WSMC Non-Floating Oil Assessment Tool

Abstract

This document describes the scientific methods implemented in the WSMC Non-Floating Oil (NFO) Assessment Tool. The tool provides incident commanders with a rapid, defensible sinking risk assessment as required by WAC 173-182-324 within one hour of spill discovery. The assessment integrates UNESCO EOS-80 seawater density calculations, PyGNOME-consistent thermal expansion modeling, Fingas (2004) empirical evaporation predictions, mass-conservation density forecasting, and sediment interaction analysis. Oil property data is sourced from the NOAA ADIOS oil library (1,456 oils). Each model, its coefficients, and its source literature are documented below to support regulatory review by the Washington State Department of Ecology.

↑ Top

1. Introduction & Regulatory Context

1.1 Regulatory Requirements

Washington Administrative Code (WAC) 173-182-324 requires the incident commander to perform a non-floating oil assessment within one hour of discovery of a spill involving oils that may become non-floating. The Revised Code of Washington (RCW) 88.46 establishes the broader framework for vessel oil spill prevention and response in Washington waters.

1.2 What is Non-Floating Oil?

Non-floating oil (NFO) is oil that has a density equal to or greater than the surrounding water, causing it to submerge or sink. Oil can become non-floating through several mechanisms:

• Inherent density — Some heavy oils (Group 4–5) are denser than water at ambient conditions.
• Evaporative weathering — Loss of light volatile fractions increases the density of the residual oil over time.
• Sediment interaction — Oil-mineral aggregate (OMA) formation increases the effective density of the oil mass.
• Temperature change — Cooler water increases oil density; if water temperature drops after a spill, buoyancy margin shrinks.

1.3 Assessment Approach

The tool calculates whether a spilled oil’s density exceeds the water density at current environmental conditions, then forecasts density changes over time due to evaporative weathering. The final risk classification (HIGH, MODERATE, or LOW) incorporates the initial density comparison, the weathering forecast timeline, NOAA laboratory weathering data, and sediment interaction risk.

↑ Top

2. Water Density Calculation (UNESCO EOS-80)

2.1 Freshwater Density

Pure water density is calculated using the fifth-order polynomial from UNESCO Technical Papers in Marine Science No. 44 (1983). This equation is valid for temperatures from 0–40°C at atmospheric pressure.

where ρ_w is in kg/m³ and T is temperature in °C.

2.2 Seawater Density

Seawater density adds salinity-dependent correction terms to the freshwater density. The UNESCO EOS-80 equation of state at surface pressure is:

where S is salinity in PSU (Practical Salinity Units), and:

2.3 Salinity Presets

The tool provides standard salinity presets for common Pacific Northwest water conditions:

Custom salinity values may be entered for site-specific conditions.

2.4 Conversion

The UNESCO equations produce density in kg/m³. The tool converts to g/mL by dividing by 1000. For example, typical Puget Sound water at 10°C and 32 PSU has a density of approximately 1.0243 g/mL.

↑ Top

3. Oil Density at Temperature

3.1 Thermal Expansion Model

When only a single density measurement is available for an oil at a given weathering state, the tool uses the thermal expansion model from NOAA’s PyGNOME oil spill trajectory model. The nonlinear form accounts for the compressibility of petroleum:

where ρ_ref is the measured density at reference temperature T_ref, and k is the volumetric thermal expansion coefficient.

The coefficient k depends on the oil density, following the PyGNOME convention:

For small temperature differences, this formula approximates the simpler linear model: ρ(T) ≈ ρ_ref − ρ_ref · k · (T − T_ref). The nonlinear form is more accurate for the larger temperature ranges (0–40°C) encountered in field conditions.

3.2 Multi-Measurement Interpolation

When multiple density measurements are available at different temperatures for the same weathering state, the tool uses linear interpolation between bracketing measurements. For temperatures outside the measurement range, linear extrapolation from the two nearest points is used.

3.3 Weathering State Selection

The NOAA ADIOS database provides density measurements at multiple weathering states (e.g., fresh, 5% weathered, 10% weathered). The tool selects measurements matching the requested weathering fraction (within a 0.5% tolerance). If no match is found, it falls back to fresh (0%) data.

↑ Top

4. Oil Group Classification

Oils are classified into groups based on specific gravity (SG) at 15°C, consistent with WAC 173-182 guidance. API gravity is calculated from specific gravity when not directly reported:

Groups 4 and 5 are the primary concern for non-floating oil assessment. Group 3 oils may also become non-floating after significant weathering.

↑ Top

5. Evaporative Weathering: Fingas (2004) Model

5.1 Background

The Fingas (2004) evaporation model is a significant departure from earlier boundary-layer-regulated models (e.g., Stiver & Mackay, 1984). Fingas demonstrated through extensive laboratory and field experiments that oil slick evaporation is not boundary-layer regulated. Wind speed, slick area, and turbulence have minimal effect on evaporation rates of thin slicks. Instead, evaporation is controlled primarily by:

1. Oil composition — characterized by the mass percentage distilled at 180°C (%D)
2. Air temperature — controls the vapor pressure of volatile fractions
3. Time — evaporation follows either logarithmic or square-root kinetics depending on oil type

5.2 Equation Forms

Fingas identified two empirical equation forms, each applicable to different oil types. Both use a 15°C reference temperature at which the temperature correction term is zero:

Logarithmic form (crude oils, heavy fuels):

Square-root form (light refined products):

where:

• %Ev = percent evaporated (by mass)
• %D = mass percent distilled at 180°C (ASTM D86 equivalent)
• T = air temperature (°C)
• t = time (minutes)
• 15 = reference temperature (°C) at which the regression coefficients were derived

The (T − 15) term is critical: at temperatures below 15°C (typical for Pacific Northwest waters), this term is negative and reduces the predicted evaporation rate. At T = 10°C, for example, the temperature correction is −5, appropriately predicting slower evaporation in cold conditions.

5.3 Equation Selection by Product Type

The square-root form is used for light refined products where evaporation kinetics are diffusion-limited. All other oils use the logarithmic form.

5.4 Distillation Data (%D at 180°C)

The %D parameter is the mass percentage of oil that distills at or below 180°C, determined from ASTM D86 distillation curves. When distillation data is available in the NOAA ADIOS database, the tool interpolates the curve to determine the fraction at exactly 180°C. Linear interpolation is used between bracketing temperature/fraction data points.

When no distillation data is available, the tool estimates %D from the product type using published typical values:

When neither distillation data nor product type is available, API gravity provides a fallback estimate: oils with API > 40 are assigned 35%, API > 30 gets 25%, API > 20 gets 15%, API > 10 gets 8%, and oils with API ≤ 10 get 3%. The conservative default for completely unknown oils is 15%.

5.5 Evaporation Cap

Evaporation is physically bounded — an oil cannot evaporate more mass than it contains in volatile fractions. When a full distillation curve is available, the maximum distillable fraction (the highest measured cumulative fraction) is used as the evaporation cap. Otherwise, a heuristic cap of min(%D × 1.5, 75%) is applied.

5.6 Inverse Function

To estimate the time required to reach a specific evaporation percentage (e.g., to estimate when NOAA lab weathering states are reached in the field), the Fingas equations are analytically inverted:

Logarithmic:

Square root:

Results are capped at 30 days (720 hours) for operational relevance.

5.7 Forecast Timeline

The tool generates evaporation forecasts at standard time steps: 1, 3, 6, 12, 24, 48, 72, 96, and 120 hours. At each step, the Fingas model predicts the percent evaporated, and the mass-conservation density model (Section 6) converts this to a predicted density.

↑ Top

6. Mass Conservation Density Model

6.1 Principle

When oil evaporates, the lighter (lower-density) fractions are preferentially lost. The residual oil is enriched in heavier fractions, increasing its bulk density. The mass-conservation model predicts the density of the residual oil after a given fraction has evaporated:

where:

• ρ_fresh = density of the fresh (un-weathered) oil (g/mL)
• F = mass fraction evaporated (0.0 – 1.0), from the Fingas model
• ρ_volatile = average density of the evaporated (volatile) fraction (g/mL)

6.2 Physical Interpretation

Consider a unit mass of fresh oil. After evaporation, the remaining mass is (1 − F). The mass of the evaporated fraction was F, with average density ρ_volatile. Since the original oil is a mixture of volatile and residual fractions, the residual density follows from conservation of mass:

ρ_fresh = F · ρ_volatile + (1 − F) · ρ_residual

Rearranging gives the formula above. As F increases, ρ_residual increases because lower-density material is removed. If the residual density exceeds water density, the oil sinks.

6.3 Edge Cases

If F ≤ 0 (no evaporation), the model returns ρ_fresh. If F ≥ 1.0 (complete evaporation), the model returns ρ_fresh as a safety bound since complete evaporation is physically unrealistic for most petroleum products.

↑ Top

7. Volatile Fraction Density Estimation

7.1 Calibration from NOAA Lab Data

When the NOAA ADIOS database contains density measurements at multiple weathering states (fresh and weathered), the volatile fraction density can be calibrated by rearranging the mass-conservation equation:

The tool uses the highest available weathering fraction for best calibration accuracy. A sanity check ensures the result falls between 0.60 g/mL (lightest petroleum fractions) and the fresh oil density.

7.2 Published Default Values

When calibration is not possible (no weathered density data), the tool uses published values from Fingas & Fieldhouse (2009):

The conservative default for unknown product types is 0.82 g/mL.

↑ Top

8. Viscosity Model

Oil viscosity varies exponentially with temperature. When multiple viscosity measurements are available at different temperatures, the tool uses log-linear interpolation:

where μ₁ and μ₂ are viscosity measurements (in mPa·s) at temperatures T₁ and T₂. This approach is more accurate than linear interpolation because the Arrhenius-type temperature dependence of viscosity is approximately linear in the logarithmic domain.

For extrapolation beyond the measured temperature range, the slope from the two nearest data points is extended. When only a single measurement is available, that value is returned directly without temperature correction.

Viscosity is reported for informational purposes but is not a primary input to the sinking risk classification. It is relevant to sediment interaction assessment, as high-viscosity oils are more likely to trap sediment particles.

↑ Top

9. Sediment Interaction & Oil-Mineral Aggregate (OMA) Risk

9.1 Background

Oil-mineral aggregate (OMA) formation occurs when suspended sediment particles adhere to oil droplets. This increases the effective density of the oil mass and can cause oil that would otherwise float to sink. OMA formation is influenced by:

• Suspended sediment concentration (load)
• Oil buoyancy margin (water density minus oil density)
• Oil viscosity (higher viscosity traps more sediment)
• Turbulence and mixing energy

9.2 Risk Assessment Method

The tool calculates the buoyancy margin as the difference between water density and the worst-case oil density (the maximum of fresh density and the highest forecasted weathered density):

The buoyancy margin is then evaluated against thresholds that depend on the sediment load reported by the assessor:

These thresholds are conservatively set. Even with low sediment load, near-neutral buoyancy (< 0.01 g/mL margin) warrants a MODERATE risk classification because minimal sediment incorporation could tip the oil to negative buoyancy.

↑ Top

10. Risk Decision Framework

The final risk classification integrates multiple factors through a staged decision process. Risk can only be upgraded (never downgraded) by subsequent factors.

10.1 Step 1: Initial Classification (Fresh Oil Density)

The primary factor is the density difference between fresh oil and water at current conditions:

10.2 Step 2: Weathering Forecast Upgrade

The Fingas model forecast is checked for the time at which oil density first exceeds water density. Risk is upgraded based on the predicted sinking timeline:

10.3 Step 3: NOAA Lab Data Upgrade

If NOAA laboratory weathering data (measured densities at advanced weathering states) confirms that the oil can exceed water density — even if the Fingas timeline does not predict it within the forecast window — the risk is upgraded:

• MODERATE → HIGH
• LOW → MODERATE

10.4 Step 4: Sediment Interaction Upgrade

If the OMA sediment risk assessment (Section 9) returns HIGH:

• LOW → MODERATE
• MODERATE → HIGH

10.5 Response Actions

The final risk level determines the recommended response actions, aligned with WAC 173-182-324 requirements and MSRC non-floating oil response capabilities (OSRO Classification #22). HIGH risk triggers immediate notification to WA Ecology, deployment of side-scan sonar, activation of MSRC NFO response, and bottom sampling. MODERATE risk places NFO resources on standby with enhanced monitoring. LOW risk follows standard surface oil response with reassessment provisions.

↑ Top

11. Data Sources

11.1 NOAA ADIOS Oil Database

The tool uses oil property data from the NOAA ADIOS (Automated Data Inquiry for Oil Spills) library, maintained by the NOAA Office of Response and Restoration, Emergency Response Division. The database contains 1,456 oils with the following properties:

• Densities — Measured at multiple temperatures and weathering states (fresh, 5%, 10%, etc.)
• Viscosities — Dynamic viscosity at multiple temperatures and weathering states
• Distillation curves — ASTM D86-equivalent boiling point distributions (15,415 data points across 276 oils)
• Pour points — Minimum temperature at which the oil flows
• Product type classification — Crude, heavy fuel oil, diesel, etc.
• API gravity — Standard industry gravity measure

11.2 Data Pipeline

NOAA ADIOS JSON data files are parsed and imported into a SQLite database. Unit conversions are applied during import: densities are normalized to g/mL, temperatures to °C, and distillation fractions to 0.0–1.0 scale. Only oils with at least one density measurement are imported.

↑ Top

12. References

1. Fingas, M.F. (2004). Modeling evaporation using models that are not boundary-layer regulated. Journal of Hazardous Materials, 107(1–2), 27–36. doi:10.1016/j.jhazmat.2003.11.007
2. Fingas, M.F. & Fieldhouse, B. (2009). Studies on crude oil and petroleum product emulsions: Water resolution and rheology. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 333(1–3), 67–81.
3. UNESCO (1983). Algorithms for computation of fundamental properties of seawater. UNESCO Technical Papers in Marine Science, No. 44. Paris: UNESCO.
4. NOAA Office of Response and Restoration. ADIOS Oil Database. https://adios.orr.noaa.gov
5. NOAA Emergency Response Division. PyGNOME: General NOAA Operational Modeling Environment. https://github.com/NOAA-ORR-ERD/PyGnome
6. Washington Administrative Code 173-182-324. Oil Spill Contingency Plan — Non-Floating Oil Assessment Requirements.
7. Revised Code of Washington 88.46. Vessel Oil Spill Prevention and Response.
8. ASTM D86. Standard Test Method for Distillation of Petroleum Products and Liquid Fuels at Atmospheric Pressure. ASTM International, West Conshohocken, PA.
9. Stiver, W. & Mackay, D. (1984). Evaporation rate of spills of hydrocarbons and petroleum mixtures. Environmental Science & Technology, 18(11), 834–840.

This document describes the scientific methods as implemented in the WSMC NFO Assessment Tool. All formulas, coefficients, and thresholds correspond directly to the source code in app/assessment.py. For questions about specific implementations, refer to the inline documentation in the source code or contact the WSMC.