Toxic Dispersion

Toxic H₂S Plume Dispersion — Engineering Fundamentals

Pasquill-Gifford Gaussian centerline, Briggs σ coefficients, ERPG / IDLH thresholds, and the limits of the steady-state plume model.

Model

Pasquill-Gifford

Steady-state Gaussian plume with Briggs σ coefficients.

Toxic endpoint

ERPG-2 = 30 ppm

EPA RMP §68 toxic endpoint for buffer-distance setting.

Default screen

Class D, u = 3–5 m/s

Neutral stability per API RP 55 and operator HSE manuals.

Use this guide when you need to:

  • Set buffer distances from quarters and roads for sour releases.
  • Compute ground-level H₂S concentration vs distance.
  • Compare predicted ppm against ERPG / IDLH thresholds.

1. Why Gaussian plume for H₂S

Hydrogen sulfide is one of the few oilfield toxics whose ERPG-2 (30 ppm, 1-hour) sits in the same range as plausible accidental releases from sour separators, flare KO drums, and gas lift lines. Operators need a defensible way to set buffer distances from quarters and roads — the Pasquill-Gifford Gaussian plume gives a closed-form, peer-reviewed estimate that satisfies EPA RMP §68 worst-case analysis and is faster than running a CFD model for every scenario.

The Gaussian plume is steady-state, neutrally buoyant, and assumes flat terrain. Those assumptions are deliberately conservative for H₂S because the gas (MW 34) is only slightly heavier than air (MW 29) — heavy-gas effects (DEGADIS, SLAB) typically matter only for refrigerated leaks (cryogenic propane, LNG) where the cold vapor is much denser than ambient.

2. The Gaussian centerline equation

For a continuous point source of mass release rate Q at height H, the centerline ground-level concentration at downwind distance x is:

C(x, 0, 0) = Q / (π · u · σy · σz) · exp(−H² / (2 σz²))

where u is the wind speed at release height and σ_y, σ_z are the cross-wind and vertical Gaussian dispersion coefficients (both functions of x and stability class). For a ground-level release (H = 0) the exponential term equals 1 and C decays as roughly 1/x for the Briggs power-law σ's.

Converting to volumetric ppm via ideal gas at receiver T, P:

ppm = C (kg/m³) · 10⁶ · R · T / (P · MWH₂S), MWH₂S = 34.08 g/mol

3. Briggs σ coefficients

Briggs (1973) gave compact power-law fits to the Pasquill-Gifford σ_y, σ_z charts. The forms used here:

ClassRural σy (m)Rural σz (m)
A (very unstable)0.22·x·(1+10⁻⁴·x)−0.50.20·x
B0.16·x·(1+10⁻⁴·x)−0.50.12·x
C0.11·x·(1+10⁻⁴·x)−0.50.08·x·(1+2·10⁻⁴·x)−0.5
D (neutral)0.08·x·(1+10⁻⁴·x)−0.50.06·x·(1+1.5·10⁻³·x)−0.5
E0.06·x·(1+10⁻⁴·x)−0.50.03·x·(1+3·10⁻⁴·x)−1
F (stable)0.04·x·(1+10⁻⁴·x)−0.50.016·x·(1+3·10⁻⁴·x)−1

Use urban coefficients when the receiver is within an industrial complex or town; the mechanical turbulence from buildings increases σ for the same x. Class D (neutral) with u = 3–5 m/s is the default screening case in API RP 55 and most operator HSE manuals.

4. ERPG / IDLH thresholds for H₂S

ThresholdConcentrationMeaning
ERPG-10.1 ppmDetection / mild odor; no impairment.
ERPG-230 ppmBoundary at which escape may be impaired after 1-hr exposure. EPA RMP toxic endpoint.
ERPG-3 / IDLH100 ppmLife-threatening after 1-hr exposure (NIOSH IDLH).
OSHA PEL (ceiling)20 ppm (ceiling; 50 ppm 10-min peak)Occupational ceiling — not for dispersion endpoint.

Note: olfactory fatigue sets in above ~100 ppm — workers cannot smell the gas, which is one of the reasons fixed H₂S detectors are mandatory wherever a release scenario can reach IDLH within the response time.

5. Model limits

  • Near field (x < 100 m). Briggs σ are calibrated to ≥100 m. Inside that zone use CFD or jet-source models (e.g., AFTOX).
  • Calm wind (u < 1 m/s). The continuous-plume assumption breaks down — use a puff model (INPUFF) or DEGADIS for stagnant conditions.
  • Heavy gas. H₂S vapor density is close to air, so the neutrally-buoyant Gaussian is acceptable. For two-phase / cryogenic releases (refrigerated LPG, NH₃, Cl₂) use DEGADIS or SLAB instead.
  • Terrain & obstacles. Gaussian assumes flat featureless terrain. In river valleys or near tall structures the plume can be channelled or deflected — verify with on-site met data.
  • Buoyancy. A hot release from a flare or vent has an effective stack height greater than the physical height (Briggs plume rise). The calculator uses H_physical only — apply your own plume-rise correction if needed.

6. References

  • Pasquill, F. (1961). "The estimation of the dispersion of windborne material." Meteorological Magazine 90, 33–49.
  • Briggs, G.A. (1973). Diffusion Estimation for Small Emissions. ATDL Contribution No. 79, NOAA.
  • AIHA ERPG/WEEL Committee, 2026 Handbook — Hydrogen Sulfide.
  • NIOSH (1994). Documentation for IDLHs — Hydrogen Sulfide. NIOSH/CDC.
  • EPA 40 CFR §68 (RMP) Appendix A — toxic endpoints.
  • API RP 55 — Oilfield H₂S Safe Operations.

Frequently Asked Questions

Why use a Gaussian plume model for H₂S?

The Pasquill-Gifford Gaussian plume gives a closed-form, peer-reviewed estimate of ground-level concentration that satisfies EPA RMP §68 worst-case analysis and is far faster than CFD for every scenario. H₂S (MW 34) is only slightly heavier than air, so the neutrally-buoyant assumption is acceptable; heavy-gas models matter mainly for cryogenic releases.

What H₂S concentration is the toxic endpoint?

ERPG-2 is 30 ppm (1-hour) and serves as the EPA RMP toxic endpoint for setting buffer distances. ERPG-3 / IDLH is 100 ppm (life-threatening after 1 hour), while ERPG-1 is 0.1 ppm. Olfactory fatigue above about 100 ppm means workers cannot smell the gas, which is why fixed detectors are mandatory.

When does the Gaussian plume model break down?

It is unreliable in the near field (x < 100 m, below the Briggs σ calibration range), in calm wind (u < 1 m/s, where a puff model is needed), for true heavy-gas or two-phase releases, and over non-flat terrain or near tall structures. Hot buoyant releases also rise above their physical stack height, requiring a separate plume-rise correction.