The RMR System Explained, Parameter by Parameter
A complete, working explanation of the Bieniawski Rock Mass Rating system — every rating table, stand-up time charts, the three separate orientation scales for tunnels, foundations and slopes that most summaries leave out, and a free calculator you can use instead of doing the arithmetic by hand.
What the RMR System Actually Does
The Rock Mass Rating system takes a rock mass that geologists would otherwise describe in loose, subjective language — "blocky," "fair," "heavily jointed" — and converts it into a single, repeatable number between 0 and 100. Z.T. Bieniawski built it in the early 1970s at South Africa's Council for Scientific and Industrial Research, and after several revisions it settled into the form most engineers use today, published in 1989 and generally referred to as RMR89.
The system works by splitting the rock mass into discrete, observable pieces of evidence. Instead of asking "is this good rock?" it asks six narrower, answerable questions: how strong is the intact material, how broken up is the drill core, how far apart are the joints, what do the joint surfaces look like, how much water is moving through the mass, and how is the dominant joint set oriented relative to whatever you're building. Each answer earns a point score from a published table, the five strength-related scores are added together, and the orientation score — almost always negative — is subtracted to arrive at the final RMR value.
Before any of the six parameters get scored, the rock mass along a project route first gets divided into structural regions: zones where the geology is reasonably uniform in terms of rock type, joint pattern, and strength. A long tunnel might pass through five or six different structural regions, each one logged and rated separately, because applying a single RMR value across genuinely different ground would hide exactly the variation the system is meant to capture.
The Six Classification Parameters
Each parameter targets a different failure mechanism. Together they cover material strength, fracture intensity, structural geometry, and water — the four things that govern almost every rock mass instability problem.
1. Strength of Intact Rock
Uniaxial compressive strength (or point-load index for stronger rock) of an unfractured sample, worth up to 15 points. This is the only parameter that ignores the fracture network entirely.
2. Rock Quality Designation
The percentage of drill core recovered in pieces longer than 100 mm, worth up to 20 points. RQD reflects fracture frequency along the borehole axis specifically.
3. Spacing of Discontinuities
The average distance between joints in the dominant set, worth up to 20 points. Wider spacing means larger, more stable intact blocks between fractures.
4. Condition of Discontinuities
A composite worth up to 30 points covering persistence, separation, roughness, infill material, and weathering of the joint surfaces themselves — the single heaviest-weighted parameter in the whole system.
5. Groundwater Condition
Worth up to 15 points, rated either from a qualitative description (dry through flowing) or from the ratio of joint water pressure to major principal stress when that data exists.
6. Orientation of Discontinuities
A subtractive adjustment that depends entirely on how the dominant joint set sits relative to the excavation — and critically, the penalty scale is different for tunnels, foundations, and slopes. Covered in full detail below.
Full Rating Tables for Parameters 1–5
These are the published Bieniawski 1989 rating bands, reformatted here for clarity rather than reproduced as a scanned table.
| UCS Range | Point-Load Index | Rating |
|---|---|---|
| > 250 MPa | > 10 MPa | 15 |
| 100 – 250 MPa | 4 – 10 MPa | 12 |
| 50 – 100 MPa | 2 – 4 MPa | 7 |
| 25 – 50 MPa | 1 – 2 MPa | 4 |
| 5 – 25 MPa | Use UCS instead | 2 |
| 1 – 5 MPa | Use UCS instead | 1 |
| < 1 MPa | Use UCS instead | 0 |
Below roughly 25 MPa the point-load test loses reliability, so a direct uniaxial compressive strength test is the recommended method for weak rock rather than converting from point-load values.
| RQD | Rating |
|---|---|
| 90% – 100% | 20 |
| 75% – 90% | 17 |
| 50% – 75% | 13 |
| 25% – 50% | 8 |
| < 25% | 5 |
| Average Spacing | Rating |
|---|---|
| > 2 m | 20 |
| 0.6 – 2 m | 15 |
| 200 – 600 mm | 10 |
| 60 – 200 mm | 8 |
| < 60 mm | 5 |
| Sub-parameter | Best (full points) | Worst (zero points) | Max |
|---|---|---|---|
| Persistence / length | < 1 m | > 20 m | 6 |
| Separation (aperture) | None | > 5 mm | 6 |
| Roughness | Very rough | Slickensided | 6 |
| Infilling | None | Soft gouge > 5 mm | 6 |
| Weathering | Unweathered | Decomposed | 6 |
These five sub-scores are added together for a combined parameter-4 score out of 30, making joint condition the single most heavily weighted parameter in the entire RMR system — worth as much as strength and spacing combined.
| General Condition | Inflow per 10 m Tunnel | Joint Water Pressure Ratio | Rating |
|---|---|---|---|
| Completely dry | None | 0 | 15 |
| Damp | < 10 L/min | 0 – 0.1 | 10 |
| Wet | 10 – 25 L/min | 0.1 – 0.2 | 7 |
| Dripping | 25 – 125 L/min | 0.2 – 0.5 | 4 |
| Flowing | > 125 L/min | > 0.5 | 0 |
The joint water pressure ratio compares measured water pressure in the joints against the major principal stress in the surrounding rock. Where piezometer data exists, this ratio is the more rigorous way to score the parameter; otherwise the qualitative description works fine in the field.
Joint Orientation: Tunnels vs. Foundations vs. Slopes
Almost every quick RMR summary online prints a single orientation table and stops there. That's incomplete. Bieniawski's 1989 revision actually defines three separate penalty scales for orientation, because an unfavorably oriented joint set is not equally dangerous to every type of structure. Click a tab below to see the scale that applies to your project.
Tunnel Orientation Adjustment
Tunnels carry the gentlest orientation penalty of the three, because a tunnel is a closed, supported opening — even an unfavorably oriented joint set still has the opposite wall and the support system working against it.
Foundation Orientation Adjustment
Foundations sit in the middle. A footing loads the rock mass directly from above, so steeply dipping joints beneath a footing create a meaningfully higher risk of differential settlement or block movement than the same joints would around a tunnel — hence a steeper penalty.
Slope Orientation Adjustment
Slopes carry by far the steepest penalty, and it isn't close. A joint set dipping out of an open slope face with nothing holding it back is a textbook setup for planar or wedge sliding — the most common and most sudden failure mode in rock slope engineering — so Bieniawski's table reflects that with a penalty up to five times harsher than the tunnel scale.
Rock Mass Classes and What They Mean
| Class | RMR Range | Description | Typical Stand-Up Time* | Cohesion | Friction Angle |
|---|---|---|---|---|---|
| I | 81–100 | Very good rock | 10 years (15 m span) | > 400 kPa | > 45° |
| II | 61–80 | Good rock | 6 months (8 m span) | 300–400 kPa | 35–45° |
| III | 41–60 | Fair rock | 1 week (5 m span) | 200–300 kPa | 25–35° |
| IV | 21–40 | Poor rock | 10 hours (2.5 m span) | 100–200 kPa | 15–25° |
| V | < 21 | Very poor rock | 30 minutes (1 m span) | < 100 kPa | < 15° |
*Stand-up time is the length of time an unsupported excavation of the stated span can remain open before collapse becomes likely. These figures come from Bieniawski's case-history correlation and should guide, not replace, real-time face monitoring.
Chart 1 — RMR Score by Rock Mass Class
Each bar shows the score range covered by that class. Wider class bands at the top mean small mapping errors matter less in very good rock; narrow bands at the bottom mean the same error can shift you a full class in poor ground.
Chart 2 — Stand-Up Time vs. Unsupported Span
Find your unsupported span on the bottom axis, move up to the curve matching your RMR score, then read across to the stand-up time axis. Both axes are logarithmic, following Bieniawski's original case-history chart.
Support Guidelines by Class
For a 10 m span horseshoe tunnel excavated by drill and blast, under 25 MPa vertical stress — Bieniawski's original reference case.
| Class | Excavation Method | Rock Bolts | Shotcrete | Steel Sets |
|---|---|---|---|---|
| I | Full face, 3 m advance | Generally not required | Generally not required | None |
| II | Full face, 1–1.5 m advance | Spot bolts, 3 m long, 2.5 m spacing | 50 mm in crown where needed | None |
| III | Top heading and bench | Systematic, 4 m long, 1.5–2 m spacing | 50–100 mm crown, 30 mm sides | None |
| IV | Top heading and bench, support close to face | Systematic, 4–5 m long, 1–1.5 m spacing | 100–150 mm crown, 100 mm sides | Light ribs, 1.5 m spacing |
| V | Multiple drifts, immediate support | Systematic, 5–6 m long, 1–1.5 m spacing, invert bolted | 150–200 mm crown, 150 mm sides | Medium-heavy ribs, 0.75 m spacing, lagging |
Worked Example: Scoring a Real Face
A road tunnel is being driven through a moderately weathered granite. Face mapping and lab testing return the following: UCS of 140 MPa, RQD of 72%, average joint spacing of 350 mm, joint surfaces slightly rough with light weathering and no infill, a damp face with minor seepage, and the dominant joint set strikes roughly perpendicular to the tunnel axis at a steep dip — rated as "fair" for tunnel orientation.
- UCS = 140 MPa falls in the 100–250 MPa band → 12 points.
- RQD = 72% falls in the 50–75% band → 13 points.
- Spacing = 350 mm falls in the 200–600 mm band → 10 points.
- Joint condition: slightly rough (3) + slightly weathered (5) + persistence 1–3 m (4) + no separation (6) + no infill (6) → 24 points.
- Groundwater = damp → 10 points.
- Subtotal: 12 + 13 + 10 + 24 + 10 = 69 points.
- Orientation (tunnel scale, fair) = −5 points.
- Final RMR = 69 − 5 = 64 → Class II, good rock.
At RMR 64 the recommended support drops close to the lighter end of Class II: occasional spot bolting in the crown, light wire mesh only where blocks are visibly loose, and no shotcrete required for most of the round. Note how much the joint condition sub-scoring carried the result — at 24 of a possible 30 points it contributed more to the final score than UCS, RQD, and spacing combined, which is typical and why field crews should never rush that part of the mapping. The marked point on Chart 2 above shows exactly where this result lands on the stand-up time curve.
Structural Regions: Why You Don't Rate the Whole Tunnel at Once
Before any scoring happens, the project route gets divided into structural regions — stretches where rock type, joint pattern, and strength stay reasonably consistent. This step matters more than most newcomers expect. Averaging RMR across genuinely different ground types produces a number that describes nothing real: a tunnel that alternates between sound granite and a sheared fault zone every 40 meters does not behave like a single "medium" rock mass, it behaves like two completely different problems stitched together.
In practice, geologists identify region boundaries from lithology contacts, major structural features like faults or shear zones, and visible changes in joint density or weathering grade observed in outcrop, borehole core, or the advancing tunnel face itself. Each region then gets its own full six-parameter rating, and support design tracks the boundaries between regions rather than a single project-wide average.
Pre-Construction Regions
Defined from surface mapping, borehole logs, and geophysical surveys before excavation begins. These set the baseline support design and contract documents.
Real-Time Face Regions
Re-confirmed or revised at every excavation round as the actual face is exposed, since real ground rarely matches predicted boundaries exactly. This is where RMR earns its keep as a live decision tool.
Where People Get RMR Wrong
Using the Wrong Orientation Scale
By far the most common error: applying the tunnel orientation penalty to a slope assessment, or vice versa, which can swing the final score by dozens of points. Always confirm which structure type you're scoring before applying the adjustment.
Rating the Worst Spot, Not the Typical Spot
Bieniawski's tables are meant to reflect typical conditions across a structural region, not the single worst fracture you can find. Cherry-picking the most damaged joint produces an overly conservative — and expensive — support design.
Skipping the Joint Condition Sub-Ratings
Estimating joint condition as a single number instead of separately scoring persistence, separation, roughness, infill, and weathering tends to compress a 30-point parameter into a rough guess, which is the single largest source of inter-observer disagreement in RMR field practice.
Forgetting Seasonal Groundwater Changes
A face mapped dry in summer can score very differently after spring snowmelt or a heavy rain event. Long-duration projects should re-check groundwater scoring across seasons rather than relying on a single early reading.
Treating RMR as a Replacement for Judgment
RMR compresses a great deal of geological information into one number on purpose, for speed and communication. It was never meant to replace an experienced engineering geologist's site-specific judgment, especially in squeezing, swelling, or heavily sheared ground where the system performs worst.
Averaging Across Structural Regions
Blending RMR scores from genuinely different ground types into one project-wide average erases the variation the system exists to capture. Score, log, and support each structural region on its own terms.
Frequently Asked Questions
Score Your Own Rock Mass in Under a Minute
Skip the manual addition. Enter your six parameters — including the correct orientation scale for tunnels, foundations, or slopes — and get the RMR score, class, and support recommendation instantly.
Open the RMR Calculator