Robertson SBT Chart — How to Classify Soils from CPT Data

Because the CPT retrieves no physical soil sample, identifying what soil type exists at each depth requires a classification framework built from the measured resistance data itself. The Robertson Soil Behaviour Type (SBT) chart, first published in 1990 and updated in 2009 and 2016, is the universally accepted solution. It classifies soils by how they behave during cone penetration — not by their grain size or mineralogy — using two normalised parameters plotted on a chart divided into nine SBT zones.

---

What Soil Behaviour Type means #

The word “behaviour” in SBT is deliberate. The CPT cone measures mechanical resistance — how hard it is to push through the soil at 2 cm/s under the prevailing stress conditions. This reflects a combination of soil type, stress history, density, and structure. Two soils with different grain sizes can produce identical CPT readings; one soil can plot in different SBT zones depending on its stress history.

SBT is therefore a proxy for soil type, not a direct measurement of it. In most natural soils it is a reliable and useful proxy — clean sand consistently plots in Zone 6, soft clay consistently in Zone 3. But in unusual geologies — cemented sands, calcareous soils, sensitive clays, heavily overconsolidated clays — SBT can misclassify. This is why the Robertson (1990) framework explicitly recommends cross-checking SBT classification against borehole samples wherever possible, particularly in unfamiliar geological environments. For how USCS classification from physical samples compares, see USCS soil classification — field guide for borehole logging.

---

The chart axes — Q_tn and F_r #

Raw q_c and f_s values increase with depth because overburden pressure increases — the same soil at 5 m and 20 m produces different raw readings. Plotting raw values would shift deeper data points to higher resistance, biasing the classification. Normalisation removes this depth dependency.

The normalised cone resistance Q_tn accounts for total and effective overburden stress:

Q_tn = [(q_t − σ_v0) / P_a] × (P_a / σ’_v0)ⁿ

The normalised friction ratio F_r is:

F_r = [f_s / (q_t − σ_v0)] × 100%

where σ_v0 is total vertical stress, σ’_v0 is effective vertical stress, P_a = 100 kPa (atmospheric pressure reference), and n is a stress exponent that varies between 0.5 for sands and 1.0 for clays. Because n depends on the SBT zone, which is itself what you are trying to determine, the Robertson (2009) procedure is iterative: start with n = 1.0, determine the SBT zone and I_c, update n, recalculate Q_tn, and repeat until convergence — typically two to three iterations.

Both axes are on logarithmic scales. Q_tn is plotted on the vertical axis; F_r on the horizontal axis. Every depth increment of CPT data produces one point on the chart. The cloud of points across a full sounding maps the soil behaviour types present throughout the profile.

---

The nine SBT zones #

Zone	Soil behaviour type	Typical Qtn	Typical Fr (%)
1	Sensitive fine-grained	< 12	< 1.0
2	Organic soils — peat	< 12	> 4.0
3	Clay — silty clay	< 70	3.0–8.0
4	Silt mixture — clayey silt to silty clay	12–70	2.0–4.0
5	Sand mixture — silty sand to sandy silt	12–70	1.0–2.0
6	Sand — clean to silty sand	70–350	0.4–1.0
7	Gravelly sand to dense sand	> 350	< 0.4
8	Very dense sand to gravelly sand	> 350	0.4–1.0
9	Very stiff fine-grained — heavily OC or cemented	Varies	> 4.0

Zones 1 and 2 are separated from Zone 3 by their very low Q_tn but contrasting F_r — sensitive clays plot at very low friction ratios because their structured fabric breaks down under cone penetration. Zone 9 occupies the high-Q_tn, high-F_r corner and typically represents aged, overconsolidated, or cemented fine-grained materials that behave stiffly despite their mineralogy. Zones 4 and 5 are the transition zones between clearly fine-grained and clearly coarse-grained behaviour — the silts and mixed soils that are hardest to characterise from CPT alone.

---

The SBT index I_c #

Rather than manually reading each data point’s zone from the chart, the SBT index I_c provides an automated numerical classification. I_c is the radius of a concentric arc centered near the origin of the Q_tn–F_r chart that passes through the data point:

I_c = [(3.47 − log Q_tn)² + (log F_r + 1.22)²]^0.5

The I_c value maps directly to SBT zones and underpins two important downstream applications. First, it determines the stress exponent n used in the iterative normalisation: n = 0.5 + 0.45 × (I_c/3.27) capped at 1.0 — so the normalisation and classification reinforce each other as the iteration converges. Second, I_c is used as the basis for automated fines content estimation and the clean-sand equivalent correction in CPT-based liquefaction procedures: layers with I_c > 2.60 are generally treated as too fine-grained to liquefy by standard simplified methods.

Approximate I_c zone boundaries:

Ic range	SBT interpretation
< 1.31	Gravelly sand to dense sand (Zones 7–8)
1.31–2.05	Clean sand (Zone 6)
2.05–2.60	Silty sand to sandy silt (Zones 4–5)
2.60–2.95	Silty clay to clay (Zone 3)
> 2.95	Clay to organic soil (Zones 2–3)

For the full treatment of normalised Q_tn, F_r, and B_q calculations, see Normalised CPT plots — Q_tn, F_r, and B_q explained.

---

How DartiGeo plots and classifies CPT data #

DartiGeo’s CPT module performs the full Robertson SBT classification automatically once raw CPT data is imported from Excel. At each depth increment, DartiGeo calculates σ_v0 and σ’_v0 from the entered soil layer profile and groundwater level, applies the iterative normalisation to compute Q_tn, F_r, and I_c, assigns the SBT zone, and produces a depth profile of SBT zones alongside the raw q_c, f_s, and R_f traces. The Q_tn–F_r chart is plotted with the nine zone boundaries and all data points overlaid, giving an immediate visual check of the classification before proceeding to parameter estimation.

SBT zone assignments also drive automatic soil-type filtering in the correlations module — ensuring that friction angle correlations are applied only to zones classifying as coarse-grained and undrained shear strength correlations only to zones classifying as fine-grained, eliminating the risk of applying inappropriate equations across the depth profile.

Download a free 14-day trial of DartiGeo →

---

Frequently asked questions #

Is SBT classification the same as USCS classification? #

No — they are fundamentally different systems. USCS (ASTM D2487) classifies soils by grain size distribution and Atterberg limits, determined from physical laboratory testing of a soil sample. SBT classifies soils by how they behave during CPT penetration, derived entirely from q_t, f_s, and stress parameters — no sample needed. In typical natural soils the two systems broadly agree: Zone 6 (clean sand) generally maps to SW or SP; Zone 3 (clay) generally maps to CL or CH. In unusual soils — cemented, calcareous, or structured — they can diverge significantly, which is why the Robertson framework recommends sample verification in unfamiliar geology.

What does it mean when CPT data plots in Zone 9? #

Zone 9 (very stiff fine-grained, heavily overconsolidated or cemented) represents material that has high Q_tn relative to its F_r — stiff despite being fine-grained in origin. This pattern is associated with heavily overconsolidated clays, cemented silts, residual soils, and some weathered rocks. It is one of the SBT zones where SBT most diverges from USCS — the material may be a clay by grain size but behaves like a stiff coarse-grained material. Borehole sampling is particularly important at Zone 9 depths to confirm the nature of the material before selecting design parameters.

Does the SBT chart work without pore pressure measurement? #

Yes — the Q_tn–F_r chart uses q_t and f_s as its primary inputs, and q_t requires u₂ for its pore pressure correction (q_t = q_c + u₂(1 − a)). In clean sands where u₂ ≈ hydrostatic, the correction is negligible and q_c ≈ q_t, so the chart works adequately without pore pressure measurement. In soft clays where excess pore pressure can be several times q_c, using uncorrected q_c instead of q_t will significantly underestimate normalised resistance and produce an incorrect zone classification. For fine-grained soils, CPTu with u₂ measurement is essential for reliable SBT classification.

What is the difference between the 1990 and 2016 Robertson SBT charts? #

The 1990 chart uses Q_t normalised with a fixed stress exponent of n = 1.0. The 2009 and 2016 updates introduced the iterative stress exponent (n varying between 0.5 and 1.0 depending on soil type) to produce Q_tn, which better handles the transition from sand-like to clay-like behaviour and improves classification accuracy in silty and transitional soils. The zone boundaries are slightly different, and the 2016 update refined the I_c definitions used in liquefaction applications. DartiGeo uses the updated normalisation framework from Robertson (2009) as the current standard of practice.

---

Related articles #

• Cone Penetration Test (CPT) — complete guide
• CPT procedure — equipment, push rate, and data output
• CPTu piezocone test — pore pressure measurement and interpretation
• Normalised CPT plots — Q_tn, F_r, and B_q explained
• Estimating soil parameters from CPT
• USCS soil classification — field guide for borehole logging
• CPT vs SPT — which test is right for your site investigation?