RFP_P2675002,_Attachment_-_Catalyst_Report_2019_FINAL.pdf

California Oil Spill Response Cost Study

November 2019

Prepared By:

Prepared For:

- ii- November 2019

This Page is Intentionally Blank

- iii- November 2019

Contents Executive Summary SECTION 1 Introduction and Study Objectives

1.1 Background

1.2 Current Regulations

1.3 Study Objectives

1.4 Model Development

1.5 Report Organization

SECTION 2 Review of Previous Research on Large California Spills

2.1 Noteworthy Historical Spills in California

2.2 US Army Corps of Engineers San Francisco Bay Spill Study

2.3 CDFW OSPR San Francisco Bay Booming Study

2.4 BSEE Oil spill Response Plan Capability Study

SECTION 3 Review of Data Sources .......................................................................................... - 16 -

3.1 California Oil Spill Volume Data .............................................................................................. - 16 -

3.2 Oil Spill Cost Data Collection ................................................................................................... - 20 -

SECTION 4 Analysis of Recent History of Spill Volumes by Category .......................................... - 28 -

4.1 Recent Spill History For Tankers ............................................................................................. - 28 -

4.2 Recent Spill History For Tank Barges ...................................................................................... - 33 -

4.3 Recent Spill History For Non-Tank Vessels .............................................................................. - 34 -

4.4 Recent Spill History Marine Facilities ...................................................................................... - 36 -

4.5 Recent Spill History For Offshore Platforms ........................................................................... - 37 -

4.6 Recent Spill History For Marine Pipelines ............................................................................... - 39 -

4.7 Recent Spill History For Small Marine Fuel Facilities .............................................................. - 40 -

4.8 Recent Spill History For Mobile Transfer Unit (MTU) ............................................................. - 41 -

4.9 Recent Spill History for Inland Production Facilities ............................................................... - 42 -

4.10 Recent Spill History For Inland Pipelines................................................................................. - 44 -

4.11 Recent Spill History For Inland Rail ......................................................................................... - 49 -

4.12 Summary ................................................................................................................................. - 53 -

SECTION 5 Model Development .............................................................................................. - 55 -

5.1 General Issues Regarding Prediction of Oil Spill Costs ............................................................ - 55 -

5.2 Limitation in Predicting Oil Behavior, Fate, and Effects.......................................................... - 56 -

5.3 Estimating Unit Spill Costs ...................................................................................................... - 56 -

SECTION 6 Model Results - Estimating Oil Spill Costs ................................................................ - 60 -

6.1 General Distribution of Spill Costs Based on Historic WorldWide Data ................................. - 60 -

- iv- November 2019

6.2 Distribution of Spill Costs in California Oil Operator Survey ................................................... - 62 -

6.3 Unit Spill Costs from Oil Spill Cost Model ............................................................................... - 65 -

SECTION 7 Discussion of Spill Costs ......................................................................................... - 66 - SECTION 8 References ............................................................................................................. - 67 - SECTION 9 Acronyms and Terminology .................................................................................... - 74 - Appendix A: CDFW Oil Spill Cost Model ........................................................................................ - 77 -

Figures Figure 1: Ephemeral/Intermittent Streams (Left); Perennial Streams (Right) in California Figure 2: Flowchart of Model Development for RWCS Estimates Figure 3: Lakeview No. 1 Gusher 1910 Figure 4: 1969 Santa Barbara Platform A Blowout Figure 5: 2015 Oil Pipeline Break at Refugio State Beach Figure 6: Maximum Possible Water Surface Exposure to Floating Oil (g/m2): Gasoline Figure 7: Maximum Possible Water Surface Exposure to Floating Oil (g/m2): Diesel Figure 8: Maximum Possible Water Surface Exposure to Floating Oil (g/m2): Crude Figure 9: Maximum Possible Water Surface Exposure to Floating Oil (g/m2): HFO Figure 10: Maximum Concentrations of Surface Oiling in 170-Day Modeled Discharge Figure 11: Maximum Concentrations of Surface Oiling in 10-Day Modeled Discharge Figure 12: Number of Oil Spills of at Least 1 Barrel in California 1998–2018 ......................................... - 18 - Figure 13: Total Annual Volume of Oil Spillage in California .................................................................. - 19 - Figure 14: 95th Percentile Tanker Spill Volumes in US by Time Period ................................................... - 29 - Figure 15: Oil Spill Volume Probability Distribution for Non-Tank Vessels, 1998-2018, N-60 ............... - 35 - Figure 16: Oil Spill Volume Probability Distribution for Marine Facilities, 1998-2018, N=307 ............... - 37 - Figure 17: Pacific OCS Region Well Map ................................................................................................. - 38 - Figure 18: Probability Distribution Function for U.S. Marine Pipeline Spills 1998 – 2007, N=2,335 ...... - 40 - Figure 19: Probability Distribution Functions for Small Marine Fuel Facilities 1998-2018, N=22 .......... - 41 - Figure 20: Oil Spill Volume Probability Distribution for Mobile Transfer Units, 1998-2018, N=98 ........ - 42 - Figure 21: Inland Production Facility Spill Volume Probability Distribution, 1998-2018, N=4,164 ....... - 43 - Figure 22: Inland Production Facility Spill Volume Probability Distribution (Excluding Produced Water), 1998-2018, N=3,178................................................................................................................................ - 44 - Figure 23: Oil Spill Volume Probability Distribution for Inland Pipelines, 1998-2018, N=207 ................ - 45 - Figure 24: Major US Pipeline Spills (>238 bbl): Five-Year Average Spill Numbers .................................. - 46 - Figure 25: Crude Pipeline Spill Volume Distribution by Time Periods .................................................... - 47 - Figure 26: US Inland Oil Pipeline Spill Number per Volume Transmission (1985-2015) ........................ - 48 - Figure 27: Major US Oil Pipeline Spill Number per Volume Transmission (1985-2015) ......................... - 48 - Figure 28: Oil Spill Volume Probability Distribution for Railroads, 1998-2018, N=363 (CA OES Data) ... - 49 - Figure 29: Oil Spill Volume Probability Distribution from US and Canadian CBR Accidents, 2013-2018, N=23 ........................................................................................................................................................ - 51 - Figure 30: CDFW-OSC Model Derivation of Geographic Unit Base Costs ............................................... - 56 - Figure 31: Basic Algorithm of CDFW-OSC Model .................................................................................... - 58 - Figure 32: Distribution of Per-Bbl Response Costs ................................................................................. - 61 - Figure 33: Log-Normal Distribution of Per-Bbl Response Costs ............................................................. - 61 -

- v- November 2019

Figure 34: Cumulative Probability Distribution of Per-Bbl Response Costs ............................................ - 62 - Figure 35: CA Survey Distribution of Per-Bbl Response Costs ................................................................ - 63 - Figure 36: CA Survey Log-Normal Distribution of Per-Bbl Response Costs ............................................ - 64 - Figure 37: California Operator Survey Cumulative Probability Distribution of Per-Bbl Response Costs - 64 - Figure 38: Hypothetical Per-Unit Volume Spill Cost Relationship with Volume and Complexity ........... - 66 -

Tables Table ES-1: Summary of Spill Analysis by OSPR-Regulated Category Table ES-2: Per-Bbl Response Cost Percentiles for Inland Production Facilities based on Responses to the California Operator Survey Table ES-3: CDFW-OSC Model Results – Range of Per Barrel Spill Costs by Oil Type for Offshore or Coastal Spills Greater than 100 bbl Table 1: Current OSPR Certificate of Financial Responsibility Amounts Table 2: Current Reasonable Worst-Case Spill Volume Definitions Table 3: USACE San Francisco Bay Modeling Study Oil Spill Types and Volumes Table 4: Total Oil Spill Scenario Costs for USACE San Francisco Bay Study (updated to 2019 US$) Table 5: California Oil Spills 1998 through October 2018 ....................................................................... - 17 - Table 6: Breakdown of California Spills by Category, 1998 – 201826 ...................................................... - 19 - Table 7: Summary of ERC Oil Spill Cost Database Information – US vs Non-US by Oil Group ................ - 21 - Table 8: Data Set for Oil Spill Cost Information–Regions by Category, 1982-2018 ................................ - 22 - Table 9: Data Set for Oil Spill Cost Information–Regions by Oil Persistence Group, 1982-2018 ............ - 22 - Table 10: Data Set for Oil Spill Cost Information–Regions by Spill Volume Category, 1982-2018 ......... - 23 - Table 11: Entities Provided with Survey Questionnaire ......................................................................... - 24 - Table 12: Survey of California Regulated Entities on Oil Spill Costs ....................................................... - 26 - Table 13: Survey of California Inland Oil Operators–Spill Volume Distribution ..................................... - 26 - Table 14: Survey of California Inland Oil Operators–Location Types1 .................................................... - 26 - Table 15: Survey of California Inland Oil Operators–Oil Types ............................................................... - 27 - Table 16: Tanker Spills in California ........................................................................................................ - 28 - Table 17: US Tanker Spill Volumes by Time Period................................................................................. - 30 - Table 18: Largest Tanker Spills in and near US Waters ........................................................................... - 30 - Table 19: Largest Tanker Spills Worldwide ............................................................................................. - 31 - Table 20: Oil Outflow Probability for Double-Hulled Tankers in Impact Accidents46 ............................. - 32 - Table 21: Oil Outflow Probability for Double-Hulled Tank Barge Impact Accidents .............................. - 33 - Table 22: Largest Tank Barge Spills in and near US Waters .................................................................... - 34 - Table 23: Oil Spill Volume Percentiles for Non-Tank Vessels, 1998-2018, N=60 (CA OES Data) ............ - 35 - Table 24: Oil Outflow Probability for Bunker Tank Impact Accidents .................................................... - 36 - Table 25: Oil Spill Volume Percentiles for Marine Facilities, 1998-2018, N=307.................................... - 36 - Table 26: Pacific OCS Region Platform Data ........................................................................................... - 38 - Table 27: Oil Spill Volume Probability for U.S. Marine Pipeline Spills, 1998-2007, N=2,335 ................. - 40 - Table 28: Oil Spill Volume Percentiles for Mobile Transfer Units, 1998-2018, N=98 ............................. - 42 - Table 29: Oil Spill Volume Percentiles for Inland Production Facilities, 1998-2018, N=4,164 ............... - 43 - Table 30: Oil Spill Volume Percentiles for Inland Production Facilities (Excluding Produced Water), 1998- 2018, N=3,178 (Based on CA OES Data).................................................................................................. - 44 - Table 31: Oil Spill Volume Percentiles for Inland Pipelines, 1998-2018, N=207 .................................... - 44 -

Executive Summary - vi- November 2019

Table 32: US Crude Pipeline Spills of 20,000 bbl or Greater (1968–2012) ............................................. - 45 - Table 33: Oil Spill Volumes for Railroad Spills, 1998-2018, N=363 (Based on CA OES Data) ................. - 49 - Table 34: Notable CBR US and Canadian Accidents with Crude Spillage 2013–2016 ............................. - 50 - Table 35: Oil Spill Percentile Volumes for US and Canadian CBR Accidents, 2013-2018, N=23 ............. - 52 - Table 36: Expected CBR Spill Volume per Incident (Loaded Unit Trains) ............................................... - 52 - Table 37: Current OSPR Reasonable Worst-Case Spill (RWCS) Volume Definitions for Rail Tank Cars .. - 53 - Table 38: Comparison of OSPR and PHMSA Worst-Case Discharge Spill Volumes for Rail Tank Cars ... - 53 - Table 39: Oil Spill Volume Summary by Category ................................................................................... - 54 - Table 40: Different Approaches to Calculation of Per-Unit Base Response Costs.................................. - 58 - Table 41: Oil Type Classifications for CDFW-OSC Model ........................................................................ - 59 - Table 42: Per-Bbl Response Cost Percentiles based on data available from 443 Spills Worldwide ....... - 62 - Table 43: Per-Bbl Response Cost Percentiles in California Operators Survey ........................................ - 65 - Table 44: California Survey Per-Bbl Response Cost Percentiles (without Produced Water) .................. - 65 - Table 45: CDFW-OSC Model Results –Spill Costs by Oil Type for Spills Greater than 100 bbl................ - 65 -

Executive Summary - 1- November 2019

Executive Summary Firms that produce, transport, or process oil and gas face the financial risk of paying for cleanup, penalties, legal services, and other expenses associated with the accidental release of product to the natural environment. Although infrequent, a large spill can be very expensive and so firms operating in California and elsewhere are required to demonstrate financial responsibility that shows they will be able to pay such costs. Under the authority of the Lempert-Keene-Seastrand Oil Spill Prevention & Response Act (CA Gov. C. §8670.1 et seq.), the Office of Spill Prevention and Response (OSPR) is required to establish financial responsibility requirements for oil spills from certain segments of industry. Senate Bill 861, among other mandates, required OSPR to establish such requirements for inland facilities. California is now considering potential revisions to these financial requirements. In Fall 2018, the Office of Spill Prevention and Response retained Catalyst Environmental Solutions Corporation (Catalyst), teamed with Environmental Research Consulting (ERC) and Greene Economics, to develop a customized oil spill cost model using current data that could be applied to spill scenarios in California for the purpose of exploring the need for revisions to current financial requirements. This report documents the background research and analysis used to develop the model and presents the results, by category of entity that is regulated by CDFW OSPR.

The current OSPR COFR amounts and RWCS volumes are based on statutory requirements or the results of the 1993 study. OSPR is currently updating its regulations. The specific objectives of this study are to:

1. Gather up-to-date spill volume and oil spill response cost data to inform development of California oil spill regulations and COFR requirements; and

2. Develop a cost model, called the CDFW Oil Spill Cost model (CDFW-OSC), that can be used as a tool by OSPR to estimate cleanup costs and damages in a variety of habitat types.

Several different sources of data were used to inform the analysis of recent spill history (1990s through 2018) and to create the relationships in the cost model developed for the study. Table ES-1 summarizes the sources of information for each category of entity regulated by OSPR regulations and indicates the section of this report where detailed analysis of spill history is described. The primary source for California spills is based on hazardous material release reporting and is publicly available on the California Office of Emergency Services website (CA OES).

ERC augmented this with a survey of inland oil producers in California. This California Operator Survey data includes 133 cases for which at least some cost data were provided. The costs were mainly for small spills at inland production facilities. Spill volumes varied from 0.25 bbl to 1,800 bbl, with an average of 44 bbl. 62% of the spills involved 5 bbl or less of spillage. Nearly a third of the spills involved produced water rather than only oil. Most of these spill cleanups would have been in a fairly limited area within the bounds of the facility. Spill costs varied from about $35 per bbl to $29,341 per bbl, with an average (mean) of $1,954.

CDFW OSPR provided data regarding vehicle spills between 2015 and 2017 to inform the analysis of mobile transfer units (MTUs).

The US Energy Information Administration (EIA) provided information on crude oil transport by rail and pipeline, and storage tank capacities gathered from oil spill contingency plans.

Finally, ERC relied on its proprietary database of oil spills and spill costs worldwide. The ERC Oil Spill Cost Database includes 443 spills from around the world, including spills from all types of sources and in different types of locations. The cases generally involve much larger spill volumes, with an average of

Executive Summary - 2- November 2019

49,500 bbl. The range of spill volumes is from two bbl to 4.9 million bbl. The spill costs average $10,697 per bbl with a maximum of $690,255 per bbl. These spills are generally more complex than the ones captured in the CA OES data. The data on these more complex, and overall more costly, spill response operations and damages from these spills are more readily available (thought often as part of litigation) than for smaller spills. For that reason, these more “notorious” incidents make up the majority of the incidents in this database.

In general, there are significant challenges in obtaining accurate records for oil spill costs. Unless a spill’s response and damage costs are part of public records (as with many non-US tanker spills funded by international funds, or in the US if the spill is federalized), the records are often confidential. The data are kept confidential by the responsible parties. Those costs that are publicized, as in media reports, are usually not broken down by cost type (response, damages, etc.), which makes it difficult to make reasonable comparisons.

Table ES-1: Summary of Spill Analysis by OSPR-Regulated Category Category Source of Spill

Volume Data Described in Section

Number of Spills Included in the

Dataset

Time Period of Spills Included in the

Dataset

Location of Analysis of Recent Spill

History in Study

Tanker ERC Oil Spill Database

150 1968-2012 Section 4.1

CA OES database 5 1998-2018 Tank Barge CA OES database 9 1998-2018 Section 4.2

Non-Tank Vessel CA OES database 62 1998-2018 Section 4.3 Marine Facility CA OES database 310 1998-2018 Section 4.4

Offshore Platform ERC Oil Spill database

10 1968-2012 Section 4.5

Marine Pipeline ERC Oil Spill database

0 (modeled) modeled Section 4.6

Small Marine Fueling Facility

CA OES database 22 1998-2018 Section 4.7

Mobile Transfer Unit (Marine)

CDFW OSPR data 2 2015-2017 Section 4.8

Inland Production Facility

CA OES database 4,164 1998-2018 Section 4.9 CA Operator Survey 157 (133 with cost) 2015-2018

Inland Pipeline CA OES database 206 1998-2018 Section 4.10 Inland Rail CA OES database 363 1998-2018 Section 4.11

As described in Section 6, the model results show that the per barrel costs for oil spill cleanup and response are dependent on the type of oil spilled and the spill location (inland to flowing water, or marine), rather than the spill source. As respondents with actual cost data to the operator survey only apply to inland production facilities, these data were used only to inform this category (responses from other categories of regulated entities provided modeled versus actual data).

There are differences of over an order of magnitude between the current California COFR unit costs and the unit costs for the smaller spills as estimated by the CDFW-OSC model. The California Operator Survey costs are considerably smaller (less than 100 bbl) and none of these spills reached flowing water.

The spills in the survey were considerably smaller than those incorporated into the ERC Spill Cost Database upon which the CDFW-OSC depends. There is a general reduction in per-unit costs as spill volumes increase. This is largely attributable to an “economy of scale” factor. This is generally true for larger complex response operations and the damages that occur from large spills. Once the equipment, Executive Summary - 3- November 2019 personnel, logistics, and overall response “infrastructure” and incident command systems are in place, the level of effort is spread out over a larger number of barrels of oil spilled and does not increase in direct proportion to each additional barrel that was spilled. Once the spill response becomes a complex operation, the unit costs increase sharply, but then drop off as the costs are spread over a greater volume.

However, for very small spills of less than 100 bbl or so, this relationship breaks down. For very small spills, like those reported in the California Operators Survey, the spill cleanup is relatively routine and can be handled by a smaller crew or even with on-site personnel trained in response. There is no complex incident command center with large numbers of state and federal officials in attendance. There is a relatively simple array of response equipment being employed.

Table ES-2 summarizes the per bbl response costs from the respondents

Table ES-2: Per-Bbl Response Cost Percentiles for Inland Production Facilities based on Responses to the California Operator Survey

Percentile Response Cost/Bbl (2019 US$) 10th $35 25th $101

50th (Median) $343 75th $1,547 90th $6,600 95th $10,000 99th $14,500

Maximum $29,341 Average (Mean) $1,954

Table ES-3 provides the results of per barrel spill costs for larger spills into water based on oil type.

These results apply to spills greater than 100 bbl which occurred either offshore or in coastal areas and entered marine or large river system environments.

Table ES-3: CDFW-OSC Model Results – Range of Per Barrel Spill Costs by Oil Type for Offshore or Coastal Spills Greater than 100 bbl1

Oil Category Per-Bbl Spill Cost

Highest Cost High Cost Medium Cost Low Cost

Non-Persistent $17,144 $13,055 $6,747 $4,615

Light Persistent $31,764 $24,183 $12,498 $8,547

Medium Persistent $38,805 $29,539 $15,268 $10,445

Heavy Persistent $70,386 $53,582 $27,700 $18,943

1 The vast majority of spill costs included in the model are marine spills which occurred in coastal and offshore environments; however, the costs also include some spills to major rivers (e.g. the Kalamazoo River spill)

Introduction and Study Objectives - 1- November 2019

SECTION 1 Introduction and Study Objectives Firms that produce, transport, or process oil and gas face the financial risk of paying for cleanup, penalties, legal services, and other expenses associated with the accidental release of product to the natural environment. Although infrequent, a large spill can be very expensive and so firms operating in California and elsewhere are required to demonstrate financial responsibility that shows they will be able to pay such costs. Under the authority of the Lempert-Keene-Seastrand Oil Spill Prevention & Response Act (CA Gov. C. §8670.1 et seq.), the Office of Spill Prevention and Response (OSPR) is required to establish financial responsibility requirements for oil spills from certain segments of industry. Senate Bill 861, among other mandates, required OSPR to establish such requirements for inland facilities. California is now considering potential revisions to these financial requirements. In fall 2018, the Office of Spill Prevention and Response retained Catalyst Environmental Solutions Corporation (Catalyst), teamed with Environmental Research Consulting (ERC) and Greene Economics to develop a customized oil spill cost model using current data that could be applied to spill scenarios in California for the purpose of exploring the need for revisions to current financial requirements. This report documents the background research and analysis used to develop the model and presents the results.

1.1 Background

Following the 1989 Exxon Valdez oil spill in Alaska and the subsequent American Trader tanker oil spill off the coast of Huntington Beach in February 1990, the California Legislature enacted the Lempert- Keene-Seastrand Oil Spill Prevention and Response Act. The Act covers all aspects of marine oil spill prevention, preparedness, and response in California and established an Administrator that is appointed by the California State Governor and is also a Chief Deputy Director of the California Department of Fish and Wildlife (CDFW). The Lempert-Keene-Seastrand Oil Spill Prevention and Response Act (Act) mandates that certain owners and operators of vessels and facilities obtain a certificate of financial responsibility (COFR) before operating in California where a spill of oil could impact marine waters. A COFR is an official written acknowledgement issued by the OSPR Administrator that the operator has demonstrated the ability to pay for cleanup costs and damages caused by a marine oil spill. The OSPR opened in 1991 as a division of CDFW that has the responsibility for protecting California’s natural resources through prevention of and response to oil spills. In 1993 OSPR commissioned a study to examine cleanup costs and monetary damages resulting from marine oil spills in order to inform requirements on the financial responsibilities of vessels and facilities.2

In 2014 Governor Jerry Brown expanded the OSPR program to cover all state surface waters at risk of oil spills from any source, including pipelines, production facilities, and the increasing shipments of oil transported by railroads. Senate Bill 861, adopted in June 2014, authorized the expansion and provided the additional statutory and regulatory authority for the prevention, preparedness and response activities in the new inland areas of responsibility. Senate Bill 861 expanded OSPR’s jurisdiction to cover oil spills to any surface waters (i.e. waters of the State), including dry washes or ephemeral/intermittent streams (see Figure 1). Consequently, the COFR requirements were expanded to apply to inland facilities, including inland oil production facilities, and transportation of oil by rail or inland oil pipelines.

2 Mercer Management Consulting 1993.

Introduction and Study Objectives - 2- November 2019

Figure 1: Ephemeral/Intermittent Streams (Left); Perennial Streams (Right) in California3

1.2 Current Regulations

The current OSPR COFR amounts as per California Code of Regulations (CCR), 14 CCR § 791.7, are in Table 1. Facility spills are classified with respect to risk to different types of waterways (marine, intermittent, ephemeral, or perennial). COFR amounts for the various spill sources are based on a function of the calculated Reasonable Worst-Case Spill (RWCS) volumes set by OSPR and a pre-determined per-barrel (bbl) cost. The current definitions of RWCS volumes are shown in Table 2.

1.3 Study Objectives

The current OSPR COFR amounts and RWCS volumes are either set forth in statute or based on the results of the 1993 study. OSPR is currently updating its regulations. The specific objectives of this study are to:

1. Gather up-to-date spill volume and oil spill response cost data to inform development of California oil spill regulations and COFR requirements; and

2. Develop a cost model that can be used as a tool by OSPR to estimate cleanup costs and damages in a variety of habitat types.

1.4 Model Development

The flow chart below (Figure 2) shows the process for building the model and outputs. As described in detail in this report, we gathered cost and spill data from publicly available databases and inquiries to OSPR-regulated entities, supplemented with existing cost data from spills worldwide to use as inputs to the project-specific model. As shown in the diagram, the cost model estimates the cleanup costs and potential damages by categorizing the individual database entries by the key factors that are correlated with spill costs. These factors are primarily (but not exclusively): spill volume, oil type, geographic location (jurisdiction and ecological/socioeconomic features), and OSPR Plan Type (rail, vessel, pipeline, facility, or transfer). Data are analyzed to establish statistical relationships between the spill factors and costs. The relationships can then be used to develop revised RWCS volumes and set revised COFR amounts where needed.

3 Southwest Environmental Response Management Application (NOAA).

Introduction and Study Objectives - 3- November 2019

Figure 2: Flowchart of Model Development for RWCS Estimates

1.5 Report Organization

This report presents the technical documentation and results of the development of the CDFW Oil Spill Cost (CDFW-OSC) model, as follows:

• Section 2 – Previous Studies on Large California Spills – provides background and cost data for historic large spills in California history, and provides context for the analysis.

• Section 3 – Review of Data Sources – describes the sources of data used to gather information on oil spills in California and oil spill response costs.

• Section 4 – Analysis of Recent Spill History – examines recent spill history for each of the source types and uses data to calculate risk percentiles.

• Section 5 – Model Development – discussed issues and limitations in modeling oil spill costs and the algorithms used to develop the oil spill cost model

• Section 6 – Model Results – Estimating Oil Spill Costs – describes our analysis of the distribution of oil spill costs in California based on historical data and compared to worldwide data and provide the output of unit spill costs from the oil spill cost model.

• Section 7 – Discussion – provides a discussion of the results and describes the effect of spill volume and location on COFR amounts.

• Section 8 – References – provides a bibliography of sources used in this report.

• Section 9 – Acronyms and Terminology – provides an index of acronyms and abbreviations used in this report as well as a brief glossary of technical terms used in this report.

• Appendix A – ERC-CDFW Oil Spill Cost Model – provides details regarding the development of the cost model and algorithms used.

Introduction and Study Objectives - 4- November 2019

Table 1: Current OSPR Certificate of Financial Responsibility Amounts

Category Sub-category

COFR

amount/calculation Minimum Maximum

Tanker all $1B $1B

Tank Barge

Large (>150,000 bbl capacity)

$1B $1B

Small (<150,000 bbl capacity)

$12,500 x (30% of total cargo capacity)

$562.5M

Non-tank Vessel

CA or federal >7,500 bbl total oil capacity;

private >6,500 bbl $300 M $300 M

CA or federal 1,001- 7,500 bbl; private 1,001-6,500 bbl

[(Total bbl capacity – 1,000) x $5,670] +

$18.9M $18.9M $50.1M

501-1,000 $18.9M $18.9M 51-500 $10M $10M 11-50 $5M $5M 1-10 $2M $2M

Marine Facility (e.g.

terminals)

$12,500 x RWCS $1M $300M

Offshore Platform Not drilling $12,500 x RWCS $1M $300M

Active drilling $12,500 x RWCS $10M $300M

Marine Pipeline $12,500 x RWCS $1M $300M

Small Marine Fueling Facility

$12,500 x RWCS $600K

MTU

$12,500 x (30% of max cargo capacity) $6.3M

Inland Facility (e.g.

production, pipelines, rail)

Risk to ephemeral or intermittent waterway

$6,000 x RWCS $100M

Risk to perennial waterway

$10,000 x RWCS $100M

Pipelines $10,000 x RWCS $100M

Rail $10,000 x RWCS $100M

Introduction and Study Objectives - 5- November 2019

Table 2: Current Reasonable Worst-Case Spill Volume Definitions

Category Sub-category RWCS amount/calculation Maximum of all plan holders Tanker all 25% of total capacity 925,000 bbl

Tank Barge Large (>150,000 bbl capacity)

25% of total capacity 81,750 bbl

Small (<150,000 bbl capacity) 20,950 bbl

Non-tank Vessel

Calif or federal >7,500 bbl total oil capacity; other >6,500 bbl

Total volume of single largest fuel tank

14,465 bbl

Calif or federal 1,001-7,500 bbl;

other

1,001-6,500 bbl 7,500 bbl

501-1,000 1,000 bbl 51-500 500 bbl 11-50 50 bbl 1-10 10 bbl

Marine Facility (e.g.

terminals)

Function of multiple factors1 31,135 bbl

Offshore Platform Not drilling Function of multiple factors2 2,107 bbl

Active drilling Daily vol for 30 days from uncontrolled blowout

NA

Marine Pipeline

Onshore Function of multiple factors3 28,267 bbl

Offshore (Leak detection + shutdown time) * max flow rate + additional leakage

3,134 bbl

Small Marine Fueling Facility

Function of multiple factors4 48 bbl

MTU Total truck tank capacity 500 bbl

Inland Facility (e.g.

production, pipelines, rail)

Production facility 10% of daily production from largest producing well 70 bbl

Pipeline Function of multiple factors5 65,856 bbl

Other Function of multiple factors1 66,861 bbl

Rail <10mph 1% bulk oil6 714 bbl Rail <25mph 5% bulk oil 3,713 bbl Rail >25mph 20% bulk oil 14,994 bbl

1 Loss of capacity of in-line, break-out, and portable storage tanks not subject to Chapter 6.67 or Chapter 6.7 of Division 20, Health and Safety Code (Aboveground and Underground Storage Tank programs administered by Certified Unified Program Agencies on behalf of CA Department of Forestry and Fire Protection, Office of the State Fire Marshal) needed for continuous operation of pipelines used to handle or transport oil; plus the amount of additional spillage reasonable expected to enter waters of the state during emergency shut-off, transfer, or pumping operations if hose(s) or pipeline(s) rupture or becomes disconnected, calculated as (maximum time to discover release + maximum time to shut down flow)*(maximum flow rate) + total linefill drainage volume; plus drainage volume from piping normally not in use.

Introduction and Study Objectives - 6- November 2019

2 Total tank storage and flow line capacity; plus portion of linefill capacity subject to loss during spill, taking into account availability, location of emergency shut-off controls and hydrostatic pressure; plus amount of additional spillage reasonable expected to enter marine waters during shut-off, transfer, or pumping operations if a hose or pipeline ruptures or becomes disconnected; plus daily production for 30 days from an uncontrolled blowout of highest capacity well.

3 (Maximum time to discover release + maximum shut-down response time)*(maximum flow rate) + (largest line drainage volume after shutdown of line section); or the largest foreseeable discharge for the line section(s) within a response zone based on maximum historical discharge; or the capacity of the single largest tank or battery of tanks within a secondary containment, adjusted for containment capacity.

4 (Maximum time to discover release + maximum time to shut down flow)*(maximum flow rate) + total linefill drainage volume 5 Loss of capacity of in-line, break-out, and portable storage tanks not subject to Chapter 6.67 or Chapter 6.7 of Division 20, Health and Safety Code, needed for continuous operation of pipelines used to handle or transport oil; plus the amount of additional spillage reasonable expected to enter waters of the state during emergency shut-off, transfer, or pumping operations if hose(s) or pipeline(s) rupture or becomes disconnected, calculated as (maximum time to discover release + maximum time to shut down flow)*(maximum flow rate) + total linefill drainage volume; plus drainage volume from pipelines normally not in use.

6 All rail RWCS have a minimum of 1 tank car.

Review of Previous Research on Large California Spills - 7- November 2019

SECTION 2 Review of Previous Research on Large California Spills In attempting to project the potential costs and impacts of a hypothetical future spill, particularly one that approaches a RWCS scenario, past cases and modeling studies provide insightful data and information. These previous case studies on historical spills and hypothetical large spill scenarios provide a perspective on the magnitude of damages and the spread of oil for large spills and make it easier to conceptualize the cleanup requirements for a large spill (i.e., many miles of shoreline, large areas of oil coverage). Modeling of spill behavior–trajectory, fate, and effects-- may also help to frame a RWCS scenario that may be larger than previous historic spills. This section reviews research to date on the costs and damages for historic and modeled large spills in California, providing comments on how these events can inform the current effort.

2.1 Noteworthy Historical Spills in California

Fortunately, very large oil spills–ones that involve millions of gallons or hundreds of thousands of barrels (bbl)–are rare events. While California did experience what is believed to be the largest on-land oil well blowout over 18 months in 1910–1911 (Lakeview Gusher Number One in Kern County) (Figure 3) and a significant offshore well blowout on 28 January 1969 (Alpha 21 Platform A off Santa Barbara), there have been no other spills over 100,000 bbl in the last 50 years. The Cymric oil field “surface expression” reported in May 2019, which is currently under investigation, spilled approximately 32,000 bbl.

Figure 3: Lakeview No. 1 Gusher 19104

In the case of the Lakeview No. 1 blowout, a pressurized oil well in the Midway- Sunset Oil Field in Kern County released 9 million bbl of crude oil over the course of 18 months. The initial flow rate was

4 By unknown photographer, 1910 - https://www.georgetownenvironmentalhistory.org/ehg-blog/field-trip-to-the-midway-sunset-oil-field-california, Public Domain, https://commons.wikimedia.org/w/index.php?curid=77935501 https://commons.wikimedia.org/w/index.php?curid=77935501

Review of Previous Research on Large California Spills - 8- November 2019 reported to be 18,800 bbl per day, but peak flow reached 90,000 bbl per day. This flow rate has never been reached in any other blowouts worldwide.

In the 1969 Santa Barbara blowout Platform A, located about six miles off the coast in the Dos Cuadras Offshore Oil Field, an estimated 80,000 to 100,000 bbl of crude oil spilled over the course of three months, with the majority of the release occurring during 28 January through 7 February, and then tapering off by April 1969 (Figure 4). The flowrate was estimated to be about 9,090 bbl per day. The spill caused oiling of the coastline from Goleta to Ventura, as well as the Channel Islands. There were reports of oiled seabirds and marine mammals. This incident was noteworthy in that the public outcry and media coverage resulted in numerous pieces of significant environmental legislation, including the development of the US Environmental Protection Agency (EPA).

Figure 4: 1969 Santa Barbara Platform A Blowout5

In neither of these incidents, particularly for the Lakewood Gusher, did responders employ the sort of comprehensive spill cleanup measures that would be expected for a spill of this magnitude today. Nor was there any real effort to assess natural resource damages and conduct restoration, which is a vital part of the aftermath of spills currently. Hence neither of these incidents provides a predictive view into the response costs of a potential very large spill in California at present.

It is important to note that for tanker spills, there has never been a worst-case discharge spill (based on the release of the entire contents of a fully-loaded large tanker) in US waters. The 1989 Exxon Valdez spill of 261,900 bbl of crude oil in Prince William Sound, Alaska, which had a significant effect on large areas of Alaska, was the largest tanker spill in the US. However, the actual release only represented about 20% of the tanker’s cargo.

There have been significant spills in California in the last 15 years that may provide some important insights, including the 2007 Cosco Busan spill in San Francisco Bay and the 2015 Refugio pipeline spill.

However, it is important to note that these spills were still only relatively moderate-sized–1,276 bbl and

5 Photo: US Geological Survey. https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html

Review of Previous Research on Large California Spills - 9- November 2019

2,500 bbl, respectively–though they did both involve extensive response efforts and noteworthy environmental impacts.

The Cosco Busan spill culminated in a $44.4 million settlement, which included $36.8 million for natural resource damages. The spill killed 6,849 birds, impacted 14 to 29 percent of the following winter’s herring spawn, and oiled 3,367 acres of shoreline habitat.6 Response costs were estimated at $72.6 million, third party claims at $8.4 million, and fines at $12.4 million.

The 2015 pipeline break near Refugio State Beach in Santa Barbara County (Figure 5) resulted in heavily oiled beach areas and important cultural resource areas for California Native Americans. Note that the spill response operations were significantly challenged by the fact that this spill occurred in the vicinity of Coal Oil Point, which has natural seeps that release an estimated 155 to 167 bbl of oil per day.7 Determining the source of the shoreline oil, pipeline spill or offshore oil seeps, required the incorporation of oil fingerprinting into the response operations.8 To date, the cost associated with the Refugio pipeline spill has included $64.5 million in cleanup or response costs, $2.5 million in third party claims, and $3.35 million in fines; there is a civil proceeding ongoing currently and there will likely be more fines and penalties assessed in the future. The Natural Resource Damage Assessment (NRDA) has not yet been concluded.

Figure 5: 2015 Oil Pipeline Break at Refugio State Beach9

Even spills an order of magnitude larger may not provide that much more information. There were a few spills in the nearly 10,000 to 20,000-bbl range in the 1970s through early 1990s. Examples include the American Trader spill off Huntington Beach in 1990, involved the release of about 9,900 bbl of crude oil with impacts to thousands of birds. The 1988 Shell Oil Martinez refinery spill of 9,500 bbl had a notable impact on fish and wildlife habitat. A 1971 two tankers collided under the Golden Gate Bridge, spilling

6https://response.restoration.noaa.gov/about/media/36-million-natural-resource-damages-settlement-cosco-busan.html; https://nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=37886&inline 7 Lorenson et al. 2011; Hornafius et al. 1999.

8 Stout et al. 2018.

9 Photo: NOAA. https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html https://response.restoration.noaa.gov/about/media/36-million-natural-resource-damages-settlement-cosco-busan.html https://response.restoration.noaa.gov/about/media/36-million-natural-resource-damages-settlement-cosco-busan.html https://nrm.dfg.ca.gov/FileHandler.ashx?DocumentID=37886&inline https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html https://response.restoration.noaa.gov/about/media/natural-seeps-historic-legacy-what-sets-apart-latest-santa-barbara-oil-spill.html

Review of Previous Research on Large California Spills - 10- November 2019

20,000 bbl of heavy oil and resulting in extensive beach oiling and large wildlife impacts. The American Trader spill resulted in two natural resource damage settlements: the settlement for recreational impacts was $11.6 million and the settlement for environmental impacts was $3.45 million. In addition, the spill resulted in cleanup costs of $19.1 million, third party claims of $25.5 million, and fines totaling over $6.5 million.

In part because these spills are all unique, and have different costs associated with the spill characteristics and historical context, additional research has focused on modeling hypothetical very large or even “worst-case discharge” spills in California to help governments prepare for such an event.

For these reasons, studies that have involved modeling of hypothetical very large or even “worst-case discharge” spills in California may prove instructive with respect to the potential magnitude and scope of effects for these types of incidents.

2.2 US Army Corps of Engineers San Francisco Bay Spill Study

One study conducted for the US Army Corps of Engineers (USACE) involved the modeling of hypothetical tanker worst-case discharges in San Francisco Bay. The purpose of the study was to determine the potential costs and damages associated with such spills so that a cost-benefit analysis could be conducted. The USACE were investigating the potential removal of rock pinnacles that presented navigation hazards for deep-draft vessels in the bay.10 The hypothetical releases included those shown in Table 3.

Table 3: USACE San Francisco Bay Modeling Study Oil Spill Types and Volumes11 Oil Type 20th Percentile 50th Percentile 95th Percentile

Gasoline (Product Tanker) 1,190 bbl 6,430 bbl 30,000 bbl

Diesel (Product Tanker) 1,190 bbl 6,430 bbl 30,000 bbl

AK Crude (Crude Tanker) 2,380 bbl 14,300 bbl 71,400 bbl

Heavy Fuel Oil (Freighter) 600 bbl 2,380 bbl 9,800 bbl

The estimated spill volumes were determined using three estimated percentile spill volumes. A percentile is developed by dividing a data set into 100 equal groups in order of the value. For example, if you had data on the height of 300 fifth graders, the first percentile would have the shortest (or tallest) three heights, the next percentile would have the next three heights and down the line to the other end.

If a spill volume is the nth percentile case, it means that n% of spills will be this size or smaller. Only 100%-n% of cases will be larger. Hence the 95th percentile case for a spill volume means that the analyst anticipates that only five percent of spills will be larger than that 95th percentile spill. For example, in Table 4, showing the USACE study framework, 80 percent of spills for gasoline tankers are expected to be larger than 1,190 bbl, 50 percent are expected to be larger than 6,430 bbl, and just five percent are expected to exceed 30,000 bbl.

The 95th percentile is often used as an “almost worst-case scenario” in oil spill studies. It corresponds with the 95th percent confidence interval used in statistical analyses. The top five percent of cases are considered statistical anomalies. The volumes of hypothetical spillage modeled in the USACE study at were based on an analysis of historical tanker accident data. Note that the 95th percentile is not a worst-case discharge, which would, theoretically, involve the release of the entire tanker’s cargo. Based on

10 Etkin et al. 2002, 2003; French-McCay et al. 2002, 2003.

11 Adapted from French-McCay et al. 2002.

Review of Previous Research on Large California Spills - 11- November 2019 tanker traffic at the time of the study (2000–2001), the largest crude tankers calling in San Francisco Bay were 215,000 deadweight tonnage (DWT). Tankers of this size would carry as much as 1.43 million bbl.

Even with double hulls, which were not universal at that time (but were present on the tanker fleet entering San Francisco Bay), the worst-case discharge would be been about 700,000 bbl. (This represents 50 percent outflow from the tanker’s cargo capacity, as would be expected with double hulls, which limit the outflow as well as reduce the likelihood of an oil release in an impact accident, such as a grounding or collision.) The researchers estimated that the 95th percentile was represented by a spill nearly an order of magnitude smaller, at 71,400 bbl based on the probability distribution function of tanker spill volumes. This means that the remaining five percent of incidents would be expected to exceed 71,400 bbl and range all the way up to a maximum of 700,000 bbl for that size tanker.

Trajectory, fate, and effects modeling using SIMAP12 were conducted to determine the behavior of the spilled oil in the 12 hypothetical spills. The extent of the spread of floating oil can be seen in Figure 6 through Figure 9 by oil type. The large area of water surface and shoreline exposed to oil extends throughout San Francisco Bay and parts of the outer coast.

Figure 6: Maximum Possible Water Surface Exposure to Floating Oil (g/m2): Gasoline13

12 The SIMAP model is described in greater detail in: French McCay 2002, 2003, 2004; French McCay et al. 2015, 2018b. Assumptions and algorithms of SIMAP are fully documented in French et al. 1996, French McCay 2002, 2003, 2004 and French McCay et al.