Input data — description of each field¶

Field-by-field reference for the RSL III app. For the operational flow see Workflow.

Stratigraphy¶

Stratigraphy tab — define the 1D column from ground level downward, down to the bedrock (seismic reference substratum, Vs ≥ 800 m/s).

For each layer¶

Field	Unit	Description
Thickness (h)	m	Layer thickness. Total sum ≥ depth to reference bedrock.
Vs	m/s	Shear-wave velocity at small strain (γ < 10⁻⁴%). From direct seismic tests: cross-hole, down-hole, MASW, ReMi, passive seismic (microtremor HVSR).
ρ (density)	kN/m³	Unit weight. For soils above the water table ρ_d (dry), below ρ_sat (saturated).
G_max	MPa	Small-strain shear modulus. RSL III computes it automatically as `G_max = ρ · Vs²` if you don't enter it.
ξ_min	%	Minimum damping (small strain). Typical: 0.5-1% (dense sands), 1-3% (clays), 2-5% (anthropic fills).
G/Gmax-γ curve	—	Modulus degradation function. Chosen from the library (see below).
ξ-γ curve	—	Damping function. Same library as the G curve.
Lithology	—	Descriptive, does not enter the calculation.

Criteria for defining the bedrock¶

The last layer must be the seismic reference bedrock. Characteristics:

Vs ≥ 800 m/s (NTC 2018 cat. A definition)
Expected linear elastic behaviour at the site's seismicity level (γ < 10⁻³ %)
Deep enough to capture the first site natural frequency: T₀ = 4H/Vs_mean → typically H = 30-100 m

Bedrock depth

If the bedrock is too shallow (e.g. 5-10 m above Vs > 800 m/s), the site does not significantly amplify and RSL III returns output similar to the standard NTC spectrum. In these cases detailed SRA is not worthwhile.

Degradation curves¶

RSL III ships with a standard library. For each layer you can assign:

Clays (Vucetic-Dobry 1991)¶

Family of curves as a function of the plasticity index PI:

PI = 0 (non-plastic clays / silty sands)
PI = 15-30 (medium-plastic clays)
PI = 30-50 (highly plastic clays)
PI > 50 (very high plasticity clays)

As PI increases, the G/Gmax curve "degrades more slowly" and ξ is smaller. Highly plastic clays are the least dissipative case.

Sands (Seed-Idriss 1970)¶

Upper / mean / lower family for sands. The mean is used by default. Upper and lower cover the natural dispersion of sands.

Darendeli (2001) — general formula¶

Synthetic literature curves, depending on PI, OCR and σ'_v (confining pressure). More modern, recommended for advanced studies.

EPRI (1993)¶

Family of curves by depth (typical of nuclear studies). Differentiates by depth range (0-15 m, 15-50 m, 50-150 m, >150 m).

Custom curves¶

Upload a γ [%], G/Gmax [-], ξ [%] table from your specific study (lab tests: cyclic triaxial, resonant column, cyclic torsional shear). CSV format with header.

Input accelerogram¶

Seismic input tab. Load the horizontal acceleration accelerogram at the bedrock.

Supported formats¶

Format	Typical origin
PEER	NGA-West / NGA-Subduction, worldwide accelerogram archive
ITACA	Italian Accelerometric Archive (INGV, Italian records)
CSV / TXT	header `time, accel` (or `t, a`) — generic
GeoStru ACC	export from Spectra or other GeoStru software

What to know about the accelerogram¶

Δt (sampling step): typically 0.005-0.01 s
Total duration: 20-40 s for real recordings, can reach 80 s for synthetic accelerograms with a long tail
a_max: peak acceleration (PGA), in g or m/s²
Magnitude Mw and distance R: useful in the technical report (spectral compatibility with NTC)

Outcrop vs within¶

⚠️ Critical point:

If the accelerogram is recorded on outcropping bedrock (e.g. rock exposed at the surface), choose "Outcrop" in the app. RSL III will apply spectral deconvolution, halving the amplitude inside the column.
If the accelerogram is already defined inside the column (e.g. from a previous SRA model), choose "Within". No deconvolution.

For archive recordings (PEER, ITACA): outcrop is the right choice in 99% of cases.

NTC-compatible spectrum¶

For NTC 2018-compliant analyses, the accelerograms must be spectrum-compatible with the site (nominal life, usage class, return period). Typically you use:

7 real accelerograms (NTC 2018 recommendation, par. 7.3.5)
Modified by scaling or spectral matching to be compatible with the target spectrum

In multi-input mode RSL III processes the 7 accelerograms and produces the mean spectrum (plus envelope / std) — ready for structural verification.

Iteration parameters¶

Advanced tab:

Convergence tolerance: default 0.5% (max γ_eff variation between successive iterations)
Max iterations: default 30 (usually converges in 4-8)
γ_eff/γ_max: default 0.65 (Idriss formulation). Some literature uses 0.5 or 0.85 for special cases.
Maximum frequency f_max: default 25 Hz (caps discretisation and aliasing). Raise for studies with high-frequency accelerometers.

The default values are recommended. Change only if you know what you're doing.

Summary: minimum to get started¶

Stratigraphy: 3-10 layers with thickness, Vs, ρ
Dynamic curves: chosen from the library (even default Seed-Idriss)
Accelerogram: 1 PEER or ITACA file, duration > 20 s

Everything else is refinement.