run your own number from a DFIT or offset microseismic before taking this one.

Sources: the Tully reported about 280 ft above the Marcellus in south-western Pennsylvania, with containment attributed to its high Poisson's ratio and high minimum horizontal stress; microseismic from central Pennsylvania showing growth about 200 ft above the treatment zone terminating at the Tully; and a microseismic-monitored horizontal finding no downward propagation through the Onondaga. See Section 10.

6 Fracturing fluid

At 8 cp this is a linear gel, not slickwater.

Fluid viscosity

8 cp. Published: slickwater 2 to 5 cp, linear gel 6 to 10 cp downhole.

Eight centipoise puts this design in the linear gel band, not the slickwater one. Published Marcellus practice describes a light linear guar gel at 10 to 20 lb per thousand gallons delivering roughly 13 to 15 cp at surface and breaking to 6 to 10 cp downhole once it reaches reservoir temperature and meets the breaker. That is exactly where 8 cp sits.

Pennsylvania Marcellus treatments are predominantly slickwater at 2 to 5 cp, where proppant transport is by velocity and turbulence rather than fluid thickness. A high viscosity friction reducer run heavy for the tail-in reaches 15 to 40 cp. Heavy crosslinked designs at 30 to 40 lb are avoided, because that much polymer does real damage in nanodarcy pores.

The choice decides how the sand is carried. A gel holds proppant in suspension, so proppant settling is slow and the fracture is propped over its full height. Slickwater relies on velocity and turbulence instead, which is why slickwater jobs are pumped fast and lead with fine sand. The viscosity entered is what the proppant transport and the settling check are computed from.

Source: published Marcellus fluid system practice and downhole viscosity ranges.

Friction reduction

50 per cent. Working range 50 to 80 per cent.

Entered as the percentage by which pipe friction is reduced, so 50 means friction is multiplied by 0.50. Applied to pipe and annulus only, never to the perforation drop, because the polymer does nothing where fluid passes through a hole rather than along a wall.

It is the single largest lever on surface pressure, and the longer and faster the job, the more it is worth. On a short heel stage at a modest rate the pipe friction is small to begin with, so the input matters less than it will on a toe stage.

Fifty per cent is conservative against the published figure. A Permian-focused laboratory study measured about 75 per cent reduction at roughly 1 pound per thousand gallons, and polymer drag reduction has a physical ceiling near 80 per cent at Virk's asymptote. Laboratory measurements run above field values, and 50 is below both.

Source: published polymer loadings and friction reducer performance. No field-measured Marcellus percentage was found. A step-down test on the well in front of you is the direct measurement.

Fluid specific gravity

1.05.

Water carrying gel. Appalachian operations reuse produced water heavily, which carries dissolved solids and sits at the top of the range.

Source: none specific. Water density at known salinity.

7 Proppant selection and mesh size

Mesh size

40/70 sand throughout.

A single cut for the whole stage. Published Marcellus practice more commonly ramps through three: 100 mesh early to seal and divert into micro-fractures, 40/70 as the transition, then coarse 30/50 as the main tail-in carrying most of the mass.

FastFrac carries one sieve cut per design by scope, so a single 40/70 job runs natively. A three-cut job has to be run three times, once per cut, reading each against the steps it is pumped in.

The first sand stage is where a single coarse cut is tested. It arrives when the fracture is at its narrowest, and 40/70 is coarse for that moment, which is precisely why the industry leads with 100 mesh. The largest 40/70 grain is 0.0165 in. against 0.0083 in. for 100 mesh. A cut too coarse for the width available bridges at the perforations, which is the near-wellbore screenout every frac design is trying to avoid.

Source: published Marcellus proppant schedules.

Proppant specific gravity - Pack porosity

2.65 and 0.37.

Ordinary silica sand. Pack porosity of 0.37 is within the published 0.35 to 0.40 for unstressed sand, and it is not a property of the mesh size: a sieved pack has the same void fraction whatever the grain diameter.

Porosity enters the pack permeability as the cube, so it moves the fracture conductivity as strongly as the grain size does.

Source: supplier data for silica sand; published proppant pack porosity.

Conductivity damage factor

1.0, the clean pack.

A factor of 1.0 assumes no crushing, no embedment, no gel residue in the pack and no multiphase flow. Values of 0.1 to 0.5 are common in practice.

It scales the reported fracture conductivity directly, so halving it halves the answer. Net stress on the pack here is only 1,400 psi at zero drawdown, because closure is 5,600 psi and pore pressure 4,200, but that climbs toward the full closure stress as the well depletes, which is when sand crushing occurs.

Source: engineering judgement. No published Marcellus retained-conductivity value was used.

8 Injection rate and the stage pump schedule

Injection rate

80 bbl/min. Published range 80 to 95 bbl/min.

Total slurry rate across all clusters, divided by the cluster count to give the rate into each fracture. At 15 clusters that is 5.3 bbl/min each, which is modest.

Rate is the strongest lever this design has, and it is set at the bottom of the published band. It drives the surface pressure, the horsepower on location and the width available to take sand, so it is the first input to move when testing the design.

Source: published Marcellus stage design practice.

The schedule. A clean pad, three ramped sand stages and a displacement. Proppant concentration is given in pounds per gallon of clean fluid, the field convention, and the volumes are slurry volumes, which is what FastFrac expects.

9 This example against current Marcellus practice

Where it is representative, and where it is lighter than what operators pump today.

Each parameter above was checked against published sources. The design is internally sound and every reservoir and rock value sits in or near its published range. As a treatment, though, it is lighter than a current Marcellus stage, and anyone adapting it should know by how much.

Lighter means one thing only here: less sand and less water per foot of lateral. Proppant intensity in pounds per foot and fluid intensity in barrels per foot are how the industry compares one completion against another, and on both measures this stage is about half of what the median Marcellus well is getting today.

The intensity figures assume a stage length the design does not carry. At 15 clusters and 20 ft of cluster spacing this would be a 300 ft stage, against a published median near 200 ft. On a 200 ft stage the same sand gives about 1,820 lb/ft, much closer to the field median. The table below uses the 300 ft figure.

A smaller stage pumped with a gelled fluid at a modest rate is a workable Marcellus treatment. If you are using this as the starting point for a modern completion, expect to raise the proppant and fluid intensity substantially, move to slickwater with a three cut proppant ramp, and lift the rate.

10 Sources for these Marcellus design parameters

The FastFrac Toolbox rock property tables Source of the Young's modulus of 6.25 million psi, the hard dense sandstone average, and the Poisson's ratio of 0.15, the hard sandstone value. The Toolbox states what these are: typical values from the author's working files kept over a career, whose original published source was not retained and which have not been re-verified. Screening values for a first estimate, to be confirmed against core, logs or a published correlation.

MSEEL, the Marcellus Shale Energy and Environment Laboratory DOE and NETL field laboratory in Monongalia County, West Virginia. Cored Marcellus interval 7,445 to 7,557 ft, minimum horizontal stress reported near 6,500 psi against an overburden near 8,800 psi, and log-derived Poisson's ratio described as abnormally low, under 0.2, in the pay zone. The stress figures are cited second-hand in a conference paper rather than read from the primary MSEEL data.

Wright and Company, development trends in the Marcellus and Utica Review covering 2022 to 2024. Median proppant intensity about 2,200 lb/ft, with ultra-high loading above 3,000 lb/ft in the West Virginia dry gas area and north-eastern Pennsylvania. Median stage length near 200 ft. Median lateral length about 11,100 ft in 2022 rising to about 12,700 ft in 2024.

Productivity drivers in North American tight and shale plays Completion trend analysis reporting Marcellus stage spacing at 198 +/- 2 ft since 2015.

Geology and geomechanics of hydraulic fracturing in the Marcellus shale gas play Journal of Natural Gas Engineering review. The Tully Limestone as the upper stress barrier, tight and of high Poisson's ratio and therefore high minimum horizontal stress, and microseismic from central Pennsylvania showing fractures rising about 60 m above the treatment reservoir and terminating at the Tully. The primary source for the containment height in Section 5.

DOE and NETL, evaluation of fracture growth above Marcellus horizontals Six horizontal wells in south-western Pennsylvania. Most microseismic events mapped below the Tully, which sits about 280 ft above the Marcellus at that location, but about 40 per cent of events were above it, all of them at least 2,000 ft below the overlying producing zones. The source for the caveat in Section 1 that containment is not absolute.

Al-Alli and Zoback, microseismic evidence from south-eastern West Virginia Stanford. Three Marcellus horizontals where vertical fracture propagation was principally upward and outside the Marcellus, in a section only about 135 ft thick. A reminder that the containment height depends on how thick your landing interval is.

AAPG Wiki, Appalachian Basin Marcellus gas play Describes good observed containment between the Tully Limestone and Onondaga Limestone intervals bounding the Marcellus Shale.

Published Marcellus reservoir and petrophysical summaries Porosity 6 to 10 per cent averaging about 8.5, matrix permeability 100 to 800 nanodarcy averaging 250 to 400, pore pressure gradient 0.45 to 0.65 psi/ft, gross thickness 50 to over 900 ft with 50 to 250 ft of target net pay, geothermal gradient 0.011 to 0.022 degF per foot.

Published Marcellus completion and fluid practice Slickwater as the standard fluid at 2 to 5 cp, linear guar at 10 to 20 lb per thousand gallons giving 6 to 10 cp downhole, high-viscosity friction reducers at 15 to 40 cp, friction reducer loadings of 0.5 to 2.0 gallons per thousand, three-cut proppant schedules, 4.5 or 5.5 in. P-110 or Q-125 production casing, entrance holes of 0.375 to 0.42 in. with 0.05 to 0.10 in. of erosion over a stage, and maximum allowable treating pressure at 80 to 85 per cent of burst.

Parameters stated in their own entries as unsourced are engineering practice rather than published Marcellus measurements: the friction reduction percentage, the cluster count, the shot count per cluster, the conductivity damage factor and the fluid specific gravity.