Home / Articles / TBDY 2018 TBDY-02
Site Classes ZA–ZF: How Does Soil Change an Earthquake?
The same earthquake, at the same distance, produces two different shakings on two different soils. TBDY accounts for this with six site classes and two factors, Fs and F1. But soft soil is not always "worse"; it depends on which period you look at.
1. Why soil matters
When a seismic wave passes from rock into soft soil it slows down. To conserve energy, a slower wave grows in amplitude; on top of that, a soft soil layer can act as a resonator at certain frequencies and amplify the shaking further. Two buildings at the same distance from the same earthquake can therefore experience very different accelerations.
One of the best-known examples is the 1985 Mexico earthquake: about 350 km from the source, damage in the parts of Mexico City built on the old lake bed was far heavier than in the parts on rock. This is why seismic codes treat the soil as a separate parameter.
2. Vs30: soil in one number
The key quantity for the TBDY site class is Vs30, the average shear-wave velocity. An important detail: the 30 metres are measured not from the ground surface but downward from the underside of the foundation or pile cap (Clause 16.4.2); in buildings with basements the two depths differ. The average is over travel times, not thicknesses — the time the wave spends in each layer is summed (Eq. 16.2):
This has an important consequence: a thin but very soft layer lowers the average much more than an arithmetic mean would. A 5 m layer at 120 m/s over 25 m at 500 m/s gives Vs30 = 30 / (5/120 + 25/500) ≈ 327 m/s, while the arithmetic mean would be 437 m/s. The two values fall in different site classes.
3. Six site classes
TBDY Table 16.1 defines six classes. The class can be set by any of three averages over the top 30 metres below the same foundation level, all computed as harmonic averages per Eq. (16.2):
| Class | Soil type | Vs30 (m/s) | (N60)30 (blows/30 cm) | (cu)30 (kPa) |
|---|---|---|---|---|
| ZA | Sound, hard rock | > 1500 | – | – |
| ZB | Slightly weathered, medium-hard rock | 760 – 1500 | – | – |
| ZC | Very dense sand, gravel and hard clay layers, or weathered, highly fractured weak rock | 360 – 760 | > 50 | > 250 |
| ZD | Medium-dense to dense sand, gravel or very stiff clay layers | 180 – 360 | 15 – 50 | 70 – 250 |
| ZE | Loose sand, gravel or soft to stiff clay layers (and the special condition below) | < 180 | < 15 | < 70 |
| ZF | Soils requiring site-specific investigation and evaluation | — | ||
Special condition for ZE: profiles containing soft clay layers (cu < 25 kPa) with PI > 20 and w > 40%, totalling more than 3 metres, are also ZE. All four conditions apply together.
ZF: (1) soils at risk of collapse or failure under seismic action (liquefiable soils, highly sensitive clays, collapsible weakly cemented soils, etc.), (2) peat and/or highly organic clays more than 3 m thick in total, (3) high-plasticity clays (PI > 50) more than 8 m thick in total, (4) very thick (> 35 m) soft or medium-stiff clays. The tabulated values are not used for these soils; a site-specific response analysis is required (Clause 16.5).
Clause 16.4.3: for shallow foundations, if more than 3 m of soil lies between the foundation base and the top of rock, classes ZA and ZB may not be assigned.
4. Fs and F1: how soil changes the spectrum
The hazard map gives two mapped spectral accelerations for rock (ZB): SS (short period) and S1 (1 second). The site class converts them into design values through two factors (TBDY Clause 2.3.2, Tables 2.1 and 2.2):
The values below were computed by the site's engine, which uses a copy of Tables 2.1 and 2.2 verified value-by-value against the Official Gazette. One location (SS = 1.0, S1 = 0.3), five soils:
| Site class | Fs | F1 | SDS = SS·Fs | SD1 = S1·F1 |
|---|---|---|---|---|
| ZA | 0.80 | 0.80 | 0.800 | 0.240 |
| ZB | 0.90 | 0.80 | 0.900 | 0.240 |
| ZC | 1.20 | 1.50 | 1.200 | 0.450 |
| ZD | 1.10 | 2.00 | 1.100 | 0.600 |
| ZE | 1.10 | 2.80 | 1.100 | 0.840 |
The most striking difference is in SD1: 0.84 on ZE versus 0.24 on ZA. Soft soil amplifies the long-period response 3.5 times relative to rock. Structures in the long-period range — tall buildings — are the most affected.
5. Is soft soil always worse?
No — and this is perhaps the most instructive part of the table. Under very strong shaking (SS = 1.5, S1 = 0.6) the factors become:
| Class | Fs | F1 |
|---|---|---|
| ZC | 1.20 | 1.40 |
| ZD | 1.00 | 1.70 |
| ZE | 0.80 | 2.00 |
On ZE the short-period factor drops to 0.8, against 1.2 on ZC. Under very strong shaking, soft soil reduces the short-period acceleration. The reason is the nonlinear behaviour of soil: at large strains soft soil yields, loses stiffness and dissipates energy, filtering out high-frequency vibration.
Yet F1 on the same row is still 2.0: long-period shaking keeps being amplified. "Soft soil is always bad" is therefore incomplete; more accurately, soft soil changes the frequency content of the shaking. Which building suffers more depends on its period.
6. Reading a geotechnical report
The site class is chosen at the start of design and is hard to question later; a wrong class silently changes every seismic force. Four things are worth checking: how Vs30 was obtained (direct seismic measurement or a correlation estimate?), whether the 30 m were measured from the foundation level, whether the profile contains a thin but soft layer (the harmonic average does not hide it, an arithmetic average would), and whether the ZF conditions — especially liquefaction potential — were examined.
References
- TBDY 2018 — Clause 2.3.2, Table 2.1 (Fs) and Table 2.2 (F1); Clauses 16.4.1–16.4.3, Table 16.1 and Eq. (16.2) (site classes); Clause 16.5 (site-specific analysis); Clause 16.6 (liquefaction). Table 16.1 and Clause 16.4 checked against the official annex PDF (2026-10-21); values of Tables 2.1 and 2.2 verified one-to-one against the Official Gazette text (site code pack, 2026-07-13).
- The Vs30 class limits were also checked against an independent academic source: ITU, "Effect of local site class and earthquake characteristics on one-dimensional dynamic behaviour" (polen.itu.edu.tr).
- Kramer, S. L. (1996). Geotechnical Earthquake Engineering, Chapter 7 (local site effects). Prentice Hall.
The numbers in this article were produced by the site's engines; code values are given with their clause numbers.