DEVELOPING A DECISION-MAKING FRAMEWORK FOR OPTIMAL MARINE OIL SPILL WASTE MANAGEMENT AND TREATMENT by Seyedeh Mahboobeh Jafari B.Sc., Shiraz University, 2014 M.Sc., Shiraz University, 2017 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NATURAL RESOURCES AND ENVIRONMENTAL STUDIES UNIVERSITY OF NORTHERN BRITISH COLUMBIA November 2024 © Seyedeh Mahboobeh Jafari, 2024 Abstract This research explores estimating off-shore oily waste, considering waste-waste compatibility due to the heterogeneous nature of oily waste. Firstly, hyperparameters for Artificial Neural Networks (ANN), Support Vector Regression (SVR), and Improved Random Forest (IRF) models are optimized to develop a comprehensive oily waste estimation model incorporating liquid, solid, and total waste types. The results show that IRF is the most accurate model, with the lowest error indices and a higher correlation coefficient compared to ANN and SVR. This study then takes a step further to propose a waste allocation framework, which is tested using information on the Bella Bella oil spill incident in British Columbia. Incorporating treatment and receiving facilities' details, such as their location and capacity, the framework distinguishes all possible waste pathways for handling the waste from source to landfill. Genetic Algorithm (GA) is introduced to optimize waste transfer processes and successfully minimize transportation costs. The results show that the model can find the most optimized path to reduce transportation costs. The model's high customization, adaptability, and capacity to consider multiple nodes make it suitable for complex waste transfer networks, demonstrating its practicality in emergency situations. Efficiently allocating resources and ensuring cost-effective waste transportation while considering facility capacities and waste compatibility, the study holds practical implications for waste management practitioners, environmental authorities, and response teams. i TABLE OF CONTENTS Abstract....................................................................................................................................................... iii TABLE OF CONTENTS ........................................................................................................................... ii LIST OF TABLES ................................................................................................................................... ivv LIST OF FIGURES .................................................................................................................................... v GLOSSARY................................................................................................................................................ vi ACKNOWLEDGEMENT ........................................................................................................................ vii CHAPTER 1 INTRODUCTION ........................................................................................................ 1 1.1 What Is a Marine Oil Spill? ............................................................................................................. 1 1.2 Major Oil Spill Incidents.................................................................................................................. 3 1.3 Oil Spill Effects on Different Sectors .............................................................................................. 5 1.3.1 Aquatic Habitat and Ecology ............................................................................................... 5 1.3.2 Wildlife ................................................................................................................................... 5 1.3.3 Economy ................................................................................................................................. 6 1.3.4 First Nations and Local Communities ................................................................................. 6 1.3.5 Tourism .................................................................................................................................. 7 1.3.6 Human Health ....................................................................................................................... 7 1.4 What Affects Oil Spill? ..................................................................................................................... 8 1.5 Oil Spill Clean-up Strategy .............................................................................................................. 9 1.5.1 Off-shore Clean-up Techniques ........................................................................................... 9 1.5.2 On-shore Clean-up Techniques ......................................................................................... 11 1.6 Review of Canadian Petroleum Industry and Oil Shipping Activities ...................................... 12 1.6.1 Canadian Marine Oil Response System and Practices ................................................... 12 1.6.2 Legislative and Regulatory Structure .............................................................................. 13 1.6.3 National Oil Spill Preparedness and Response Regime .................................................. 14 1.7 Objectives and Significance of this Study ..................................................................................... 14 1.8 Organization of The Thesis ............................................................................................................ 15 CHAPTER 2 LITERATURE REVIEW ......................................................................................... 16 2.1 Oil Spill Waste Management and Modeling ................................................................................. 16 2.2 Research Gaps in Oily Waste Estimation and Waste Management ........................................... 19 CHAPTER 3 METHODOLOGY .................................................................................................... 21 3.1 Data Collection and Assumption ................................................................................................... 21 3.2 Artificial Intelligence (AI)-based Model Development ................................................................ 23 3.2.1 Artificial Neural Network (ANN) ....................................................................................... 23 3.2.2 Support Vector Regression (SVR) ..................................................................................... 24 3.2.3 Improved Random Forest (IRF) ........................................................................................ 25 ii 3.3 Waste Estimation Model ................................................................................................................ 26 3.3.1 Model Evaluation ................................................................................................................ 29 3.4 Waste Management/Transfer Framework ................................................................................... 30 3.4.1Problem Description ................................................................................................................. 30 3.4.2 Mathematical Modeling ...................................................................................................... 32 3.4.3 Objective Function and Constraints .................................................................................. 35 Appendix A ............................................................................................................................................ 38 CHAPTER 4 RESULTS and DISCUSSION .................................................................................. 43 4.1 Hyperparameter Optimization ...................................................................................................... 43 4.2 Waste Estimation Model ................................................................................................................ 45 4.3 Waste Allocation Framework ........................................................................................................ 46 CHAPTER 5 CONCLUSION AND RECOMMENDATIONS ..................................................... 54 5.1 Conclusion ....................................................................................................................................... 54 5.2 Recommendations ........................................................................................................................... 57 REFERENCES .......................................................................................................................................... 59 iii LIST OF TABLES Table 1.1 Twenty-two major oil spill incidents the world has seen………………….13 Table 3.1 AI-based model’s hyperparameters, their description and search space used in the Bayesian optimization ………………………………………….………37 Table 3.2 Notations used in the designed waste management framework…………….41 Table 3.3 Description of the decision variable used in the designed waste management framework…………………………………………………………………...43 Table 4.1 Optimized hyperparameters of ANN waste estimation model …….…….…52 Table 4.2 Optimized hyperparameters of SVR waste estimation model …………...…53 Table 4.3 Optimized hyperparameters of RF waste estimation model……………..….54 Table 4.4 Evaluation of the AI-based waste estimation models………………….……55 Table 4.5 Name and location of currently operating facilities in British Columbia.......56 Table 4.6 The Detailed information on oily waste handling facilities in British Columbia…………………………………………………………...……….58 Table 4.7 The identified path from source to the landfill with volume based on the Bella Bella oil spill incident in British Columbia…………………..….……61 iv LIST OF FIGURES Figure 1.1 Comparison between the number of tanker spills and growth in crude oil and other tanker trade between 1970 and 2020 (Adapted from ITOPF (2021)).... 11 Figure 1.2 Number of medium and large tanker spills from 1970 to 2021 (Adapted from ITOPF (2021))…………………………………………………...………….. 12 Figure 1.3 Canadian Coast Guard Regional boundaries (Adapted from Government of Canada (2022)…...…………...………………………………………………22 Figure 3.1 Parameters affecting the volume of generated off-shore oily waste ……..….31 Figure 3.2 Schematic structure of the ANN model…………………...…..................…...33 Figure 3.3 A Random Forest schematic view…………………….……….……………..34 Figure 3.4 Schematic view of possible transportation paths between each two nodes in the framework…………………………………………………………………….42 Figure 4.1 Results of Bayesian optimization for AI-based optimization models: a) ANN, b) SVR, and c) RF ………………………………….………………..……....53 Figure 4.2 The location of waste handling facilities in British Columbia……………….56 Figure 4.3 The common practice of oily waste management in British Columbia…...…57 Figure 4.4 Genetic Algorithm minimization cost graph……………….…………..….…61 v GLOSSARY Abbreviations AI ANN CC CCG DFO EPA IMO IRF ITOPF MPRI NOAA NSERC NP-hard PPE RMSE RMAE SRM SVR WCMRC Artificial Intelligence Artificial Nueral Network Corrolation Coefficient Canadian Coast Guard Fisheries and Oceans Canada Environmental Protection Agency International Maritime Organization Improved Random Forest International Tanker Owners Pollution Federation Limited Multi-Partner Oil Spill Research Initiative National Oceanic and Atmospheric Administration Natural Sciences and Engineering Research Council of Canada Non-deterministic Polynomial-time hard Personal Protective Equipment Root Mean Square Error Relative Mean Absolute Error Structural Risk Minimization Support Vector Regression Western Canada Marine Response Corporation vi ACKNOWLEDGEMENT I stand at the end of a very long and challenging journey, one that truly tested my determination and resilience. This thesis is a culmination of academic effort and a testament to the power of hope, resilience, and unwavering support. To my beloved family and friends, though miles apart, your love and encouragement were always close at heart. Your belief in me, even during the darkest times, fueled my determination to overcome challenges and power through. I extend my deepest gratitude to my supervisors, Dr. Jianbing Li and Dr. Youmin Tang and my esteemed committee member, Dr. Chris Opio. Your guidance, patience and support steered me towards overcoming challenges. I am profoundly grateful for your mentorship and support. The research was financially supported by the Multi-Partner Oil Spill Research Initiative (MPRI) of Fisheries and Oceans Canada, as well as the Natural Sciences and Engineering Research Council of Canada (NSERC). I would like to express my sincere gratitude to the following individuals whose invaluable assistance greatly contributed to the success of this research: Dale Bull, Senior Environmental Emergency Response Officer at the BC Ministry of Environment and Climate Change Strategy; Jeff Brady, Deputy Superintendent of Environmental Response at the Canadian Coast Guard; Scott Wright, Director of Response Readiness at Western Canada Marine Response Corporation (WCMRC); and David Ellwood, Regional Commercial Manager at Terrapure Environmental. Their provision of information and data played a crucial role in facilitating the progress and outcomes of my research. To all who find themselves navigating their storms, may my thesis stand as a humble symbol of hope and an invitation to persist and dare to dream, even when the odds are saying otherwise. Thank you to each and everyone of you for being a part of this journey. Mahboobeh vii CHAPTER 1 INTRODUCTION 1.1 What Is a Marine Oil Spill? Over the past two decades, there have been ongoing debates surrounding the transition to more sustainable energy sources. Despite this, many industries continue to rely on fossil fuels. The rapid growth of industrialization and economic expansion has led to a rise in the transportation of fossil fuels, consequently increasing the risk of oil spill incidents. A marine oil spill incident occurs when petroleum hydrocarbons are accidentally released at sea due to human error or equipment failures (Li et al. (2014); Beyer et al. (2016)). Tankers, offshore platforms, drilling rigs, and subsea piping lines are identified as the primary sources of such spill incidents (Li et al. (2014)). The International Tanker Owners Pollution Federation Limited (ITOPF) is a London-based organization that compiles statistics on global oil demand and incidents of oil and chemical spills. They also provide technical support services for oil spill response plans. This information is made available to tanker owners, their insurers responsible for covering oil pollution incidents, and international organizations like the International Maritime Organization (IMO), as adapted from ITOPF (2021). In essence, ITOPF's core services encompass spill response, analysis of claims and damages, training, contingency planning, and advisory and informational support. Figure 1.1 depicts a long-term analysis of global trends in oil demand versus tanker spill incidents. As shown, despite an increase in tanker movements from 1970 to 2020, the number of oil spill incidents has consistently decreased over these years. This positive trend is attributed to advancements in the shipping industry, coupled with stricter regulations and a sustained commitment to enhancing maritime safety and environmental protection through investment and exploration of innovative solutions. 1 Figure 1.1 Comparison between the number of tanker spills and growth in crude oil and other tanker trade between 1970 and 2020 (Adapted from ITOPF (2021)). Investigating the causes of oil spill incidents offers valuable insights for managerial tasks and risk analysis. An analysis conducted by ITOPF spans tanker spill incidents from around 1970 to 2021. Spills exceeding 700 tonnes, those ranging from 7 to 700 tonnes, and those below 7 tonnes are categorized as large, medium, and minor, respectively. Figure 1.2 illustrates the trend and volume of spill incidents in the medium and large categories during this period. As shown, the total volume of oil released due to tanker spill incidents in 2021 is approximately 10,000 tonnes, with the majority classified as medium spill incidents. While annual fluctuations are present, the overall trend shows a decline in the number of oil spill incidents. This reduction is attributed to positive shifts within the shipping industry, reinforced regulations, and governmental support for research initiatives. The following sections will provide more details regarding marine oil spill incidents, causes, environmental impacts and related regulations. 2 1970 1976 1979 1982 ^■>700 tonnes 1988 1991 1994 1997 2009 of spills per year 7-700 2012 2018 decade Figure 1.2 Number of medium and large tanker spills from 1970 to 2021 (Adapted from ITOPF (2021). 1.2 Major Oil Spill Incidents Even though the number of oil spill incidents has decreased significantly, primarily due to stringent regulations, it's essential to recognize that a single oil spill incident can still lead to catastrophic events. A summary of the twenty most significant oil spill incidents worldwide is presented in the following table. This table includes renowned incidents such as Exxon Valdez and Prestige, although they are listed further down the rank column in Table 1.1. The subsequent section will provide detailed information about the consequences of oil spill incidents. 3 Table 1.1 Twenty-two major oil spill incidents the world has seen. Rank Ship name Year Location Spill size (tonne) 1 Atlantic Empress 1979 Off Tobago, west India 287,000 2 ABT Summer 1991 700 nautical miles off Angola 260,000 3 Castillo de Bellver 1983 Off Saldanha Bay, South Africa 252,000 4 Amoco Cadiz 1978 Off Brittany, France 223,000 5 Haven 1991 Genoa, Italy 144,000 6 Odyssey 1988 700 nautical miles off Nova Scotia, 132,000 Canada 7 Torrey Canyon 1967 Scilly Isles, UK 119,000 8 Sea Star 1972 Gulf of Oman 115,000 9 Sanchi 2018 Off Shanghai, China 113,000 10 Irenes Serenade 1980 Navarino Bay, Greece 100,000 11 Urquiola 1976 La Coruna, Spain 100,000 12 Hawaiian Patriot 1979 300 nautical miles off Honolulu 95,000 13 Independenta 1979 Bosphorus, Turkey 94,000 14 Jakob Maersk 1975 Oporto, Portugal 88,000 15 Brear 1993 Shetland Islands, UK 85,000 16 Aegean Sea 1992 La Coruna, Spain 74,000 17 Sea Empress 1996 Milfor Haven, UK 72,000 18 Khark 5 1989 120 nautical miles off the Atlantic 70,000 coast of Morocco 19 Nova 1985 Off Khark Islan, Gulf of Iran 70,000 20 Katina P 1992 Off Maputo, Mozambique 67,000 21 Prestige 2002 Off Galicia, Spain 63,000 22 Exxon Valdez 1989 Prince William Sound, Alaska, USA 37,000 4 1.3 Oil Spill Effects on Different Sectors 1.3.1 Aquatic Habitat and Ecology Aquatic habitats are the first areas heavily impacted by oil spill incidents. The marine environment is a complex system of plants, animals, and their surroundings. When the environment is harmed, it often affects species in the food chain, leading to further consequences for other species (IMO (2001)). In open water, fish and whales can swim away from an oil spill by going deeper or in different directions. However, animals like turtles, seals, and dolphins, which usually live near the shore, are at a higher risk. They are affected by oil on beaches or by eating prey in oil. In shallow water, oil can also harm plants like seagrasses, kelp, and coral reefs, which many species rely on for food, homes, and breeding spots (IMO (2001)). Oil spill incidents can also happen in swamps and marshes, where there's less water movement, making the situation worse than in flowing water. Lakes and ponds are also at risk because more types of animals can be exposed to oil. This is especially true for migrating birds, as they can spread the contamination over a larger area. While rivers are usually less affected than lakes, if a river is a drinking water source, it directly threatens human health. This concerns many communities, including indigenous communities, that rely on water bodies for their needs. 1.3.2 Wildlife The species most impacted by oil spill incidents are birds, mammals, and plants inhabiting marine environments and adjacent shorelines. These organisms face various 5 threats, including direct physical contact with oil, toxic contamination, depletion of food sources or habitats, and reproductive challenges (IMO (2001)). Different species exhibit varying degrees of tolerance to oil ingestion. For instance, oil spill incidents and their vapours can be fatal to seabirds. Despite their potential resilience, once exposed to oil, these birds can accumulate it within their bodies, leading to infiltration of their nervous system, liver, and lungs. Consequently, oil enters the food chain, posing a risk of contamination to predators. 1.3.3 Economy Marine-based industries, including port operations, fisheries, tourism, and aquaculture enterprises, often experience significant adverse effects from oil spill incidents. These impacts manifest through direct losses of products due to mortality or habitat destruction, as well as restricted access resulting from harvesting bans and area closures. Furthermore, economic losses stem from decreased market demand, driven by concerns over the safety of products tainted by oil contamination (Li et al. (2014); Beyer et al. (2016)). These losses reverberate throughout the fisheries supply chain, affecting docks, processors, and supply businesses (Beyer et al. (2016)). 1.3.4 First Nations and Local Communities First Nations represent a particularly vulnerable segment of society, highly susceptible to oil spill incidents' consequences. These communities often rely on fisheries and land resources from water bodies, making them particularly vulnerable to the impacts of oil contamination in terms of polluted soils from oil pipelines and tainted seafood (Li et al. (2014)). 6 In addition to First Nations, other indigenous communities and rural locals with strong ties to the natural environment for subsistence and cultural practices are also affected by oil spill incidents. While compensatory frameworks exist for these communities, delays in compensation disbursement may adversely affect trust in governmental authorities, potentially leading to significant social unrest and societal disruption. 1.3.5 Tourism Tourism stands out as a particularly vulnerable sector in crises and disasters. This vulnerability arises from its interrelations with other industries, such as transportation and accommodation, hotels, airlines, and car rentals. Additionally, tourism is heavily influenced by external factors such as political stability, currency exchange rates, and weather conditions (Beyer et al. (2016)). Oil spill incidents near shorelines and areas populated by humans pose aesthetic concerns, necessitating safety measures due to toxic volatile vapours. Cleaning such spills is more costly and may extend over longer periods, resulting in significant losses and market decline for businesses and properties situated near beaches and waterfronts, such as restaurants, hotels, and recreational facilities. Over time, tourism in the affected area may experience crises or collapse, as these events divert tourist traffic to alternative destinations. 1.3.6 Human Health Human beings can be impacted by an oil spill incident in three significant ways, including disruption of ecological processes, resulting in direct harm. This includes ingesting seafood contaminated with oil toxins, which can lead to various health issues. For instance, consuming seafood bio-accumulated with oil toxins or breathing in oil vapours can cause 7 direct harm. Economic stressors affect individuals working in fields such as fishers and the tourism industry. According to Aguilera et al. (2010) and Major and Wang (2012), inhalation of vapours or consuming contaminated seafood can lead to harmful health effects ranging from dizziness and nausea to certain types of cancers and issues with the central nervous system. Although the long-term effects of hydrocarbon toxicity on humans are less studied, they have been associated with severe DNA degradation, cancers, congenital disabilities, reproductive defects, irreversible neurological and endocrine damage, and impaired cellular immunity (Binet et al. (2003); Aguilera et al. (2010)). 1.4 What Affects Oil Spill? Once an oil slick forms on the water's surface following an oil spill incident, it undergoes various weathering processes, including photolysis, evaporation, dilution, formation of oil-water emulsions, and biodegradation (Dave and Ghaly (2011)). Specifically, the formation of oil-water emulsion leads to significant changes in water interfacial tension, density, and viscosity. Selecting the most effective method for spill clean-up largely depends on the characteristics of the oil and environmental factors. Therefore, understanding factors such as the quantity and type of spill, weather and ocean conditions, age of the spilled oil, and ocean behaviour is crucial. The most common marine oil spills include bunker crude oil, refined petroleum products and by-products, and waste oils (Dave and Ghaly (2011)). 8 1.5 Oil Spill Clean-up Strategy When the oil spills on top of the water's surface, it will slowly drift toward the shore due to tidal activities. Spill clean-up operations must be employed to prevent a catastrophic situation after an incident. Over the last 50 years, oil spill clean-up technologies have developed extensively. The clean-up processes at the spill location (at sea) and the shore are called off-shore and shoreline clean-ups, respectively. 1.5.1 Off-shore Clean-up Techniques Off-shore response techniques are divided into mechanical/physical, chemical, biological and in-situ burning. (Dave and Ghaly (2011)). a) Physical techniques primarily involve spatially controlling the oil slick using physical barriers, thereby keeping the oil's physical and chemical characteristics unchanged. Commonly used physical barriers globally include booms, skimmers, and adsorbent materials (Fingas (2016)). Booms are floating barriers designed to restrict the movement of oil, ultimately facilitating higher oil recovery through skimmers or other response methods. They come in three main types: fence booms, curtain booms, and fireresistant booms. Fire-resistant booms are typically employed in conjunction with insitu burning. Fence and curtain booms serve a similar function, with approximately 60% submerged underwater and only 40% floating on the surface. Booms are typically around 15 meters long and can be interconnected as needed. While fence booms are lightweight, easy to handle, and reliable in calm waters, their stability in rough conditions with high waves and strong winds is limited. 9 Skimmers represent the second physical response technique commonly used alongside booms to enhance oil recovery efficiency. Depending on the skimmer type, they exhibit high stability in rough ocean conditions and can recover up to 90% of the oil. However, they are less effective when dealing with oil mixed with dispersants. Regardless of their type, all skimmers are made of oleophilic materials, which attract oil slicks that adhere to the surface. This oil can then be scraped or squeezed from the surface and collected in a small storage tank. Utilizing adsorbent materials offers another physical approach to handling oil spill incidents without altering the oil's characteristics. Hydrophobic sorbents are employed for clean-up after skimming as a final step in response operations to capture any remaining oil on the water. b) Chemical methods complement physical techniques to expedite clean-up, particularly in spill locations near shorelines or sensitive marine environments. Dispersants and solidifiers are commonly employed chemicals in oil spill clean-up, altering the characteristics of the oil. Dispersants typically consist of surfactants, which break down the oil slick into smaller droplets, facilitating faster biodegradation as they are dispersed into the deep-water column. On the other hand, solidifiers transform the oil phase from liquid to a rubber-like substance, enabling more accessible collection on the water's surface. In rough seas, solidifiers can effectively utilize wave energy to enhance dissolution in the water, resulting in a higher rate of solidification (Dave and Ghaly (2011)). c) In-situ burning represents a straightforward and rapid clean-up method. However, its usage is limited due to concerns over human health and environmental impacts associated with burning by-products, including residues and the emission of thick plumes of black smoke. This method is particularly effective in snowy 10 conditions, such as clean-ups after pipeline leaks or when oil spills occur on top of ice. One crucial requirement is that the oil slick be sufficiently thick to sustain burning and prevent it from cooling. d) Biodegradation serves as the final option for marine oil spill clean-up. While environmentally friendly and safe, its application is restricted by the slow degradation process, leading to prolonged exposure and less suitable for environments with low microbial activity. 1.5.2 On-shore Clean-up Techniques On-shore oil spill clean-ups are more straightforward than off-shore and do not require special equipment. The clean-up process can be summarized into three stages based on (ITOPF (2021)) as follows: A) Stage 1- Emergency phase: This phase revolves around a collection of floating oil near the shoreline and transfer to temporary storage, usually on the shore. B) Stage 2 – project phase: Stage 2 is a complementary phase to Stage 1, often combined. This phase involves collecting oily contaminated materials left on the shoreline. In smaller projects, any remaining oil might be left to degrade naturally. C) Stage 3- Polishing phase: this stage includes the final clean-ups and removal of oil stains if required. Removing bulked oil and treating oil-contaminated beach materials (Stages 1 and 2) typically involves using skimmers to collect the oil, which is then transferred using vacuum trucks to subsequent handling facilities. The oil removal process can be done using 11 mechanical equipment or manual methods. Pressure washing is commonly employed to ensure thorough cleaning for hard-to-access areas along the shore. 1.6 Review of Canadian Petroleum Industry and Oil Shipping Activities Canada is among the world's top oil producers, with approximately 98% of its oil exported to the United States (Mohammadiun et al. (2021)). Alberta has the largest oil sand reserves and crude oil production, followed by Saskatchewan, which produces approximately 487,000 daily barrels. Canada relies on reliable transportation methods such as railways, pipelines, trucks, and oil tankers to transport the produced oil to its destinations, including the US and other parts of the world. The selection of transportation methods depends mainly on factors such as volume and destination. Regarding marine oil shipping, around 87% of Canadian oil is transported through the Atlantic coast, the Great Lakes, the Gulf of St. Lawrence and St. Lawrence Seaway, and associated ports. In comparison, the remaining 13% is shipped through Pacific coastal ports. Seven major ports accommodate significant oil tanker traffic, including the Port of Vancouver, Port of Montreal, Port de Quebec, Newfoundland Off-shore, Port of Saint John, Port of Hawkesbury, and Nova Scotia. Given Canada's strategic geographical location, multiple international transit routes, and the continuous growth of industrial activities, there has been an increase in oil tanker transit along the Canadian coastline, consequently heightening the risk of oil spill incidents. 1.6.1 Canadian Marine Oil Response System and Practices The Canadian Coast Guard (CCG) operates as a strategic agency under Fisheries and Oceans Canada (DFO) and is entrusted with the task of providing timely and effective 12 responses to incidents involving ship-source or unknown-source pollutants in Canadian waters (Government of Canada (2022)). To facilitate efficient program delivery, Canada has been divided into three regions known as the Western, Atlantic, Central, and Arctic Coast Guards. These regional divisions were established in October 2012 to ensure swift administration of response efforts. Figure 1.3 illustrates the boundaries of the Canadian Coast Guard regions as determined in October 2012. Figure 1.3 Canadian Coast Guard Regional boundaries (Adapted from Government of Canada (2022)) 1.6.2 Legislative and Regulatory Structure Under federal legislation and various international agreements, the federal government is responsible for cleaning up any pollutants spilled in Canadian waters. 13 1.6.3 National Oil Spill Preparedness and Response Regime Since 1995, Transport Canada has taken the lead of federal regulatory agencies responsible for the regime based on the partnership between industry and government. This regime provides guidelines and regulatory structures for preparedness and response to marine oil spill incidents. In this regime, potential polluters pay for readiness. The three pillars of oil spill clean-up are Prevention, Preparedness and Response. 1.7 Objectives and Significance of this Study Understanding the quantity of generated oily waste is essential for proactive preparedness for oil spill response, enabling the determination of necessary resources such as materials and labour required for urgent response. This helps decision-makers allocate resources effectively to address oil spills promptly, as the longer the response is delayed, the further the oil spreads across the water body. This spread necessitates increased use of resources such as products and labour, significantly increasing the volume of oily waste generated. Therefore, prompt action is vital to minimize waste accumulation and effectively address the environmental impact of oil spill incidents. Therefore, the objectives of this study are twofold: a) to develop an AI-based model to effectively and accurately estimate the volume of generated oily waste, and b) to develop a waste management framework for handling the generated waste based on factors such as waste type and the availability of treatment and receiving facilities, as well as landfills, to minimize the cost of waste handling. 14 1.8 Organization of The Thesis The thesis structure is as follows: Chapter 2 consists of a comprehensive literature review. The methods, materials, results, and discussion were described in Chapters 3 and 4. Chapter 5 explains the conclusion of the entire study and the recommendations for future studies. 15 CHAPTER 2 LITERATURE REVIEW 2.1 Oil Spill Waste Management and Modeling In recent years, the focus on adequate pre-planning, risk assessment, and advancements in response techniques within the realm of oil spill management has intensified. This heightened attention stems from recognizing oil spills' profound environmental and economic consequences. Researchers such as Marta-Almeida et al. (2013) and Azevedo et al. (2014) have made vital contributions to this field, whose models have emerged as valuable tools for addressing these challenges. The models proposed by Marta-Almeida et al. (2013) and Azevedo et al. (2014) have benefitted various stages of oil spill management. Specifically, these models have played instrumental roles in pre-planning activities, facilitating more informed decision-making processes. Additionally, they have proven invaluable in conducting comprehensive risk assessments, enabling stakeholders to anticipate and mitigate potential environmental and socioeconomic impacts (Bejarano and Mearns (2015)). Moreover, these models have contributed significantly to advancing response techniques during oil spill incidents. By providing insights into the behaviour and trajectory of spilled oil, as well as the effectiveness of different response strategies, they have empowered responders to devise more effective and efficient cleanup operations (Lehr et al. (2000); Ghanbari et al. (2021)). Previous studies have underscored these contributions (Peterson et al. (2003); Dave and Ghaly (2011); Barron (2012)), which have highlighted the critical role of modelling approaches in enhancing preparedness and response capabilities. Through their empirical analyses and case studies, these researchers have demonstrated the tangible benefits of integrating modelling frameworks into existing oil spill management protocols (IMO (2001)). 16 Managing oily waste represents a crucial aspect of pre-planning and response strategies in oil spill waste management. Oily waste necessitates specialized handling and disposal procedures due to its hazardous nature and potential environmental impact. Despite the considerable attention given to oily waste management, a significant gap exists concerning estimating oily waste volumes generated during spill incidents. While preliminary guidelines for oily waste management and factors influencing waste generation rates have been documented since the early 2000s (IMO (2001); IPIECA and IOGP (2014)), more emphasis should be placed on the accurate estimation of oily waste volumes. Metcalf (2014) attempted to address this deficiency by developing a general waste management plan to investigate the impact of cleanup strategies on waste generation. However, further advancements were reported as necessary to improve the accuracy and reliability of oily waste estimation methodologies (Beegle-Krause (2005); Bergstra (2012)). Recent studies have turned to innovative approaches, such as self-learning or semisupervised learning techniques, to bridge this gap and enhance the model's accuracy. These methodologies leverage labelled and unlabeled data to improve model performance (Tkalich (2006)) iteratively. Combining information from unlabeled data with labelled data expands the labelled training dataset over successive iterations until the entire dataset is labelled. Such techniques are the foundation for machine learning algorithms, including Random Forest (RF). The applicability of self-learning approaches has been demonstrated across various scientific disciplines, including GIS and remote sensing (Fatehi and Asadi (2017); Zhao et al. (2017); Lottes and Stachniss (2017)); medical diagnostics (Kourou et al. (2015)), and groundwater investigation (Sameen et al. (2019)). These studies have highlighted the 17 potential of self-learning methods, particularly in data-scarce environments, as they require fewer labelled training data than traditional supervised learning models. Another aspect of marine oil spill response waste management involves efficiently handling the waste. Oily waste is categorized as hazardous waste, characterized by toxicity, corrosiveness, ignitability, and chemical reactivity, primarily stemming from industrial and manufacturing processes. In many countries, the manufacturing industry contributes over 75% of hazardous waste, emphasizing the criticality of its management due to environmental and human health risks. Efficiently managing hazardous waste involves collecting, transporting, treating, recycling, and disposing of it safely and cost-effectively (Xu et al. (2014); Ghanbari et al. (2021)). The integration of location and vehicle routing decisions in hazardous waste transportation and disposal was explored in the early 1980s. Various models have emerged to optimize objectives like travel time, transportation, and disposal risk. Alumur and Kara (2007) addressed optimizing hazardous waste management systems with treatment and disposal facilities, introducing realistic constraints on the properties of waste types. Alumur and Kara (2007) incorporated waste-technology compatibility constraints, while Peterson et al. (2003) considered waste-waste compatibility and disposal centers, considering government regulations and air pollution standards as constraints. Few studies have directly addressed the vehicle routing problem, while Peterson et al. (2003) tackled waste management with multiple incompatible waste types. Integrating inventory control into location and routing decisions has yet to be explored. Addressing uncertainty is crucial, as it affects planning decisions. Zhao et al. (2017) explored waste generation amount uncertainties and transportation costs. Metaheuristic 18 approaches like Genetic Algorithms (GA) have gained traction due to the Non-deterministic Polynomial-time (NP)-hard nature of location-routing problems. Despite advancements, several gaps still need to be addressed. Simultaneous optimization of location, routing, and inventory plans under uncertain conditions, multiperiod planning, and application of multi-objective solution approaches for medium to largescale stochastic problems warrant further exploration. This comprehensive literature review outlines these aspects, providing a foundation for marine oily waste management framework research. 2.2 Research Gaps in Oily Waste Estimation and Waste Management Understanding the magnitude of oily waste production during oil spill incidents is crucial for proactive response planning, facilitating the allocating necessary resources such as materials and labour required for prompt intervention. Rapid response is imperative to mitigate oil spread across water bodies, minimizing resource consumption and waste generation. Therefore, timely action is crucisl to reduce waste accumulation and effectively mitigate the environmental repercussions of oil spill incidents. Hence, this study aims to achieve two primary objectives: firstly, to develop an AIbased model for the estimation of oily waste volume, and secondly, to formulate a comprehensive waste management framework considering factors such as waste type and the availability of treatment and disposal facilities with the ultimate goal of optimizing and/or minimizing waste management costs. Based on the existing literature, this research introduces a novel, Improved Random Forest (IRF) model for marine oily waste estimation. By linking self-learning methodologies with the Random Forest algorithm, IRF aims to enhance predictive accuracy regarding oily 19 waste generation in marine contexts. To ensure the efficacy of the IRF model, this study employs Bayesian optimization for fine-tuning hyperparameters and conducts thorough performance evaluations through cross-validation. Furthermore, two conventional machine learning algorithms, namely Artificial Neural Network (ANN) and Support Vector Regression (SVR), are examined for comparative analysis to discern the optimal method for marine oily waste estimation. Moreover, a comprehensive framework for managing oily waste has been developed to streamline the allocation of various waste volumes to suitable facilities, considering their capacities and geographical locations to minimize transportation costs. This model provides responders with a structured pathway for transferring waste volumes from generation points to treatment facilities, receiving stations, and ultimately landfills, considering compatibility between different types of waste and the capacities of the respective facilities. This model provides responders with optimized routes for waste transfer, ensuring efficient management even during facility disruptions or the availability of additional facilities. 20 CHAPTER 3 METHODOLOGY This research aims to conduct comparative assessments of various well-known AIbased models, namely Artificial Neural Networks (ANN), Support Vector Regression (SVR), and Improved Random Forest (IRF), for estimating off-shore oily waste following oil spill clean-ups. ANN is often considered a primary choice among researchers due to its versatility and success in addressing many problems. Meanwhile, SVR has gained increased attention in recent years for its effectiveness in handling estimation problems comprehensively. The following paragraphs will provide data collection and background information on the abovementioned models. This section will then continue explaining how to develop and evaluate estimation models. 3.1 Data Collection and Assumption Understanding the factors influencing the volume and composition of waste generated during oil spill incidents is crucial for estimating waste. However, accessing information about oil spill incidents presents significant challenges due to limited valid sources. This study compiles a comprehensive database detailing the physical and chemical characteristics of previous oil spill incidents globally to address this issue. Extensive efforts were made to gather information through literature searches, surveys, and consultations with relevant agencies such as the Western Canada Marine Response Corporation (WCMRC), DFO, and BC Ministry of Environment and Climate Change Strategy as provided in Appendix A. This database considers various aspects of oil spill incidents, including quantity of spilled oil, location, oil type, viscosity, ocean conditions, wind speed, water temperature, response techniques, duration of off-shore response operations, and volume of recovered oil 21 and waste (Figure 3.1). These factors significantly influence response strategies, personal protective equipment requirements (PPE), and the volume of generated waste. Spilt oil Oil recovery Oil Type (Viscosity) Generated oily waste (liquid and solid) Offshore response time Ocean temperature and Wind speed Response technique (booms, dispersant, in situ burning, skimmer and absorbent) Figure 3.1 Parameters affecting the volume of generated off-shore oily waste In cases where climatic data, such as ocean temperature, were unavailable in oil spill response reports, assumptions were made based on past 20-year average water temperatures for the incident location reported on different websites such as National Oceanic and Atmospheric Administration (NOAA), US Environmental Protection Agency (EPA), and ITOPF. Sixty oil spill incidents were ultimately selected as input data for the research model. The selection criteria prioritized oil spill incidents with 90% of the required information available. While AI-based models typically benefit from larger datasets, the decision to use 60 oil spill incidents for testing purposes was deemed sufficient. This approach helps prevent 22 model overgeneralization and ensures efficient pattern recognition by excluding erroneous or irrelevant data. Notably, the model's adaptability allows for optimization based on new and existing data, promising more accurate waste volume estimations as additional information becomes available. 3.2 Artificial Intelligence (AI)-based Model Development 3.2.1 Artificial Neural Network (ANN) Artificial neural networks are complex mathematical systems that mimic the human brain's neural system. In brief, ANNs comprise three different layers: input, output, and hidden. Each layer consists of neurons or nodes as processing elements of the network. Neurons on the input layer will distribute the input information to the neurons of the next layer, hidden layer(s), where the information will be processed and then transferred to the output layer, where the results are produced (Zhang and Friedrich (2003)). Figure 3.2 shows a schematic structure of a simple ANN model. In training such a network, weighted connections among layers only occur in a forward direction from the input to the output layers. These interconnections are then adjusted by distributing errors through the layers to produce the most accurate outputs, which is called backpropagation. Hidden layers are advantageous over traditional logistic regression analysis techniques by modelling the interactions and relations among all the input parameters. An essential consideration in designing ANNs is choosing the number of neurons at each layer and the number of hidden layers, commonly known as hyperparameters (Zhang and Friedrich (2003); Wang et al. (2009)). This study has applied a new approach to tuning these parameters, which will be discussed in section 3.2. 23 Figure 3.2 Schematic structure of the ANN model 3.2.2 Support Vector Regression (SVR) Another studied model is SVR, initially designed for classification problems (Sain and Vapnik (1996)), but with the introduction of a new parameter, Vapnik’s Ɛ intensive loss function, its application has been extended to solve non-linear regression estimation (Smola and Schölkopf (2004)), and time series forecasting as well (Thissen et al. (2003); Lin et al. (2006)). SVR implements Structural Risk Minimization (SRM) rather than empirical risk minimization implemented by most traditional ANNs, which ultimately results in providing an answer that is always unique and globally optimal (Lin et al. (2006)). Similar to ANN, SVR uses an implicit feature space mapping from the dimension of the data to a possibly infinite feature space, which provides a non-linear representation of the modelled data, performed through a kernel function (Smola and Schölkopf (2004); Sadri et al. (2012); Goyal et al. (2014)). Defining kernel function and its parameters is not straightforward in the SVR network, as some settings might be prone to over-fitting or underfitting. Therefore, different kernel functions should be tested to choose the one associated 24 with the lowest error in the validation stage. This study uses a Bayesian optimization model to determine the most competent kernel function. 3.2.3 Improved Random Forest (IRF) RF, the third studied model in this work, is a classification-based regression model introduced by Breiman (2001). It is a supervised learning algorithm that uses an ensemble learning method for regression. RF operates by creating and combining many decision trees (Figure 3.3). Decision trees handle high dimensional data well, overlook irrelevant descriptors, control multiple mechanisms of action, and are amenable to model interpretation ( Svetnik et al. (2003)). Figure 3.3 A Random Forest schematic view 25 The steps of training the RF regression model can be summarized in the following steps: 1. About two-thirds of the data will be randomly selected as a training dataset called a bootstrapped dataset. 2. The remaining one-third will fall into an out-of-bag dataset, which will be used to measure the error. 3. The model's outcome is the average result of all the trees (Cutler et al. (2007)). 3.3 Waste Estimation Model As mentioned, artificial intelligence and machine learning techniques can significantly expedite decision-making processes by analyzing subtle data patterns from previous oil spill incidents. No single machine learning algorithm consistently outperforms others, as their effectiveness depends on the nature of the problem and data formats. Hence, this thesis compares ANN, SVR, and RF for off-shore oily waste estimation. Generally, developing estimation models involves establishing a formula between input and output parameters that accurately represents the problem's nature and can be extended to estimate new variables within the input parameter range the model is trained with. The effectiveness of AI-based models hinges on the proper combination of hyperparameters, which control the model's overall performance. Typically, hyperparameters are selected through trial and error, a time-consuming process with loose guidelines on their numerical range. Moreover, hyperparameters interact with each other, complicating the optimization process. To address this challenge, robust optimization methods are employed to find the optimal configuration of each AI-based model. Traditionally, grid and random search 26 methods have been used for this purpose. Grid search optimizes all possible model configurations, while random search involves iterative runs of evaluating models with randomly selected hyperparameters. However, these methods overlook historical information from previous evaluations. This study uses Bayesian optimization to find the best combination of hyperparameters by minimizing an error index (objective function) within predefined search ranges (constraints). Unlike trial-and-error approaches, Bayesian optimization leverages past evaluations to create a probabilistic model mapping hyperparameters. The Bayesian optimization procedure involves several steps, including model initialization, acquisition function optimization, model updating, and hyperparameter selection. For detailed information on Bayesian optimization, readers are referred to Snoek et al. (2012). 1. Build a surrogate probability model of the objective function; 2. Find the hyperparameters that perform best on the surrogate; 3. Apply these hyperparameters to the actual objective function; 4. Update the surrogate model incorporating the new results; and 5. Repeat steps 2–4 until max iterations or time is reached. Table 3.1 presents the name of hyperparameters along with their search range for optimization purposes for ANN, SVR and RF. To accurately compare the models, the objective function is set to minimize the Root Mean Square Error (RMSE) for all models (see Eq. (3.1)). Mean Square Error (MSE) and RMSE are widely used error indices for evaluating model performance. RMSE is chosen over MSE mainly because RMSE gives higher weight and punishes significant errors. 27 n  ( Oi − Si ) 2 RMSE = = Eq. (3.1) i =1 n Where O and S denote the observed and estimated values, respectively, to achieve a more reliable estimation of oily waste, a 5-fold cross-validation method is employed alongside Bayesian optimization to tune hyperparameters finely. This process involves randomly dividing the data in the training subset into five equal-sized parts. One part is reserved for validating the model, while the remaining four are used for training. The crossvalidation process is repeated five times for each run of Bayesian optimization. The estimated hyperparameters and their accuracy are then derived by averaging the results across all runs (Diamantidis et al. (2000)). Table 3.1 AI-based models’ hyperparameters, their description and search space used in the Bayesian optimization process Model Hyperparameter Search space Hidden Layer size (1, 20) Learning rate (10-3, 1) Kernel Function (Gaussian, RBF, Polynomial) Kernel scale (10-3, 103) Number of estimators (trees) (100, 5000) Max_features (1, 7) ANN SVR RF Max_depth (The maximum number of splits a (1, 20) tree should make before it makes a prediction) 28 3.3.1 Model Evaluation In this study, the three most common statistical error indices, including RMSE, Relative Mean Absolute Error (RMAE), and Spearman correlation coefficient, are considered to compare the performance of waste estimation models at the test stage as follow: n  ( Oi − Si ) 2 RMSE = Eq. (3.2) i =1 n n  Oi − Si RMAE = i =1 n  100 Eq. (3.3)  Oi i =1 n 6 Correlation coefficient = 1 − d 2 i Eq. (3.4) i =1 2 n (n − 1) Where n is the sample size, O and S denote the estimated waste from the AI-based model and the actual generated waste, respectively. This study uses Spearman correlation, a non-parametric correlation factor, to measure the association between the estimated waste from the AI-based model and the actual generated waste. Unlike other correlation measures, Spearman correlation does not assume a normal distribution of the variables. Instead, it evaluates based on the difference in their statistical ranks (as elaborated in Snoek et al. (2012)). 29 3.4 Waste Management/Transfer Framework Off-shore oily waste management faces more constraints related to cost and resources compared to response plans, emphasizing the importance of careful pre-planning to maximize resource efficiency. This planning can be divided into two main components: waste estimation and designing an efficient waste management framework. This section will evaluate the most effective framework for waste transfer, considering factors such as route number and cost. Accurate waste estimation is crucial for responders to arrange temporary storage and allocate resources effectively before an incident occurs. Given its significant cost implications, waste management plays a pivotal role in contingency planning for oil spill response. Efficient and optimized allocation of generated waste relies on several critical factors, including the location of the oil spill incident and available facilities, the compatibility of waste types with treatment facilities, the capacities of these facilities, the treatment rate of treatment facilities, and the number of generated points (oil spill incidents). 3.4.1 Problem Description Industrial hazardous waste management involves handling waste at its source, transporting it to treatment or receiving facilities, and ultimately disposing of it in landfills if necessary. Through consultations with hazardous waste management contractors in British Columbia (BC), such as Terrapure, it has been recognized that waste transportation constitutes the most costly and time-consuming aspect of waste management. This challenge is exacerbated when dealing with multiple types of waste, as considerations must be made for compatibility with facilities and transport vehicles, with some waste types posing risks of dangerous reactions if co-transported. Hence, understanding the available waste treatment network is crucial for optimizing waste transportation efficiency. The key focus of this 30 optimization lies in determining the most effective routing system to minimize transportation costs. BC has been selected as a case study for this study due to its relevance to the research funded by DFO for waste management purposes on the West Coast. On a typical day without spill incidents, approximately three to four trucks transport around 60 tonnes of waste to landfills in BC and Alberta. Consequently, proactive planning is essential to transport waste to each facility based on its current waste load and capacity to prevent congestion and additional storage costs. The efficient utilization of temporary storage or receiving facilities' capacity constitutes the second component of this study. Each facility is designed to accommodate a finite amount of waste. Lastly, the third component involves selecting the optimal facility based on its capacity and location among other potential facilities. Therefore, a network incorporating all these components can aid decision-makers in promptly managing generated waste, mainly since prolonged collection and transportation times result in increased contamination and storage costs. The formulation of this model aims to optimize transportation routes to wastecompatible facilities using waste-compatible vehicles to minimize total costs. Several assumptions are considered in designing the model, including the impact of uncollected waste at generation nodes and unprocessed waste at treatment facilities leading to additional storage costs, the determination of the number of vehicles at each generation node, and ensuring that the amount of waste transferred by cars does not exceed their capacity. 31 3.4.2 Mathematical Modeling With the assumptions outlined, a non-linear model is developed to identify the optimal waste management strategy from the generation node to the final destination (landfill), considering facility capacities and waste types to minimize transportation costs. Before delving into the model's design, defining the terms and notations utilized for the modelling process is essential, as detailed in Table 3.2. Based on the assumption above, a non-linear model is programmed to find the optimum way of handling waste from the generation node to the ultimate destination (landfill), considering the capacity of facilities and waste type to minimize the transportation cost. Before designing the model, the terms and notations used for modelling are stated in Table 3.2. Table 3.2 Notations used in the designed waste management framework Index Parameter Value/unit Description t Number of treatment facilities r Number of receiving facilities s Number of generation nodes (source) l Number of landfills Vsol 40,000 m3 Truck capacity (solid waste) Vliq 60,000 m3 Vacuum truck capacity (liquid waste) TR 10% Treatment rate at solid treatment facilities 5 cc Transportation cost CAD/km.tonne TCt 750 tonne/day Capacity of treatment facility number t 32 RCr 1200 tonne/day Capacity of receiving facility number r LCr Inf Capacity of landfill facility number l DIStr km Distance between two nodes As previously stated, the primary objective function of the model is to identify the most optimized path within a marine oily waste management network. Consequently, all feasible paths between each node for waste originating from a particular node until it reaches its designated destination are considered design variables (Figure 3.4). For instance, the volume of waste transferred from source 1 to treatment facility one is denoted by ZST(1,1), as per the notation outlined in Table 3.3. Subsequently, additional decision variables are defined based on the number of generation nodes (sources), treatment plants, receiving facilities, and landfills. Figure 3.4 Schematic view of possible transportation paths between each two nodes in the framework 33 Table 3.3 Description of the decision variables used in the designed waste management framework Decision Variable Description ZST(1,1) The volume of waste transferred from source 1 to treatment plant 1 ZST(1,2) The volume of waste transferred from source 1 to treatment plant 2 … … ZST(1,t) The volume of waste transferred from source 1 to treatment plant t ZST(2,1) The volume of waste transferred from source 2 to treatment plant 1 ZST(2,2) The volume of waste transferred from source 2 to treatment plant 2 ….. … ZST(s,t) The volume of waste transferred from source s to treatment plant t ZSR(1,1) The volume of waste transferred from source 1 to receiving facility 1 …. … ZSR(s,r) The volume of waste transferred from the source s to the receiving facility r ZSL(1,1) The volume of waste transferred from source 1 to Landfill 1 … … ZSL(s,l) The volume of waste transferred from source s to Landfill l ZTR(1,1) The volume of waste transferred from treatment plant 1 to receiving facility 1 ... … ZTR(t,r) The volume of waste transferred from the treatment plant t to the receiving facility r ZTL(1,1) The volume of waste transferred from treatment plant 1 to landfill 1 ... … ZTL(t,l) The volume of waste transferred from the treatment plant t to the landfill l ZRL(1,1) The volume of waste transferred from receiving facility 1 to landfill l … … ZRL(r,l) The volume of waste transferred from the receiving facility r to the landfill l 34 3.4.3 Objective Function and Constraints As discussed in the previous section, the complexity of the problem involves numerous decision variables that cannot be readily solved. These issues fall under the Nondeterministic Polynomial-time hard (NP-hard) problems, which are computationally challenging and time-consuming to address using traditional algorithms. This challenge is exacerbated by the exponential growth rate of feasible solutions over time, a phenomenon known as combinatorial explosion (Hoang (2008)). Consequently, various meta-heuristic algorithms have been developed to tackle such optimization problems effectively. One prominent meta-heuristic optimization approach is the Genetic Algorithm (GA), which has successfully addressed complex optimization challenges. Originating in 1975, GA is inspired by natural selection in biological evolution. In GA, a population of feasible solutions evolves through selection, recombination, and mutation, akin to genetic alterations aiming to enhance solution accuracy. The critical steps of a GA optimization problem are as follows: • Initialization: A population of feasible solutions is randomly generated. • Evaluation: Each solution is assessed against a fitness function, typically representing an error index quantifying its effectiveness in addressing the problem. • Selection: A subset of solutions is chosen based on their fitness function values, expecting that subsequent generations will yield improved solutions. • Recombination: Selected solutions undergo recombination operations such as crossover or mutation to generate a new population of solutions. 35 • Replacement and iteration: The new population replaces the previous one, repeating steps 2 to 5 for a specified number of generations or until a satisfactory solution is attained based on the fitness function. GA's main advantage lies in its ability to handle many feasible solutions while remaining time-efficient. However, GA's performance heavily relies on the randomness of the initially generated solutions, underscoring the importance of running the model multiple times to achieve a robust solution. In designing an optimized transportation framework, the primary objective function aims to minimize transportation costs and determine the optimal volume and path for waste generated from a predefined source location to the ultimate destination (landfills). The objective function can be formulated as follows: Min ( f z ) = s t s st st s sr s =1 t =1 t r l  Z  DIS  cc +   Z  DIS  cc +  Z  DIS  cc + r sr sl s =1 r =1 t Eq. (3.5) sl s =1 l =1 l r l  Z  DIS  cc +  Z  DIS  cc +  Z  DIS  cc tr tr t =1 r =1 tl tl rl t =1 l =1 rl r =1 l =1 Z Total = Z s1 + Z s 2 + ...+ Z s Eq. (3.6) Where the variables and decision parameters are described in Table 3.2 and Table 3.3. The optimization model's constraints are outlined s follows: 1. Whenever a collection vehicle enters a node, it must exit it towards another destination. 36 2. Compatibility between wastes for transfer within a collecting vehicle is not a concern. 3. Waste types must align with the technology at treatment facilities and the types accepted by receiving facilities and landfills. 4. The quantity of waste allocated to a route and/or facility must be equal to or greater than zero. 5. The total waste volume across all directions must match the overall volume that needs to be managed, excluding considerations for waste type. 6. The volume of waste being transferred to a facility should be at most of the facility's capacity. 7. some waste residues can be directed to a receiving facility or a landfill following treatment. 37 Y (1190 barrels) N Dispersent (Y/N) Bioreme diation (Y/N) Other Y N In-situ burning (Y/N) Capacity Number Reference Volume of recovered oil From absorbents From skimmers Other Quantity of generated liquid waste (Barrels) N N N 1000 (Barrels) 9000 (Barrels) • Eathern Berms used N N N N 14,000 (Barrels) • Bardge deck w/ 2 vacum tra ilers • Low pressure washing at shoreline • Severe wave actions and thick in seawe ed: Booms and skimmers inefficient. • Vacuum trucks designed for thick liquid to remove oil from piers • Hot water effective removal at rocky shoreline • Oily seaweed raked and transfered with front end loaders manually • Eathern Berms effective for clogging wtlands until high tides and surf applied 20000 (tons) ~ 80,000 (tonnes) ~ 20,000 (tonnes) • 7000 personelle (military) Y N N Y Multiple, non specified weight Multiple Multiple 4 21 (Days) 27.5 11 (mph) South West Y Y N Y oleophilic chalk 650 (Metric tons) 2.5 (Mile) multpile (Years) 12.4 Y Y N 26 (kmh) North East 1400000 (Barrels) • Trenches dug for high tide oil collection • Eathern Berms Used Y N N Y 25 (Miles) • ARAMCO 21 107 (Days) 22.7 3 (m/s) N Y • United States Coast Guard • Saudi Arabias Ar abian American Oil Company • Meterorological and Enviromental protenction Admin • Royal Commission for Jubayl and Yanbu • Various International help N Y 72 (Days) Kuwait Crude Oil • USCG Marine Saftey Office Galveston • Rie del-Peterson Enviromental Sevice • Malin Envirometnal • Clean Gulf Association • O'Brien Oil Pollution Services • Garner Ennviromental Servies and N N N Y Multiple, non specified weight 2 locations Multiple • 2 Open Water Oil Containment Recovery System Skimming Barriers • 7 Marco Class V • 3 large off shore 15 + • 25 support boom / tow vessels 55 (Days) 10 (Days) 0 N N Y 35 (knots) N • ITOPF N Y 1 (Day) • No. 5 Oil (Vaccum Oil / Catalytic Feed Stock) 14 (Days) 9000000 (Barrels) 38 N N Y 2800 (Barrels) 3,126 (Barrels) 50 (Barrels / 100 yrds) • 2,800 (cubic yards) Including Sorbents • 2180 (tons) of oiled logs • 2000 + volunteers • Oil Spill Case Histories: Summaries of signifcant U.S and Interna tional Spills, NOAA Hazardous Materials Response and Assessment Division • Radar set up for skimnemrs to work continusouly for 5 weeks • 6000 drift cards deployed between spill • 3 Cana dian Barges supplied cranes to and coast remove trapped oil from logs • In suit burning 2 attemps, w/ material • High pressure water cannons to clean composed of fine grained fumes silicia onshore logs from 2 Foss T ugs particles treated w/ silane (product • Oiled fishereis cleaned with sorbants, Tullanox 500), Sustained bunred failed cages , nets and catwalks replaced by ARCO N N N N • Multiple, non specified weight • one location 3,0000-6,0000 (squa re yard) 17800 (Feet) N 4 + Vessels • 15 Vessels • 2 Tanks • 3 Log Barge • 2 Other Barges • 1 Helicopter • 11 Vaccum Trucks 9 7 (Years) 8.2 Y Y N South-Southe ast 10 (feet) N • U.S Coast Guard • NOAA • Woods H ole Oceanographic Institute • National Marine Fisheries Sevices • Univeristy of Rhode Island • U.S Geologial Survey Y N • No. 6 Fuel Oil • Cutter stock 10 (Days) 183000 (Barrels) 183000 (Barrels) Tank Vessel 29 mile s south of Nantucket Island, Massachusetts, U.S.A 15-12-1976 Agro Merchant 126 (Days) 8.2 N Y N 10 (konts) East and West 6-10 (feet) Y • Canadian and U.S. Coast Guard • ARCO Spill Response Team, Long Beach, California, U.S.A Y N Alaska North Slope Crude Oil 1 (Days) 814000 (Barrels) 5690 (Barrels) Tank Vessel Port Angeles Harbour, Washington, U.S.A 21-12-1985 ARCO Anchorage • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • Oil Spill Case Histories: Summaries of • ITOPF signifcant U .S and International Spills, • Cedre NOAA Hazardous Materials Response and • Guidlines for Oil Spill Waste Assessment Division Minimization and Managment, • Operational a spects of the response to the IPECA Report Series. Vol. 1 ARCO Anchorage oil spil, Levine Robert A. • Effectiveness of Mechanical 1987. pp 3-7 Recovery for Large Offs hore Oil Spills, 2021 Dagmar Schmidt Etkin, Tim J. Nedwed. 0.00 0.00 10,000 (cubic meters) ~ 15,000 (tonnes) ~ 30,000 (tonnes) • Highpressure hot water washing Y Y N Y Multiple 49 (Days) 0 (Days) 17.9 Y Y N 11.9 (mph) • Portuguese Navy • International T anker Owners Pollution Federation • European Economic Community Task Force N Y Mexican Maya Crude Oil 49 (Days) 235000 (tonnes) 175000 (Barrels) Tank Vessel • Tank Vessels • Facilite s Pipelines • Platforms 16476 (Barrels) Madeiran Archipelago, Portugal Persian Gulf, Kuwait • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of signifcant U.S and International of signifcant U.S and International • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of signifcant U.S and International Response and Assessment Division of signifcant U.S a nd International of signifcant U .S and International of signifcant U.S and International Response and Assessment Division of signifcant U.S and International of signifcant U.S and International Spills, NOAA Hazardous Materials • Effectiveness of Mechanical Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials • IOTPF Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Respons e and Assessment Division Recovery for Large Offshore Oil Response and Assessment Division Response and Assessment Division Response and Assessment Division • Guidlines for Oil Spill Waste Response and Assessment Division Response and Assessment Division Spills, 2021 Dagmar Schmidt Etkin, Minimization a nd Managment, Tim J. Nedwed. IPECA Report Series. Vol. 12 63,000-78,000 6500 (bbl) • Sea Barrier N N N Vacum T rucks (Y/N) Quantity of generated solid waste offshore storage 2000 (Feet) , 24" oil snare N 3 lcoations 500 + (Feet) multiple Type/brand of sorbent pad N Multiple (non-specified) 2 steamboats • 2 Barges • Tugboat 29.3 Y Y N 140 (mph) Multiple, non specified weight • Gaima 1 6 Tugboats 100 (Days) 19 (Days) 61 (Days) 3 (Days) 3 (Days) 3 (Days) 16.3 11.9 (km/h) • USCGM Arine Saftey Office Group Los Angeles-Long Beach • USCG Atlantic Strike Teams • Hess Oil Virgin Islands Corpoeration Power Plant Personal, Contracted by U.S Coast Guard Y Y N Y N Alaskan North Slope Crude Oil Aragon 29-12-1989 • Tank Barge • 4 Cargo Tanks 19-01-1991 Arabian Gulf / Kuwait Galveston Bay, Texas, USA 28-07-1990 Apex 3417 Barge, Apex 3503 Barge • Arabian Light Crude • Iranian Crude • Buncker C 1619048 (Barrels) 591458 (Barrels) 0 (Days) 1619048 (Barrels) Tank Vessel Brittany, France 16-03-1978 Amoco Cadiz 9458 (Barrels) Tank Vessel Huntingoton Beach, Californaia, U.S. A 07-02-1990 American Trader N Y Heavy Cr ude Oil 0 (Day s) 10000 (Barrels) 5 Storage Tanks Port Alucroix Limetree Bay, St Croix, U.S. Virgin Islands 20-09-1989 Amerada Hess Oil Co. Storage Tanks No. Sorbent pad Type/brand of absorbent boom No. absor bent boom No. conta inment boom Type of skimmer No. Skimmer Onshore Offshore 26.7 N Cloudy (Y/N) Y 15-20 (knots) East and North East Clean Gulf Assoication Y N 1 (Day) No. 6 Fuel Oil 7 (Days) 11900 (Barrels) • Venezuelan Merey Crude • Pilon Crude. 20.3 N Y Sunny (Y/N) Rainy (Y/N) 26 (km/h) Y Genoa Port Authority N Y Kuwait Crude Oil Wind velocity Wave height Wave period Calm (Y/N) Sudden (Y/N) Continouse (Y/N) Number of responders Response Method No. response vessels Response time Ocean Temp. (C) Weather condition Ocean condition Response Agency Incident type Oil Pr oduct Time lag for starting response No. of cargo 2 (Days) 65000 (Barrels) 7350 (Barrels) Volume of spilt oil. Tank Vessel Savannah River, Garden City, Georgia, U.S.A 2 Tank Vessels Tank Vessel Calcasieu river Bar Channel. 11 miles SE Cameron Louisianna, U.S.A Geona, Italy Vessel Type Amazon Venture 05-12-1986 Location Alvenus 10-10-1977 Date of incident (dd-mm-yyyy) 30-07-1984 Al Rawdatain Name of Incident Appendix A Arrow (Barr els) Assimi N Y Light Iranian Crude Oil 0 (Days) 379000 (Barrels) 379000 (Barrels) Tank Vessel Oman 07-01-1983 1,905 (Barrels) • 150 people • Ice, damns and ducks, hinderd fruther spread of the oil N N N Y Multiple, non specified weight 20 000 (feet) Multiple Locations • Tug boats • 10 Barges • 11 vaccum trucks • 3 cranes • Dozen small work boats • 6 helicopters 34 (Days) N N Y • Vessel moved 200 miles from shore sink • Spill and fire started two different explosions burning a majoority of cargo N N N N • 2 tug boats • Overflight watch 0 (Days) 28 (Days) 24.1 N Y Y 17 (knots) North 1 (meter) Y • U.S Coast Guard, Pitsburg, • Oman Council: Concil of National Srtike Force LANTAREA Conservation from the enmviroment Strike Team and prevention of pollution • U.S Enviroment Protenction (CCEPP) Agency (Regions III, IV, V) • Sultanate of Oman Navy (SON) • U.S Departments of Interiot • Sultanate of Oman Air Force (Philadelphia, Chicago) (SOAF) • National Oceanic and Atmosphere • Royal Oman Marine Police N Y Diesel Fuel 2 (Days) 23810 (Barrels) Facility Monogahela River, West Elizabeth, Pennsylvania, U.S.A 02-01-1988 Ashland Petroleum Company 0 (Barrels) 0 N/A N/A • No countermeasures towards response were taken. N N N N N?A N/A N./A N/A N/A N/A N/A • 5 Airplanes • 7 Merchant vessels 0 (Days) 56 (Days) 0 N N Y Y • U.S Coast Gaurd • Canadain Coast Gaurd N Y Unleaded Gassoline 252429 (Barrels) 252429 (Barrels) Tank Vessel Southeast of Cape Race, Newfoundland, Canada 22-04-1988 Athenian Venture (gallons) 1000 (gallons per hour) 190,000 519, 302 (gallons) 400 (cubic yards) • Moping tehcniques for beach cleanup. Type: Mark II-9D-Pt, Mark-II-4E) • Filter Fence and sorbent pads were palced to collect drifint oil (3 sorbent pillows) N N N Y Sawdust, Sorbent C 3 sorbent pillows, non specified weight 9600 (Feet) • 30' by 110' Barge style Exon Skimmer • Clean Channel Industry Skimmer • M/V Lady Alice (capacity 170 gallons) 2 28 (Days) 31 (Days) 20 Y Y Y 30-35 (knots) Southeast 3-4 (feet) N • U.S Coast Guard • Brine Service Company • Clean Channel Industries N Y • Louisiana Crude\ • Buncker C 2 (Days) 10000 (Barrels) Tank Vessel Upper Galvestone Bay, Huston Ship Channel, Texas, U.S.A 09-03-1973 Bayou Lafousche / Barge PC 2901 Betelgeuse Borag N N N N 2000 (Feet) • Rheinwerft Skimmer 1 tugboats 19.5 Y N International Tanker Owners Pollution Federation N Y No. 6 Fuel Oil 6 (Days) 213690 (Barrels) Tank Vessel Keelung, Ta iwan 05-02-1977 30 (tons) 20,000 (meters cubed) • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • Hay spread to absorb oil at shoreline • Explosion created a fire that burned oil • 2 four inch oil mops until Januaray 9th • Locals collected oil mannually with • Using escallop-dredging boats to recover buckets suiken oil N Y (260 Barrels) N Y Hay Multiple, non specif ied weight approx. 250 (Meters) • Gulf Oil Company Bay 1 • 10 escallop-dredging boats • Tugboat(s) • Aircrasft w/ diserpa nt sparying capabilites 10.3 Cork County Council Y Y Mixed Arabian Crude Oil 300000 (Barrels) 14720 (Barrels) Tank Vessel Bantry Bay, Ireland 08-01-1979 • Oil Spill Case Histories: Summaries of signifcant U.S and • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of International Spills, NOAA Hazardous Materials Response of signifcant U.S and International • Oil Spill Case Histories: Summaries signifcant U .S and International Spills, and Assessment Division Spills, NOAA Hazardous Materials of signifcant U.S and International • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of NOAA Hazardous Materials Response • Arrow and Kurdistan marine oil spills Crown Land Oily Response and Assessment Division Spills, NOAA Hazardous Materials of signifcant U.S and International signifcant U.S and International Spills, and Assessment Division Waste Management Sites. Nova Scotia Enviroment • The Ashland Oil Spill of Januaray Response and Assessment Division Spills, NOAA Hazardous Materials NOAA Hazardous Materials Response and • Investigations in Ba ntry Bay • Kurdistan and Arrow Oily Waste Disposal Sites 1988: EPA Perspective • T anker Asimi - A Case History, Response and Assessment Division Assessment Division Following the Betelgeuse Oil Tanker Hadleyville, Little Dover, Fox Island and Sandpoint. Nova • Ashland Oil Spill, Floreedde, PATerence M, Hayes Disaster. By R. J. R. Grainger, C. B. Scotia Transportation and Infras tructure Renewal Case History and Response Duggan, D. Minchin and D. O'Sullivan. Hydrogeological Investigation Evaluations 37000 2300000 (liters) • Oil slicks dispersed by wave action and diserpants • Steam Cleaning of oil wharves and boats • Oil removed by the abosption through peat moss placed ijn water ; recovered with slick lickers (skimming system) ; used at locations where high viscosity of oil remained • Gravel and sand mix beaches cleaned with mevchanical equiment which mixed oil deeper into the ground ; Scrappers effective on snad beaches N Y (10 tones) N N • Slick Lickers: Oleophillic- Belt Type Multiple (see cell AJ:14) 66 (Days) -0.7 Y Y N 34-37 (knots) N • Department of Military, Transporatation, and Coa st Gaurd of Canada N Y Bunker C Oil 10 (Days) 77000 (Barrels) Tank Vessel Chedabuctio Bay, Nova Scoita, Canada 04-02-1970 0 (Days) • Nigeria n Crude • Blended Crude (high naphtha content) 2 (Days) Other Reference Volume of recovered oil insigifiga nt 101141 (Barrels) • Hay Filter Dam N • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries Response and Assessment Division of signifcant U.S and International of signifcant U.S and International • Cleanup Efficiency and Biological Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Effects of a Fuel Oil Spill in Cold Response and Assessment Division Response and Assessment Division Weather. The Bouchard #65 oil spill in Buzzard Bay. By Eric Schrier 2080-4160 (liters) From absorbents 77,070 (liters) From skimmers Other 576 000 (kg) 19990 (liters) 19000 (liters) Quantity of generated liquid waste Capacity 1 pla toon like float Quantity of generated solid waste offshore storage Number • 6558 personelle (highest rate) N • Oiled sand eemoved by front end loaders • Hay was spread on beaches to absorb oils N • Ice effectively contained the oil and prevented shoreline impact • Ice Removal • Drilling hole s into pocket of oil under ice to recover Bioremediation (Y/N) N Y (5,500 gallons) N Dispersent (Y/N) N N Y (15,000 lirers) In-situ burning (Y/N) Y Y Y Vacum Trucks (Y/N) Hay • 16 mnen (onshore cleanup) • 9 firefighters • Coast Guard Cutter Valiant • 400 ppl beach cleanup • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • Burhmah Agate - Chronology and containment operations. Timothy W Kana, Edmond P. Thompson, Robert Pavia • NOAA Incidents 6700 (Barrels) 162100 (Barrels) • 14 Cpast Guard personnel • 2 Coast Guard P ollution inverstigators • other representatives 3095 (Barrels) 6927 (Barrels) 1000 (barrels) 2 • 250 men • Washing techniques / absorbant materials to clean bayous N Y (2000 Barrels) N N • Straw (27,000 bales) • Urefoam Pads (450: 19"x16") Multiple, non specifed weight • J M Boom • Navy Boom • Kein Filtration Boom 14 (Miles) • 6 Swedish Skimmers • Chase Skimmers 12 • 60 Vessels including: - Containment Barges - 5 Skimmer Boat - Skimmer Barge - Oil Herder #32 drilling barges 48 (Days) 11-22 (knots) N • Chevron Oil Compan • Regional Repsonse Team • U.S Coast Guard N Y Crude Oil 22 (Days) 65000 (Barrels) Platfrom 11 Miles East Mississippi River Delta, Loisiana, U.S. A 10-02-1970 Chevron Main Pass Block 41 • Cleaning strated with asrobant materials and skimmers • All wells on platform were capped • T wo chemical diserpants sprayed on platform. N N N Y absorbant materials Multiple, non specified weight Multiple Large Exxon Skimmer Multiple • Fireboats • Mulitple Vessels: - M/V Captin F.L - Fransworth • Four Non-Specifed Boats 30 (Days) 26.1 Y Y Y 33 (knots) N • U.S.A Coast Guard, and Pacific Gulf Strike Team Y Y • Santa Maria Crude Oil • Cata lytic Cracher Feedstock 110,000 (Barreks) 203000 (Barrels) Tank Vessel Deer Park, Texas, U.S.A 01-09-1979 Chevron Hawaii Concho 368 (Barrels) 20,191,100 (Barrels) 285 (tones) • Off loading operations aided by High Capacity Mohn Pumps • 655 Barrels of concentrated and 952 barrels of sta ndard diserpants applied effectively by vessels and aircraft • Onshore high pressure water washing , Cooper Pegler beach spray unit and backpack sprayers • Manual cleanup onshore N Y (1607 Barrels) N N Corinthos N N N N foam N/A N/A Multiple 22.2 5.9 (mph) • U.S. Coast Guard Philadelphia and Atlantic Strike Team • Philadelphia Fire Department • U.S. Army and Navy Y N Algerian Crude Oil (Type 2) 77 (Min) 315000 (Barrels) 266000 (Barrels) Tank Vessel Delaware River, Marcus Hook, Pennsylvania, U.S.A 18-09-1992 35,714 (Barrels) • Off loading and diving Operations • Lightering with an Air-Deliverable • Ashphault-like residue on shoreline Anti-Pollution Transfer System manually removed with shovels Pumps, and Framo and Thune Eureka cold pumping systems. N N N N • JBF Dynamic Incline Plane 3001 • 3003 Self Propelled Skimmers • Bantry Bay Skimmer with a monosuction pump • 2 Komara Miniskimmer • 2 Oceanpack • Seakskimmer Multiple 2 • Barge(s) • ADAP TS 6 (Days) • U.S. Coast Guard and Atlantic Strike Team • Ocean Salvors Company Y N No. 6 Fuel Oil (Type 4) 0 (Days) 207269 (Barrels) 1786 (Barrels) Tank Vessel Kill Van Kull, New York, U.S.A 19-01-1981 6 • British Petrolum Boats • 37 vessels with diserpant spraying capabilites • Tugboat (United Towing Tug Gaurdsman) • Multiple Vessels: 4 (Days) 2 (Days) 12 • Her Majesty's Coast Guard • British Potrolum Y N Heavy Iranian Crude Oil 0 (Days) 257250 (Barrels) 21990 (Barrels) Tank Vessel Irish Sea, South Wales 12-10-1978 Chritos Bitas E-24 1000 (Gallons) 185 (Gallons) • Salvage operations to recover suken cargo N N N N 13 (Days) 201 (Days) 12.4 Y Y N 0 0 0 0 0 • Caping of Rig • Wave Action • 2000 plastic wrapped drift cards N N N N N/A N/A N/A N/A N/A N/A N/A N/A 8 (Days) 6 Y n N 4-6 (feet) gale force winds 5-6 (feet) N • Red Adair • Norwegian State Pollution Control N Y Ekofisk Crude Oil 25 (knots) East N • Ma rine P ollution Control • U.S. Coast Gaurd United States • Navy Supervisor of Salvage • Captin of the Port New York Y N • No. 6 Fuel Oil • Diesel Fuel 0 (Days) N/A 20000 (Barrels) No. 6 Fuel 2 (Days) 202381 (Barrels) Platfrom Norway, North Sea 22-04-1977 Ekofisk Bravo Oil Field 71 (Barrels) Tank Vessel Block Island Sound, Fishers Island, New York 22-11-1985 Eleni V 03-09-1988 ESSO Puerto Rico 1000 (tonnes) 100 tons •Blow up of aground bow se ction of vessel 8000 barrels Y Y N M N/A N/A N/A N/A Multiple (non Specified) N/A Non Specifed • 5 Tugs • 22 Vessels with spraying capabilites N/A 11.95 Y N N 11.1 (mph) Y Dutch and England Authorites N Y Heavy Fuel Oil 0 (Days) 117, 280 (Barrels) 52500 (Barrels) Tank Vessel 10 (Barrels) 55 (gallon) 2 • Diving and suction operation • Oil disapated with river currents. N N N N Multiple, non specified weight • Utility boats 29.4 N N N 3.9 (mph) South Y • U.S. Coast Gua rd and Athlantic Strike T eam • Marine Saftey Office New Orleans Y N Carbon Black Feedstock 23000 (Barrels) Tank Vessel Norfolk, Southeast Coast of England Mississippi River, Louisiana, U.S.A 06-05-1978 • Lighterning • Oil snares to pick up oil on and in be tween ice chuncks • Mannual Removal: Shovels, pitch forks, rakes, plastic bags, steam generators, high velocity water streams on onshore impacts N N N Y Oil Snares Multiple, non specified weight N/A N/A 300 + (feet) • Lockhead Skimmer • Macro Skimme • Cold Weather Skimmer Multiple • Tug boats • ADAP TS 10 (Days) 3 N N N 17.1 (km/h) N • U.S. Coast Gaurd Atlantic Strike Team • Sea Land Enviromental Enginering Company N Y No. 6 Fuel Oil 1 (Day) 60000 (Barrels) 10000 (Barrels) Tank Barge Hudson River, New York, New York, U.S. A 04-02-1977 Ethel H (II) 39 • Oil Spill Case Histories: Summaries of signifcant U.S and International • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials of signifcant U.S and International • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: of signifcant U.S and International Response and Assessment Division • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of signifcant U.S and International • Oil Spill Case Histories: Summaries Summaries of signifcant U.S and Spills, NOAA Hazardous Materials •Alpline Geophysical Asooicates. Oil of signifcant U.S and International Response and Assessment Division of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International Spills, NOAA Hazardous Materials of signifcant U.S and International International Spills, NOAA Response and Assessment Division Poltuion Inciendent, P Latform Spills, NOAA Hazardous Materials • NOAA Incident News Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Response and Assessment Division Spills, NOAA Hazardous Materials Hazardous Materials Response • NOAA Incidents Charlie, Main Pass Blaock 41 Field. Response and Assessment Division • Effectiveness of Mechanical Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division • NOAA Incident News Response and Assessment Division and Assessment Division • Cabo Pilar Grounding and Oil United Stated Enviromental • NOAA Incident News Recovery for Large Offshore Oil • Cedre • NOAA Incident News Spill. Capt. Francisco Pizarro, MNI Protection Agency National Service Spills, 2021 Dagmar Schmidt Etkin, Center for Enviroment Publications. Tim J. Nedwed. 1971 N/A N/A N/A 1200 (plastic bags) per day • Coretext 9527 (582 gallons) and Slickgone LTE (3380 gallons) of disperpants used 612 (feet) of OWOCRS 2 ADAPTS N N Y N N N/A N?A N./A N/A N/A • Beaches cleaned with Vacalls • Most oil burned in ship or surroudning water • Lightering • Beach Cleanup done by manual removal with front end loaders and dump trucks N Y Y N/A Multiple (non-specifed weight) Type/brand of sorbent pad Hay N/A Multiple, non specifed weight N/A 10612 (Feet) N/A • (4) Skimming barriers (OWOCRS) • (1) Lockhead skimmer • Marco Class V Skimmer • 2 OWOCRS 73 (Days) 6.3 Y Y N N • Wifsmuller Salvage • Chilean Maritme Authoraty Y N ENAP Crude Oil 2 (Days) 40900 (Barrels) Tank Vessel Putna Davis, Chile N/A N/A Multiple, non specified weight Cabo Pilar 08-10-1988 6 • Tugboats • Cranes • Outriggers • 11 differnt vessels rotated between OWOCRS • Multiple Vessels: 27 (Days) 92 (Days) No. Sorbent pad Multiple Multiple Y Y • Cle an Water Inc. • Smit International Inc. • U.S Navy Supervisor of Salvage N Y Good Year Boom (containment) Multiple (non-specified) No. absorbent boom 27.6 North East • Regional Response Team • U.S Coast Guard Atlantic Stike Team• • Marine Saftey Office P ittisburgh Y N Type/brand of absorbent boom Multiple (non-specified) • Navy Marco Skimmer • Lockhead Clean Sweep Skimmer • Endless Rope Skimmer 3 • Vacum Trucks • Skidmounted Vacum Units • Tug boats • Barges 24 (Days) No. containment boom Type of skimmer No. Skimmer Onshore Offshore Y Cloudy (Y/N) 5.1 Y 7-22 (knots) 1.2 (meters) N • U.S. Coast Guard • U.S Enviromental Protection Agency • Cannon Engineering • Coastal Servies Inc • Jetline Services • USCG Atlantic Strike T eam N Y N Y N Rainy (Y/N) Sunny (Y/N) Wind velocity Wave height Wave period Calm (Y/N) Sudden (Y/N) Continouse (Y/N) Number of responders Response Me thod No. response vessels Response time Ocean Temp. (C) Weather condition Ocean condition Response Agency Incident type Kuwait Crude Oil Transmix : • Gasoline • Kerose ne • No. 2 Fuel Oil No. 2 Home Heating Oil 1 (Day) Oil Product Time lag for starting response 400,000 (Barrels) 76191 (Barrels) 254761 (Barrels) Tank Vessel No. of cargo Pipline 1790 (Barrels) 73600 (Barrels) Galveston Bay, Texas, USA Allegheny River, Knapp Run, Pennsylvania, U.S.A Tank Vessel 01-11-1979 Burmah Agate 30-03-1990 Buckeye P ipline 1932 (Barrels) Sa o Sebastiao, Sao Paulo, Brazil 09-01-1978 Brazilian Marina Tank Barge Buzzards Bay, Massachusetts, U.S.A 28-01-1977 Bouchard #65 Volume of spilt oil. Vessel Type Location Date of incident (dd-mm-yyyy) Name of Incident • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • Response to the Janurary 1990 Arthur Kill Heating Oil Spill. Lt. Brian G. Bubar, J. R. Czarnecki 140,00 (Gallons) 19,000 (Barrels) • 150 Exxon personnel • 400 contract personnel • Trenches were dug for late r vaccuming • Slow Release Fertilizer applied to few impacted shorline • Eathern Berms Y N N Y pompoms Multiple, non specified weight 400,000. (feet) 60,000 (feet) • Self propelled Skimmer • Ma rco Skimmer • JBF Skimmer 10 • 40 vacuum trucks • 70 boats • 4 helicopte rs 72 (Days) 72 (Days) 5.1 20.2 (mph) • Alyeska Pipline Service Company • Exxon Company • Federal State and Local Agencies: • Including Burrard Clean Operations • Exxon Company • U.S. Coast Guard • Clean Harbors Cooperative • National Oceanic Atmosphere Adminastration • New York State Department of Enviromental Conservation • New Jersey State Department of Enviromental Protection • New York City Parks and Recreation • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • IOTPF • 25 Years After the Exxon Valdez Oil Spill: NOAA’s Scientific Support, Monitoring, and Research • 30 Years After Tht Exxon Valdez, Grea t State of Alaska • Exxon Valdez Alaska Oil Spill - CDC • Guidlines for Oil Spill Waste Minimization and Managment, IPECA Report Series. Vol. 12 • WCMRC 3700 (Barrels) 10,000 (tonnes) 25000 (tones) • Barge 12,000 (Gallons) 1 • 11,000 personnel • Fish Hatcheries Boomed • Use of Hopper Dredge • Hoses spraying Seawater flushing Oil from Shorelines • Natural Cleaning P rocesses • Eathern Berms applied Y Y (45,000 gallons) Y (15,000 gallons) Y pompoms Multiple, non specified weight Rolled P ads, Snare Multiple 100 + (miles) • Weir Skimmers • Oleophillic Disc Skimmers • Egmolap Brand Paddle Belt Skimmer (Egmolap II) • Rope Mop Skimmers Multiple •1,400 Vessels • 100 Aircraft 891 (Days) 4 (Days) 5 Y Y N up to 70 (knots) 18 (feet) N N Y Prudhoe Bay Crude 35 (Hours) 53 000 000 (Gallons) 240500 (Barrels) Tank Vessel Bligh Re ef, Prince William Sound, Alaska, U.S.A 24-03-1989 Exxon Valdez Y N No. 2 Heating Oil 0 (Days) 13500 (Barrels) Pipline Arthur Kill, New York 02-01-1990 Exxon Bayway Refinery • Oil Spill Case Histories: Summaries of signifcant U.S and International Spills, NOAA Hazardous Materials Response and Assessment Division • Investigations in Bantry Bay Following the Betelgeuse Oil Tanker Disaster. By R • NOAA incident News • Oil Exploitation and Marine Pollution Evidence fomr thr Niger Delta, Nigeria. S.O Aghalino* and B. Eyinla. P g 177-180. 0 (Barrels) N/A N/A • 8,800 gallons Gold Crew deserpants • Sand beaches cleaned by wave and tide action • Some oil evaporated in the water • Oil retained in Santana mangrooves N Y (8,800 gallons) N N N/A N/A N/A N/A N/A N/A N/A • 2 rigs 5+ (Months) N/A 29.2 Red Adair Corperation N Y Nigerian Crude 12 (Days) 200000 (Barrels) Platfrom Niger Delta, Forcados, Nigeria 17-01-1980 Funiwa No. 5 N N • Majoirty of oil burned off, no cleanup required • Wave action • Vessel Towed for lightering Y (10,900 gallons) N • Oil sprayed with detergent when on water but clsoer to shoreline • Bio Errosion from variety of browsing molluscs and chiton rapidly cleaned oil reefs • Lightrerning and skinking vessel. Dispersent (Y/N) Bioremediation (Y/N) Other N Reference From absorbents From skimmers Other 10 (Days) Jupiter 8.6-10 (mph) Y • Bangor County Fire Department • USCG personnel • Williams Boots and Coots Company N Y Unleaded Gassoline 30 (Min) 20000 (Barrels) Tank Vessel Saginaw River, Ba y City, Michiagan, U.S.A 16-09-1990 6 8060 (Barrels) 2500 (cubic yards) 7900 (feet) • Oil recovered on wate r by mopcats and vacuum trucks • Cutting Oiled vegitation aided by inner marsh grass cleaned by tide • Offloading • Change in tide naturally cleaned marshes • Low pressure washing from man made futures into sorbent booms collected by surface skimmers • Sea walls cleaned by hand wiping and scrubbing • Cource grain sand beaches cleaned with rakes and shovels. • 442 personnel N N N Y 2000 (feet) N N N Y Multiple, non specified weight 59,695 (feet) • DELBAY Skimmer • Navy Skimmers 11 (Days) 13.3 15000 (Barrels) Oily water emulsion 55 (gallon) Multiple Oil Drums • 200 + personnele • When booms were short in supply, discarded floating hoses from oil loading systems were used for precautionart booming intake of area N Y N Y • disgarded SMB hose. • Light booms • Navy Booms Multiple Multiple 8 (Months) 35,700 (Barrels) 26140 (cubic yards) • numerous volunteers • army representatives • Manual removal of oil and debris • Diving Opera tions attemped to recover sunken oil and control underwa ter le akage. N N N Y Multiple Multiple 14.9 643 (Barrels) • Manual cleanuop of beached include the use of front end loaders , raking of oiled shells and seaweed, and sand removal w/ dump trucks • Low-P ressure washing • Polyurethane sheeting covered picnic tables, shoreline and grass in cleanup N N N Y Multiple, non specified weight 10400 (feet) • U.S Navy Skimmer Multiple • Barges • CG-32303 water side pollution partol boat 26.7 • No Reported cleanup up oiled shoreline needed N N N Y Multiple Multiple N/A 7 (years) 13.8 N/A N/A N/A N/A N/A • Wave action dispe rsed the oil enough N N N N N/A N/A N/A N/A N/A N/A N/A • 2 cutters: - Mallow - Jarvis 5 (Days) 262 (Barrels) • Flush and foam sewers • 770 (barrels) off loaded from lightering • manual removal at shorelione N N N Y Multiple 6 (Months) 7.3 40 • Oil Spill Case Histories: Summaries of signifcant U.S and International • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries of signifcant U.S and International Response and Assessment Division of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International Spills, NOAA Hazardous Materials • The Hasbah 6 (saudi Arabia) Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Response and Assessment Division Blowout: The Effecrts of an Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division • NOAA Incident News International Oil Spill As Experienced in Qatar. Joseph A. CM. van Oudenhove 0 (Barrels) N/A Quantity of genera ted liquid waste Volume of recovered oil N/A Capacity Quantity of genera ted solid waste offshore storage Number N N/A N/A Y Sea Grass, Sargassum Type/brand of sorbent pad In-situ burning (Y/N) Multiple, non specified weight No. Sorbent pad N/A N N/A Type/brand of absorbent boom N/A N/A N/A N/A 41 (Days) • 27 Vacum trucks • Barge • Tug • 47 small boats • 32 vechicles 0 (Days) • 6 Utility Boats • Coast Guard Cutters • 4 aircraft • Tugboats 7 (Days) 0 (Days) Vacum Trucks (Y/N) N/A No. absorbent boom Multiple (non Specified) N/A No. containment boom N/A Type of skimmer • Esso Margarita Tanker • Foam T ankers • Pump Truks No. Skimmer Onshore 22.2 N N Y 30 N Y N Offshore SE N • Regional Response Team • Science Advisoiry Group • U.S Coast Guard N Y Light Crude Oil 237600 (Barrels) Tank Vessel Pacific Ocean, 50 miles North Lisianski Ha waiian Islands, U.S.A 17-01-1977 Irene's Challenge 21.7 (km/h) N • Enviromental Protection Service • Regional Enviromenta l Emergeinces Team • Canadian Coast Guard • Atmospheric Enviroment Service N Y Bunker C 0 (Days) 175000 (Barrels) 43900 (Barrels) Tank Vessel 18.2 (km/h) N • U.S Coast Gaurd • Minnesita P olution Control Agency • U.S Enviromental Protection Agency, Region V • Minnesota Department of Emergency Management • Minnesota Department of Natural Resources Y N Crude Oil 0 (Days) 40476 (Barrels) Pipline Grand Rapids, Minnesota, U.S.A 93 km north-east of Sydney (Nova Scotia), near Ca pe Breton Island, Nova Scotia, Canada Lakehead Pipeline Company 03-03-1991 Kurdistan 15-03-1979 MCN-5 Mega Borg 29 (Days) • The OSC • The Olympic Tug and Ba rge Co. of Seattle Wa shington N Y 10-15 (Knots) Y • U.S Coast Guard in Galveston • USCG Saftey Office N Y less than 1 (Hour) Angolan Palanca Crude Oil 1 (Week) 819,500 (Barrels) • Heavy Cycle Gas Oil • Intermediate Fuel Oil • Marine Diesel Oil 100000 (Barrels) 1604 (Barrels) Tank Vessel Gulf of Mexico, 57 miles, Southeast of Galveston, T exas, U.S.A 08-06-1990 • 9874 (Barrels) Heavy Cycle Gas Oil • 524 (Barrels) Intermediate Feul Oil & Marine Tank Barge Guemes Channel, Shannon Point, Washington, U.S.A 31-01-1988 Nestucca N • U.S Natioanl Park Servie • Washington State Department of Ecology • U.S Coast Guard • Canadian Coast Gaurd Y N Buncker C 8 (Day s) 69000 (Barrels) 5500 (Barrels) Tank Barge Grays Harbour, Washington, U.S.A 23-12-1988 • Workers used squeegees to push oil along ice ro a removal location • Ice blocks were sprayed with hot water towash out oil and recovered with skimmers N N Y Y Multiple Multiple 16 (Days) N/A N/A (diving ooperations) no oil was recovered • Salvage Operations N N N N Multiple 0 0 • 809,500 (Barrels) Ightering • 13,023 (Barrels) Tar onshore • Numerous Volunteers • Much of the oil evaporated or burned (50%) • Firefighting vessels equiped with sea water and foam application • Lightering • Manual oil pick up on shoreline Y Y (11,300 gallons) Y Y 36" 6000 (feet) 12 + • U.S Navy Supervisor of Salvage Skimmers 3 • 2 cutters • 6 Firefighting vessels • Over 50 commercial vessels • USCG Air Eye Aircraft with radar • 2 tug boats • 2 commercial C-130 aircraft • Large inverted plane type skimmer • Oleophillic belt type skimmer • Small Oleophillic Rope-Type Skimmer • T ug Boat(s) 12 (Days) 32-27 N N Y N N N 585 (tons) 11,344 (Barrels) 118 (cubic yards) • Onshore manual cleanup • Sorbents used to scarope and absorb oil from boulders, ellgrass, and beach sand • Oil form dama ge tank pumped into • Passive cleaning using anchored oil an empty tank snares continued for two months • Filter Fence in marsh areas within the Olympic National Park effectivelt protecting the area after shoreline deemed clean • On the West Coast Tra il oiled logs burned on oiled gravel burning to N N Y Y non specified N Multiple, non specifed weight • Petromesh • Oil Snares • Pompoms 4400 (feet) 11 (Days) 0 (Days) 29.6 N N N 11.1-10.2 (mph) Multiple, non spe cifed weight 0 0 0 119 (Day s) 0 (Days) 9.9 N Y N Y • Corpus Christi Area Oil Spill Association • Miller Enviromental Service • Garner Marine Services • O'Brien Oil Pollution Services • United States Coast Guard • Atlantic Strike T eam Y N Beatrice (North Sea) Crude Oil 1 (Hour) 625,000 (Barrels) 15350 (Barrels) Tank Vessel Inner Harbour, Corpus Christi, Texas, U. S.A 13-07-1988 Nord P acific 8 (tons) Open water Oil Containment and Recovery System • 300 personnel from Nataional Guard • Numerous Volunteers • Lightering Operations • Offshore Manual Removal N N N Y Oil Snares Multiple, non specifed weight • Goodyear 36" • 36' Sea Curtain Multiple • Applications and 8 pounds bioaugmentation INOC 8162 mixed with water, following 20 ounces of fertilixer Mira cle-Gro and repaeted aplication of Miracle-Gro a week later. Applied to oiled tidal wetland grass bla des with hand spraying equipment Y N N N non specified Multiple, non specifed weight None None 3 • Inverte d Plane T ype Skimmer • Oleophilic Belt Skimmer • Oleophilic Rope-Type Skimmer Multiple • Skimming Vessel - North Sounder - Petro Retrieve • T ug 2 (Weeks) 0 (Days) 17.8 5.5-6.3 (mph) Y • L & J Vaughtco International • U.S Fish and Wildlife Service N Y Crude Oil 1 (Week) 20 (Barrels) Platfrom Se al Beach, California, U.S. A 31-10-1990 Seal Beach Well Blowout • Suction Skimmers • Floating Weir Skimmers • Delaware Bay and River Cooperative Skimmer Delbay (belted incline plane system) • Clamshell buckets • Hopper Barge • Fishing Vessel w/ stern tawl net 1 (Year) 4 (Day s) 0 Y Y N 8.4 (mph) light winds Y • Multi Agency Local Response Team • U.S Coast Guard • Atlantic Strike Team No. 6 Oil (heavy industrial grade) 0 (Day s) 452,000 (Barrels) 73310 (Barrels) Tank Vessel Delaware River, South of Marcus Hook, Pennsylvania 24-06-1989 Presidente Rivera • Oil Spill Case Histories: Summaries of signifcant U.S and International • Oil Spill Case Histories: Summaries Spills, NOAA Hazardous Materials of signifcant U.S and International Response and Assessment Division Spills, NOAA Hazardous Materials • Arrow and Kurdistan marine oil Response and Assessment Division spills Crown Land Oily Waste • Oil Spill Case Histories: Summaries • The tank barge MCN-5: Lessons • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries Management Sites. Nova Scotia of signifcant U.S and International in Salvage and response operations. of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International Enviroment Spills, NOAA Hazardous Materials Cdr. Gregory N. Yaroch & Lt. Cdr. Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials • Kurdistan and Arrow Oily Waste Response and Assessment Division Gary A. Reiter Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division Disposal Sites Hadleyville, Little • Sunken Oil Oil Dete ction and Dover, Fox Island and Sandpoint. Recovery. API TECHNICAL Nova Scotia Transportation and REP ORT 1154-1 FIRST Infrastructure Renewa l EDITION, FEBRUARY 2016 Hydrogeological Investigation 2,500 (Barrels) 1,004,000 (plastic bags) 55 (gallon) Multiple • Bow section towed 200 miles from Nova Scotia and sunk • Stern section towed to Port Hawkesbury Nova Scotia to recover remaining oil N N N N sorbent material Multiple, non specified weight Multiple • Barge (s) -0.1 Y Y N 6-10 (feet) 17 (km/h) N • Turkeish Navy • Diector of the Marmara Sea District • International T anker Owner Pollution Federation Ltd. Y N Es Sider Crude Oil 714,760 (Barrels) 687785 (Barrels) Tank Vessel Instanbul, Turkey 15-11-1979 Independenta 7 (mph) 9.5 (mph) • Enviromental Coasta l Polliution Cleanup Service • Need a Diver contracted by the U.S Coast Gaurd • Regional Re sponse Team • USCG Gulf Strike T eam N Y Bunker C Oil, Light Diesel 1 (Day) 952 (Barrels) Non-Tank Vessel Port Sutton ChannelT ampa Bay, Florida, U.S.A 10-05-1978 1-2 (feet) 6.5 (mph) • Harbour Master in Ge noa Italy • Interntiona l Tankers Owners Pollution Federation • Itallian Coast Guard • Smit - Tak contracted by ARAMCO • Bahrain Petrolum Company under ordered from the Bahra in Government • Qatar General Petrolum Corporation • Gulf Area Oil Companies Mutual • U.S Coast Guard • Hakensack Meadowlands Development Commission 5 (m/s) N Y Iranian Heavy Crude Oil 1000000 (Barrels) 142857 (Barrels) Tank Vessel Genoa Ita lly Y N Crude Oil 5 (Days) 100000 (Barrels) Platfrom Gulf of Arabia: 250 km NW Qata r, 140 kkm N Saudi Arabai Haven 11-04-1991 Y N No.6 Fuel Oil 0 (Days) 47619 (Barrels) Hasbah 6 02-10-1982 10 (knots) Southwest 17-21 (knots) North-Northwest Y • U.S Coast Guard • OSC, Fedrally Funded Various Responders form near by Platforms off the Gulf of Mexico Y Y N Ninian Crude Oil 0 (Days) 530659 (Barrels) 10357 (Barrels) Facility Hackensack, New Jersey, U.S.A Delaware River, Marcus Hook, Pennsylvania, U. S.A Tank Vessel 26-05-1976 Hackensack Estuary, T ank Farm 28-09-1985 Grande Eagle Y N Cloudy (Y/N) 24.22 13.7 (mph) N N/A Y N 30 (Mins) Arabian Crude Oil Rainy (Y/N) Sunny (Y/N) Wind velocity Wave height Wave period Calm (Y/N) Sudden (Y/N) Continouse (Y/N) Number of responders Response Method No. response vessels Response time Ocean Temp. (C) Weather condition Ocean condition Response Agency Incident type Oil Product N/A 119000 (Barrels) No. of cargo • Venezuelan Crude Oil • Diesel Fuel 7000 (Barrels) Time lag for starting response Tank Vessel 37700 (Barrels) Gulf of Mexico 15-08-1975 Globtik Sun Tank Vessel Eleuthera Island, Bahamas 07-03-1968 General Colocotronis Volume of spilt oil. Vessel Type Location Date of incident (dd-mm-yyyy) Name of Incident N N Dispersent (Y/N) Bioremediation (Y/N) Other N In-situ burning (Y/N) Reference Volume of recovered oil From absorbents From skimmers Other 7830 (Barrels) N N N N 26191 (Barrels) • 7 divers N N N N Multiple, non specific weight Multiple Multiple IXTOC 4 200 000 (ft) Y Y Y Y Y ( ~1290450 bbl) Y Y N/A Multiple, non specific weight • Good Year Boom • Multiple (non Specified) Y • Sand Berms • Oil Mops • Over 10 (miles) • T iger Booms • Long double chamber booms 13 300 000 (ft) • ~ 1,000 (barrels) N/A 0 0 23-02-1977 Hawaiian Patriot • Cedre • ITOPF • Effectiveness of Mechanical Recovery for Large Offshore Oil Spills, 2021 Dagmar Schmidt Etkin, Tim J. Nedwed. 0 0 0 0 0 N/A 0 0 • Wind and Wave action • Evaporation N N N N N/A 0 N/A 0 0 N/A 0 • Merchantman Philippine Bataan (Vessel) 0 (Days) 0 (Days) Y Y N N N/A Y N light Indonesian Crude Oil N/A 99,000 (tonnes) 50,000 (tonnes) Tanker 580Km West of Hawaiian Islands 41 • ITOPF • Other Significant Oil Spills in the Gulf of • NOAA Incident News "Deep Water Horizon"• "BP Mexico. NOAA Office of Response and Gulf of Mexico Spill Response Accelerating". BP Restoration & Emergency Response Division. • Cedre • The Worlds Worst Oil Diasters. • Cedre. • Effectiveness of Mechanical Recovery for Large CNBC. 2010 • NOAA Incident News Offshore Oil Spills, 2021 Dagmar Schmidt Etkin, Tim • Cedre • Ixtoc I: A Case Study of the World's Largest J. Nedwed • IT OPF Oil Spill. Allen P ress on behalf of Royal • Smithsonian Ocean Pollution: The Gul Spill • Herald Journal, Nov 11, 1988 Swedish Academy of Sciences. A. Jernelöv, O. • WCMRC Effectiveness of Lindén. • • Report of Investigation into the Circumstances mechanical recovery for large • Ixtoc I Oil Blowout. S.L Ross, C. W. Ross, F Past Spill Incidents, British Columbia Surrounding the Explosion, Fire, Sinking and Loss of offshore oil spills. D. Etkin, T. Lepine, R.K Langtry Eleven Crew Members Aboard the Mobile Offshore Nedwed • Ixtoc I Oil Spill. H. Fernadez. Stanford 2017. Drilling Unit: Deepwater Hroizon . United States Coast • World Infromation Systems. Ixtoc I Well Gaurd Blow-out and oil spill. • Comparison and Assessment of Waste Generated • WCMRC 27,000 (liters) 161,000 (bbl) 0 [784000 - 833000} (Barrels) Pumping at well head ~ 792399 (barrels) 0 Removed at wellhead: 23,000 (metric tonnes) 10,000 cubic yards 0 N/A 162,260 (tonnes)*up to 2014 89,200 (tonnes) * Up to 2014 N/A N/A • Natural Dispersion N N N N N/A N/A N/A N/A N/A N/A several throusand (feet) N/A Multiple • 'Frank Mohn A/A FRAMO ACW-400 Skimmer(s) • Open Water Skimmer • Shell Oil CO. SOCK Skimmer • Cyclonet 150 open Sea Skimmer • Vortex Weir Skimmer In 2010: • Coast Guard Helicoptors in 2010: 26 Skimming Vesssels • Russian Weather Ship, Passat, • Bulk Carrier, Maritime Wisdom 0 (Days) 0.3 Y Y N • (400' x 100' ) x 26 barges ROV: TREC • T ug Boat • 186' supply. Vessel • Supply Vessels • Multiple Barges 9 (months) 25- 27 N N N • 7000 various vessels: - USCG cutters - Private offshore supply vessels 1455 (Days) 87 (Days) 21.2 N N N 25 (feet) 44 (mph) > 11 (sec) N • Canadian Coast Guard Y N 0.6 (feet) 12 (knots) NW Sept 20, 1979 Y • Petroleos Mexicanos (PEMEX) • Oil Mop Inc. • NOAA • Red Adair • Martech International of Houston • Daivaz • Conair Aviation • United States Coast Guard N Y North Sea Crude Oil 1 (hour) Crude Oil 132,157 (tonnes) N/A 925 099 (Barrels) Tanker / Vessel Canada 10-11-1988 T/V Odyssey •8 (days) • USCG: 2 (Months ) ~ 3,332,000 (Barrels) Well Bay of Campeche, Mexico 03-06-1979 6 (Knots) 218 degreees SSW <1 (Foot) Y • British Petrolum • NOAA • USCG • EPA • Transocean • Smit Salvage Americas • National Response team • U.S Coast gaurd Gulf Strike T eam N Y Monaco Crude Oil Night of 700,000 (gallons) diesel 4,900,000 (Barrels) Oil Rig Macondo Prospect, 66km of the coast of Louisianna, Gulf of Mexico 20-04-2010 Deepwater Horizon • ROV used to seal one of smallest valves causing leaks • Pumping the leak using congtainment chambers • Drilling of 2 releif wells • Vessel Initally sourrounded with 3 (fialed) • Injecting concrete into well to slow spillrate layers of containment boom • tool tube was successfully inserted into the riser lying • Corexit 9527 Diserpant • Off Loading • Flushing of beaches in select areas on the seafloor, methanol was injected to prevent the • Manual Removal on Texas Shorelines captuing run off with booms. formation of methane hydrates • Capping of Wellhead • Natural Flushing of Water • Containment cap to slowley close the wells valves • Natural Degration • High energy onshore flushed by and stop the leak completely. • Pumping 100,000 steal and lead balls into the tidal cycles anf subsequent storms • Injecting drilling fluids into the well head ocean casuing flow to slow • Dumping various objects into the well head to cap • Volunteers • + 47,000 P ersonelle (2010) • Martech response included 50 personnel on • National Gaurdsmen • including 13 WCMRC staff site • Inmantes from the • 1,100 Louisiana National Guard troops State Adult Correctionsl Institute N N N Y Multiple, non specifed weight 3,000 (feet) Multiple Multiple 2 (Days) •Western Canada Marine Response Corporation • QM International Enviromental and Industrial Services •Department of NAtional Defence • Vancouver Oile Driving • Diesel Oil 30,000 (liters) Barge Plumper Bay, Esquimalt Harbour, BC, Canada 8-05-2016 Esquimalt Harbour • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries • Oil Spill Case Histories: Summaries On-Water Clean-Up In Plumper of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International of signifcant U.S and International Bay, Vancouver Pile Driving: Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Spills, NOAA Hazardous Materials Marine Genral Contractors Response and Assessment Division Response and Assessment Division Response and Assessment Division Response and Assessment Division 714 (Barrels) 25000 (gallons) 95000 (gallons) Quantity of generated liquid waste 14,000 (Gallon) 35,000 (plastic bags) Capacity 1 Quantity of generated solid waste offshore storage Number N N N N None None None None • Draining of fresh water marshes • Oil tracking via over head flights • Weed Cutters removed oiled • Manual onshore removal using vegetation avoiding to damage root shovel, rakes, sorbents, • Lightering and rhizomes in fresh and salt water andscrubbing. • Sinking of vessel marshes • Unmanned remotely operated • Sealing of small cracks in sucken • High P ressure warm water Vechicle to survey and the use of its bottom of barge with wooden plugs washibg of rocky intertidal shoreline mechanical arms to a 3-inch suction • No cleanup action occured • Filter fences w/ pompomp were hose inot the suken vessels portholes used across sloughs which were to to pump oil from vessel narrow for skimmers. Sorbent Y pompoms Type/brand of sorbent pad Vacum Trucks (Y/N) • Ro-Boom Multiple, non specifed weight Multiple, non specifed weight Multiple Multiple • 7 rescue ships. • Several USCG vessels • U.S Coast Guard Cutter Sedge • Tug Salvage Chief • Remotely Operated Vechicle 3 (Days) 15.7 N Y Y • 2 tugs • Remotely Opperated Vechicle No. Sorbent pad Multiple • Federal On 5-10 (knots) N •Various Cleanup Contractors Scene Cordinator • Atlantic Strike Team • U.S Coast Gaurd N Y 4 (Days) 23 (Days) (Min) 195238 (Barrels) No. 2 Heating Oil 11. 0 (Days) 36 (Days) 5.2 Y Y N 50-60 (knots) 20-25 (feet) N • United Marine Tug and Barge, Inc. Seattle • U.S Coast Gaurd N Y Diesel Fuel 12 (Days) 6873 Tank Vessel Newport, Rhode Island 23-06-1989 World Prodigy 38 (Days) Type/brand of absorbent boom Multiple No. absorbent boom • Marco Class I Skimmer • U.S Navy Maro Class V Skimmer • Supplied by Clean Bay Inc. • Weir / Pump Skimmer Multiple 107 (Days) No. containment boom Type of skimmer No. Skimmer Onshore Offshore 10.1 N Cloudy (Y/N) N 5.9 (mph) Y 0 (Days) 5.5-6.3 (mph) • Canadian Coast Guard • U.S Coast Guard (Canadian / U.S joint Marine Pollution Contigency Plan) • Members of the Nuu-Chah-Nulth tribe in British Columbia • U.S Coast Gaurd • Clean Bay Inc. • National Oceanic and Atmosphere Adminastration. • Enviromental Protection Agency • U.S Navy • U.S Navy Supervisor of Salvage • California Depoartment of Fish Y N Y • Intermediate Fuel Oil • Diesel Oil • Lube Oil • Bilge Oil N Y San Joaquin Valley Heavy Crude Oil Rainy (Y/N) Sunny (Y/N) Wind velocity Wave height Wave period Calm (Y/N) Sudden (Y/N) Continouse (Y/N) Number of responders Response Method No. response vessels Response time Ocean Temp. (C) Weather condition Ocean condition Response Agency Incident type Oil Product 0 (Days) 47620 (Barrels) Time lag for starting response No. of cargo 7143 (Barrels) • 6500 (Barrels) Intermediate Fuel Oil • 2166 (Barrels) Diesel Oil Barge 47620 (Barrels) Non-Tank Vessel Facility Aleutian Islands, Alaska 8700 (Barrels) Neah Bay, Washington, U.S.A Carquinez Straits, Martinez, California, U.S.A 26-12-1988 UMTB 283 Volume of spilt oil. 22-07-1991 Tenyo Maru 23-04-1988 Shell Oil Complex Vessel Type Location Date of incident (dd-mm-yyyy) Name of Incident • IPIECA Oil Spill Waste Minimization and Management 2,900 (m^3) 49,900 (m ^ 3) * Soil: 73,200 (m^3). * Debris, non harzourdous: 1,600 (tonnes) * Debris, hazardous 9,200 (m^3) N N N Y Mutliple N Y Crude oil with benzene diluent 3,100 m ^3 Pipeline Marshall, Michigan, USA Kalamazoo River Pipeline • WCMRC • Marinetraffic.com Marine transportation safety investigation report M16P0378. Transport Saftey Board Canada 110,131 (liters) Oily water 119,000 (L) diesel / lubercants: Salvage operation • 200 crew • Salvage and Pumping Operations N N N Y Sorbent Pads Seven 208 L drums N/A Multiple (non specified) 300 m • (2x ) Lamor LORS-2 Skimmers (32.8 t/hr total) 2 • 45 Vessels including mentioned: - WCMRC Clowhom Spirit 1 - WCMRC vessel Eagle Bay - CCG John P. Tully - Small Heiltsuk vessel - North Arm Diligent 40 (Days) 11 Y Y N 10 (knots) SE 0.5 (meters) 3 (meters) N • WCMRC • Canadian Coast Guard • Kirby Offshore Marine Operating LLC, Houston, TX, U.S.A. • Environment and Climate Change Canada • Fisheries and Oceans Canada • British Columbia Ministry of Y N Diesel Fuel 18 (hours) 129,000 ( Liters) 100,000 (Litrers) • WCMRC • Chief: Diesel spill shows cleanup weaknesses. Times Colonist 16,000 (liters) 11,000 (liters) • Shoreline Manual Cleanup N N N Y Sorbent Pads Multiple, non specific weight N/A Multiple (non specified) Multiple (non specifed) NA Multiple >4 (months) 4 (Days) high winds N • Dpartment of National Defence • WCMRC • Canadian Coast Guard • Esquimalt and Songhees First Nations • Keystone Enviroment Y N Diesel Fuel Same Day [20,000 - 30,000] (Liters) Spud Barge Esquimalt Harbour, British Columbia, Canada Edge Reef, off of Athlone Island, Entrance of Northern BC Seaforth Channel, Canada Tug Boat 08-05-2016 Plumper Bay 13-10-2016 Bella Bella (Nathan E Stewart) • WCMRC • Marine T raffic • City of Vancouver • Clear Seas • Canadian Ship Source Oil Pollution Fund • T ECHNICAL REVIEW OF THE M/V MARATHASSA FUEL SPILL ENVIRONMENT AL IMPACT ASSESSMENT REPORT. Canadian Department of Fisheries and Oceasns • A Modeling Study on the Oil Spill of M/V Marathassa in Vancouver Harbour. Journal of Marine Science and Engineering 1,400 (liters) • Vancouver Volunteer Corps , 4000 personelle N N N Y Multiple (non specified) Multiple Multiple • 3 WCMRC Skimming vessels 18 (Days) 16 (Days) • Canadian Coast Guard • WCMRC • Emergency management of British Columia • Port Metro Vanocuver • City of Vancouver • Department of Fisheries adn Oceans • T ransport Canada • Enviromental and Climate CHange Canada Y N Bunker C Fuel 12 (hours) 81000 (t DWT) 2,700 (Liiters) Bulk Carriuer English Bay, Vancouver, British Columbia, Canada 05-04-2015 M / V Marathassa • WCMRC • Salvage operations 90 (Days) Y Y N N • WCMRC • Canadian Coast Guard • Gitga'at Nation • WCMRC Fisherman Oil Spill Emergency Team • Various wildlife advisors and shoreline assessment teams N Y Bunker C Fuel 67 (years) Army Ship 100 nautical miles south of Prince Rupert, BC, Canada 1905-07-05 M.G Zalinski Inlet Drive • WCMRC 210,000 (liters) Multiple (non-specified) 90 (Days) • WCMRC • Canadain Coast Gaurd • Island T ug and Barge Co. Y N Synthetic Bitumen Same Day 2434,000 (Liiters) Pipeline / Strom drains Baret Highway to Burrard Inlet, BC, Canada 24-07-2007 • WCMRC Multiple (non-specified) Multiple • WCMRC • Canadian Coast Guard Y N Black Oil Same Day 50,000 (Liters) Carrier Squamish, British Colubia, Canada 04-08-2006 Westwood Anette • WCMRC Multiple (non-specified) Multiple • WCMRC • Canadian Coast Guard 1 (Day) Vessel Gil Island, British Columbia Canada 21-03-2006 Queen of the North • Government of British Columbia Spill Incidents, Sunken Vessel Bligh Island • Arine Traffic 48,511.305 (kg) 60 (tonnes) Pumping of HFO and marine diesel • Salvage operations: hot tapping • High Speed Sweep System Sand Bags • RO Boom • GRS Boom • General Purpose Boom • Deflection Boom over 16,000 (feet) 5350 (feet) • T riton 20 Skimmer • Triton 60 Skimmer 2 • T ug (Atlantic Raven) • ETV AT L EAGLE – On Water Branch Director • CCGS Moorhen (RHIB) • CCG 750 • CCGER 735 221 (Days) 221 (Days) • BC Ministry of Enviroment and Climate Change Strategy (ENV) • Emergency Management BC • Canadian Coast Guard • Enviroment and Climaate Change Canada (ECCC) • Westerm Canada Marine Response Corpoeration N Y Fuel / Oil 18-17 hours Vessel Bligh Island, Bristish Columbia, Canada 04-12-2020 MV Schiedyk, Bligh Island (DGIR: 203178) Braer N Bioremediation (Y/N) Other Y Dispersent (Y/N) Reference Volume of recovered oil From absorbents From skimmers N/A N/A N/A 8,500 (tones) 4000 (tones oil) 60,000 (tones) Lightering 25,000 (tonnes) 11,000 (tonnes) • 500 Peronel at sea 200 (tonnes) 6,700 (tonnes) • Submereged Oil Spill Repsonse • Diver Puping Operations 21 (barrels) Diving Operations 2,610 (tonnes) 37,610 (tonnes) • 40, 000 workers • Numerous Volunters • 21,000 troops, public officials, residents and volunteers • Lightering Operations •Manual Onshore Recovery Y Y Y Y N/A Multiple, Non Specified Weight N/A Multiple, Non Specified Weight Multiple, Non Specified Weight N/A Multiple • More than 100 vessels • 15000 private fishing boaats ~ 245 (days) N N N 14-16 (m/s) NW 2-4 (meters) N Amazzone N N N N N/A N/A N/A N/A 5000 (meters) • Sirene 20 skimmers • Pollutank inflatable reservoirs 0 13,000 (tonnes) Nakhodka • Skimming Vessel 8-10 (m/s) NE, ENE 2 (meter) • Petroloes de Venuzuela • Maraven • Lagoven N Y Venezuelan crude oil 6 (hours) 75,000 (tonnes) 3,600 (tonnes) Tanker / Vessel Maracaibo Channel, Venezuela l 28-02-1997 Nissos Amorogos 8,700 (kL) • 1,000 - 1,500 (tonnes) Lightering • 6,600 (kl) onshore mechanical recovery 50,000 (tonnes) N/A • Oil Drums • 200,000 Personell 40,000 (tonnes) • 550 personelle • colelcting emulsions from sea using clam shell buckets • 2 Recovery Systems • Sieving, beach cleaning with • Tracked excavating machines used to reove machines and surf washing sucken oil techniques sued to remove tar balls form beaches • Using a casueway built to the incident to remove oil via vechile Y Y N Y N/A N/A N/A N/A Multiple (non specified) Multiple Over 80 vessels • Crane Barges • Several hundred Fishing Boats • ASUWA of the Fukui Oil Storage Co • dredger/oil recovery vessel 154 (Days) 54 (Days) 15.6 Y Y N 20 m/s NW 0 - 6 (meter) N • Japan Coast Guard • Maritime Disaster Prevention Centre • Japaneese Contractors • Maritime Saftey Agency • Japan Marine Diaster Provention Center • Pertoleum Association of Japan N Y Intermediate Feul Oil 2 (days) 19,000 (tonnes) 6,240 (tonnes) Oil Tanker (Double Hull) West side, Honshou, Japan 02-01-1997 42 • Comparison and Assesment of Waste • Comparison and Assesment of • Comparison and Assesment of Generated Durning Oil Spills. 2014 Waste Generated Durning Oil Spills. Waste Generated Durning Oil Spills. Internatioanl Oil Spill Conference. T. 2014 Internatioanl Oil Spill 2014 Internatioanl Oil Spill Wadsworth Conference. T. Wadsworth Conference. T. Wadsworth • Comparison and Assesment of • International Oil Polution Compensation • Sustainabledevelopment.un.org • NOAA Incident News Waste Generated Durning Oil Spills. Fund • Moller T.H. 1997. The Nakhodka • Cedre 2014 Internatioanl Oil Spill • P.Masciangoili, G. Febres, M.E. Viale-Rigo. Oil Spill Response - T he Techinical • K. Lee, T.Kim. Influence of Tidal Conference. T. Wadsworth Nisso Amorgos Oi lSpill Experience. Advisers P ersepctive Current, Wind, and Wave in Hebei • Cedre International Oil Spill COnference • Internation Symposium on Marine Spirit Oil Spill Modeling. Journal of Proceedings. (march 1 1999). Oil Spill Response Proceedings. July Marine Scindde and Engineering. • A Response Guide for Sunken Oil Mats 16-17 1997, Toyko. The Nippon DOI:10.3390/jmse8020069 (SOMs): Formation, Behavior, Detection and Foundation Recovery. NOAA N/A N/A • 54 975 (Barrels) Lightering • 20, 000 (tons) Mechanical Recovery 0 28,000 (tonnes) N/A N/A • Due to exteme spring weather conditions, containment and recovery was haulted. • Plan P olmer (Land) was activated. • Plastic sheets were used to cover • Manual and Mechanical Clean Up promenades, jetties, walls, and sand at the top of several beaches in the Cotes-du-Nord Department. • Mechanical Onshore Recovery • Oil sand Seperation Machines Over 800 Wokrers: • 3000 personelle moblaized on 200 • Volonteers cleanup sites • Firefighters • Soldiers Y Y N Y • Granual Mineral Sorbents (Ekoperl 33) Yes, Non specific Weight N/A N/A > 16 500 (meters) • Egmolap skimmer Multiple • French Naval Vessels • Skimmer Vessels via French Navy Partol Boats • Abeille Languedoc • Abeille Flandre • rafts • British Voyager Multiple ~ 20 (Days) 7 (Days) 0 N N N Force 12 Winds 10-12 (meter) N • French Authroites • French Navy Army • Cedre Y N crude oil, paraffinic 48 (hours) 32,000 (tonnes) 21,000 (tonnes) Oil tanker off the coast of Finistère, Western Brittany, France 30-01-1988 76 (Days) 512 (Days) 0 N N N Force 11 NW ~ 26 (Feet) N • French Authorities • French Navy Army • Polmar Plan • ITOPF under the International Oil Pollution Compensation (IOPC) Fund • Departments of Finistere and Cotes-du-Nord • Korean Navy • Korean Coast Guard • Korean Army • Korea Marine Pollution Response Corporation • Various National Resposne Agencies • Ministreis of Maritime and Y No. 6 Feul Oil 48 (hours) 190,580 (Barrels) 98, 955 (Barrels) Oil Tanker Breton Coast of Britinnay France N N Y Tanio 07 -03 - 1980 Y N Light Arabean Crude Oil N N N/A N/A N/A N/A 30 (km) • Vessel-Submerged Oil Recovery System 1 Year N N Y Y • Uniterd Stated Coast Guard • NOAA Y N 4 (Days) 260,000 (tonnes) 10,000 (tonnes) Oil Tanker Port of Incheon on the west coast of South Korea, south of Seoul 07-12-2007 Hebei Spirit • Comparison and Assesment of • Comparison and Assesment of • Comparison and Assesment of Waste Generated Durning Oil Spills. • Comparison and Assesment of • Comparison and Assesment of Waste Generated Durning Oil Spills. Waste Generated Durning Oil Spills. 2014 Internatioanl Oil Spill Waste Generated Durning Oil Spills. Waste Generated Durning Oil Spills. 2014 Internatioanl Oil Spill 2014 Internatioanl Oil Spill Conference. T. Wadsworth 2014 Internatioanl Oil Spill 2014 Internatioanl Oil Spill Conference. T. Wadsworth Conference. T. Wadsworth • Texaco UK Conference. T. Wadsworth Conference. T . Wadsworth • Cedre • Braer: The huge oil spill that • Guidelines on Waste Management • US Department of Interior. M/T • Cedre • Times Shetland survived. BBC durning a shoreline pollution Athos I Crude Oil Spill • ITOPF • Sea Prince Incident and Changes •The Braer Oil Spill: R. Perry. incident. Cedre, Arcopol, Marine • Cedre • J. Micheal Spills of Nonfloating of Response Scheme After Incident. International Oil Spill Conference Enviroment and Renewable • E.Levine, G. Ott.Roles of T he Oil: Evaluationof Response L. Bong-Gill. • ITOPF Energies, Atlantic Area translationsal Enviromental Unit: Success and the Techniquires. International Oil Spill • The effects of the M/V Sea Prince • The Barrier Oil Spill, 1993, Ch 36. Programme, Euoropean Union, Athos 1 spill response. International Conference Prceedings (2008). accident on maritime safety R. J. Law, C. F. Moffat Enviromental Agnency of Wales, Oil Spil lCOnference Proceedings. management in Korea. Dong-Oh Cho Pembrokeshire County Council. 900 (tons) a few tons 81,498 (tons) Pumpung operations Other 0 2,000 (tonnes) Quantity of generated liquid waste 2,000 (tonnes) 5,100 (tonnes) Capacity • 166,905 total personelle Y Y N Y 2000 (meters) N •Joint Resposne Center • Marine Pollution Control Unit • Oil Spill Response Limited • Tenches were dug parallel to sea to • Rock Barrier instead of the use of catche oil and water mixture. booms •Lightering Operations Y Y N N N/A N/A N/A N/A Quantity of generated solid waste offshore storage Number • Lightering of vessel • On shore mechanical recovery, with portable high pressure pumps N Multiple Type/brand of sorbent pad In-situ burning (Y/N) 239, 678 (kg) No. Sorbent pad Y N/A Type/brand of absorbent boom Vacum Trucks (Y/N) N/A No. absorbent boom 13, 766 (meters) No. containment boom N/A N/A • Trawl Skimmer • Screw Skimmer Type of skimmer 5 (Days) N/A • Over 500 Vessels 5 (months) 90 - 100 (mph) Hurricane force winds Multiple No. Skimmer Onshore 19 (Days) Y Offshore Y N Sunny (Y/N) Cloudy (Y/N) 40 (m /sec) Rainy (Y/N) 8 - 10 (feet) Typhoon force winds Wind velocity N • Marine Polution Control Unit • Shetland Island Council • Pusan Maratime Police Korean • National Maratime P olice Agency • Petroleum Association of Japan • East Asia Response Limited Y N N 13,000,000 (gallons) Venezuelan crude oil Crude Oil Oil Tanker 1000 (tonnes) • Gulfaks Curde Oil • Bunker C Y N Arabian Crude Oil Wave height Wave period Calm (Y/N) Sudden (Y/N) Number of responders Response Method No. response vessels Response time Ocean Temp. (C) Weather condition Ocean condition Response Agency Incident type Continouse (Y/N) 1 (Day) Time lag for starting response Oil Product 85,000 (T onnes) No. of cargo 73,000 (tones) Oil Tanker 1 (Day) 85,000 (tones) 5,000 (Tonnes) Delaware River, Philedelpha, U.S.A Milford Haven Passage, United Kingdom Athos 1 26-11-2004 15-02-1996 Sea Emperess 17 (hours) Oil Tanker • 84,7000 (tonnes) •1500 (tonnes) Oil Tanker Volume of spilt oil. Quendale Bay, Shetland Isles 5- 01 -1993 Vessel Type port of Yosu, South Korea Date of incident (dd-mm-yyyy) Location Sea P rince 23- 07-1995 Name of Incident Erika • Comparison and Assesment of Waste Generated Durning Oil Spills. 2014 Internatioanl Oil Spill Conference. T. Wadsworth • IMO/UNEP: Regional Information System – Operational Guidelines and Technical Documents, Mediterranean Oil Spill Waste Management Guidelines, REMP EC, May 2011 (RIS/D/12). • Guidelines on Waste Management durning a shoreline pollution incident. Cedre, Arcopol, Marine Enviroment and Renewable Energies, Atlantic Area translationsal Programme, Euoropean Union, Enviromental Agnency of Wales, Pembrokeshire County Council. 1,200 (tones) emulsifed 1,100 (tonnes) 200,000 (tonnes) • > 5000 personelle & volunteers N N N Y Mop Nets Multiple, Non specifed Multiple, Non specifed • French Atlantic Maritme Prefect • French Authroities • Spain Authroites • Total 20,000 (tons) Bay of Biscay, France 12-12-1999 • Comparison and Assesment of Waste Generated Durning Oil Spills. 2014 Internatioanl Oil Spill Conference. T. Wadsworth N/A N/A 250 (tonnes) Pumping Oeprations 6,800 (tonnes) booms and earth moving machinery 965 (tonnes) 10,750 (tonnes) * 220 responders Y N N Y Multiple, Non specifed Multiple, Non specifed N/A N/A • 15 German, Danish and Swedish. Vessels Y Y N SW N • Danish, Swedish and German Authorites • Danish Civil Defence Corps • EC task Force Y N Heavy Feul Oil 3 (Days) 30,000 (tonnes) 2,7000 (tonnes) Oil Tanker Baltic Sea Bteween German and Danish Waters 28-03-2001 Baltic Carrier Prestige 52,512 (tones) 115,619 (tones) onshore mechanical recovery 567 (m^3) 3980 (m ^3) • Surf Washing • Diving Operations • High P ressure Hot water washing SE to NE • SASEAR • Tragsa • Department of Atlantic Coast • Departmental Equipment Agency , Maritime Service and Maritime Affairs •REMPEC N Y Intermediate Feul Oil 150 10,000 to 15,000 (tons) South of Lebanon, Mediterrranean Sea 13-07-2006 JYEH (Lebanon) • 3512 (tonnes) IFO • 152 (tonnes) Marine Deseil Oil Pumping Operations • Pumping Operations N N N • IFO • Mairine Deseil Oil • 50 (tonnes) •530 (tonnes) Cargo Vessel North of Tregastel Cotes d'Armor, Fracne, Western Channel 18-01-2007 MSC Napoli • Comparison and Assesment of Waste Generated • IMO/UNEP: Regional Information System Durning Oil Spills. 2014 Internatioanl Oil Spill – Operational Guidelines and Technical Conference. T. Wadsworth Documents, Mediterranean Oil Spill Waste • Guidelines on Waste Management • IMO/UNEP: Regional Information System – Management Guidelines, REMPEC, May durning a shoreline pollution Operational Guidelines and Technical Documents, 2011 (RIS/D/12). incident. Cedre, Arcopol, Marine Mediterranean Oil Spill Waste Management • Guidelines on Waste Management durning Enviroment and Renewable Guidelines, REMPEC, May 2011 (RIS/D/12). a shoreline pollution incident. Cedre, Energies, Atlantic Area translationsal • Guidelines on Waste Ma nagement durning a Arcopol, Marine Enviroment and Programme, Euoropean Union, shoreline pollution incident. Cedre, Arcopol, Marine Renewable Energies, Atlantic Area Enviromental Agnency of Wales, Enviroment and Renewable Energies, Atlantic Area translationsal Programme, Euoropean Pembrokeshire County Council. translationsal Programme, Euoropean Union, Union, Enviromental Agnency of Wales, Enviromental Agnency of Wales, Pembrokeshire Pembrokeshire County Council. County Council. 50,000 (tonnes) 159,300 (tonnes) 1 (Day) 64,000 (tons) Cape Finisherre Galica, Spain 13-11-2002 CHAPTER 4 RESULTS and DISCUSSION 4.1 Hyperparameter Optimization In this section, the optimized values for the hyperparameters of Artificial Neural Networks, Support Vector Regression, and Random Forest models are presented. As previously stated, this study aims to develop an estimation model for three types of oily waste: liquid, solid, and total solid and liquid waste. Bayesian optimization has been employed to fine-tune the hyperparameters of the estimation models listed in Table 3.1 to minimize the models' RMSE index. The scripts for all estimation models, including the hyperparameter optimization, were implemented in MATLAB. Figure 4.1 illustrates the results of Bayesian optimization for ANN, SVR, and RF models. Additionally, Tables 4.1, 4.2, and 4.3 provide detailed reports on the outcome of the optimization process and the bestdetermined values for the hyperparameters of ANN, SVR, and RF waste estimation models, respectively. Table 4.1 Optimized hyperparameters of ANN waste estimation model Type of waste Best Hidden layer size Best Learning rate Solid Waste 8 0.04087 Liquid Waste 8 0.03352 Total Waste 9 0.03156 43 a b c Figure 4.1 Results of Bayesian optimization for AI-based optimization models: a) ANN, b) SVR, and c) RF Table 4.2 Optimized hyperparameters of SVR waste estimation model Type of waste Best Kernel Function Best Kernel scale Solid Waste RBF 1.127 Liquid Waste RBF 1.127 Total Waste RBF 1.127 44 Table 4.3 Optimized hyperparameters of RF waste estimation model Type of waste Number of trees Best max feature no. Best split no. Solid Waste 3270 6 5 Liquid Waste 112 3 13 Total Waste 259 1 6 4.2 Waste Estimation Model After optimizing the hyperparameters, the models were tested with the tuned settings, and the results of the estimation models are summarized in Table 4.4. It is observed that the error values decrease progressively from ANN to SVR and finally to RF. Notably, RF exhibits the lowest reported error indices (RMSE and RMAE) across all types of waste. Furthermore, RF demonstrates a higher correlation value (0.77) between the estimated and observed wastes compared to ANN and SVR, which average at 0.4 and 0.59, respectively. This superiority of RF can be attributed to its increased randomness and minimal risk of overfitting due to the large number of decision trees involved. Additionally, RF provides a more robust output estimation by ensuring low correlation among individual trees, achieved through diversification of the forest by limiting the number of input parameters for each tree. The RF-based model developed for oily waste estimation holds significant applicability to numerous oil spill incidents, as the database can be continuously updated, and the model will automatically adjust accordingly. This model is of immense value to oil spill response practitioners and waste disposal contractors by providing crucial insights for waste management planning. 45 Table 4.4 Evaluation of the AI-based Waste Estimation Models Model ANN SVR RF Type of Waste RMSE (tonne) RMAE (%) CC Solid 9.98 4.86 0.39 Liquid 9.98 4.88 0.39 Total 9.98 4.80 0.41 Solid 3.69 5.32 0.61 Liquid 3.69 6.09 0.55 Total 3.69 5.32 0.61 Solid 1.20 0.61 0.77 Liquid 1.22 0.66 0.77 Total 1.25 0.61 0.77 4.3 Waste Allocation Framework The Bella Bella oil spill incident in British Columbia, Canada, serves as a case study to assess the proposed framework outlined in the methodology section. To this end, comprehensive details regarding treatment and receiving facilities, including their capacities and distances from one another, have been compiled and presented in Table 4.5. This data was acquired through numerous meetings with representatives and operation managers of Terrapure, the primary waste contractor on the West Coast. The geographical locations of each facility are indicated on the map depicted in Figure 4.2. 46 Figure 4.2 The location of waste handling facilities in British Columbia Table 4.5 Name and location of currently operating facilities in British Columbia Treatment/Recycle facilities Receiving facilities (Transfer stations) Safety Clean: Delta Aveitas: Maple Ridge Secure Energy: Richmond GFL: Victoria, Nanaimo, Port Alberni, Surrey, Sumas Environment: Burnaby Cranbrook, Kelowna, Prince George Stericycle: Surrey Safety Clean: Delta Secure Energy: Richmond, Fort St. John Sumas Env: Burnaby, Kamloops Stericycle: Surrey As depicted in Figure 4.2, most facilities are situated in the lower mainland area or on Vancouver Island, underscoring the importance of pre-established frameworks for waste 47 transport in the event of emergency incidents occurring in the central or northern regions of the province. Through research and discussions with the primary waste contractor for the province, it has been determined that the treatment path for generated oily waste varies depending on whether it contains a high concentration of BTEX or bunker oil, as illustrated in Figure 4.3. Liquid high BTEX oily waste is primarily transported to oil refineries in Alberta via pipelines. At the same time, its solid form can undergo processing within the province, being stabilized at a soil processing facility in Princeton before being transported to designated landfills. In cases where the generated solid waste stems from a bunker oil spill incident, it is typically transported directly to landfills without segregation. Conversely, the liquid waste undergoes treatment at specialized facilities, where processes such as dewatering, emulsion breaking, and physical separation are employed to separate oil from water. In many instances, the separated oil can subsequently be sold. Liquid Through pipeline to Alberta Oil refinery BTEX Solid Liquid Processing facilities Bunker oil/BTEX free Solid EnviroGreen, Princeton, BC Soil processing facility Landfill (Silverberry/Great valley) Dewatered (heated tanks) Chemical (emulsion breaking) Physical separation Usually as is (rarely segregated or recycled) Landfill (Silverberry/Great valley) Figure 4.3 The common practice of oily waste management in British Columbia 48 Wastewater treatment Reselling the oil Considering the heterogeneous nature of oily wastes, it is imperative to consider waste-waste compatibility. Furthermore, the transportation of each waste type necessitates specific types of trucks and treatment facilities. To address these requirements, comprehensive details regarding all operational facilities have been compiled and are presented in Table 4.6. Table 4.6 The detailed information on oily waste handling facilities in British Columbia ID Name Type City S1 Bella Bella Oil Spill Incident Bella Bella T1 GFL Victoria T2 GFL Surrey T3 Safety Clean Delta Treatment Facility T4 Secure Energy Richmond Sumas T5 Burnaby Environmental T6 Stericycle R1 Aveitas Surrey Maple Ridge R2 Safety Clean Delta R3 Secure Energy Richmond R4 Secure Energy Fort St. John Sumas Receiving Facility R5 Burnaby Environmental Sumas R6 Kamloops Environmental R7 Stericycle Surrey 49 R8 GFL Victoria R9 GFL Nanaimo R10 GFL Port Alberney R11 GFL Surrey R12 GFL Cranbrook R13 GFL Kelowna R14 GFL Prince George L1 Silverberry L2 Great Valley Landfill Fort St. John As outlined in the methodology section, the optimization of decision variables surpasses the cognitive capacity of human processing, classifying these issues as NP-hard problems that are nearly insurmountable without computational aid. GA emerges as a robust solution for optimizing decision variables involved in waste transfer processes, effectively integrating the capacities of diverse handling facilities into the model without necessitating intricate programming. This accelerates processing time and enhances the model's practicality, which is a critical consideration in emergency scenarios. The real-world application of this methodology is exemplified through the analysis of the Bella Bella oil spill incident in British Columbia, where estimated solid waste volume and associated cleanup costs serve as pivotal parameters. However, detailed information regarding the incident volume is private. Therefore, based on the most recent study published on the estimated solid waste in the Bella Bella oil spill incident, the volume of generated solid waste is estimated to be 13.8 × 103 tonnes. According to the published articles, the cleanup costs of the Bella Bella oil spill incident amount to around $2.7 million. The cost breakdown is reported as follows: 50 • Cleanup costs: $2.7 million • Fines for non-compliance: $500,000 • Environmental testing during spill: $50,000 • Labor expenses: $500,000 • Office equipment, boats, etc.: $100,000 • Hidden costs: $100,000 • Transportation and disposal expenses: $1.45 million • Disposal fee (approx.): $100/tonne of solid waste • Total transportation cost: Approximately $1 million After running the model, the transportation cost was minimized, as shown in Table 4.7. The optimization process took approximately 30 minutes, resulting in a total minimized cost of 662,600 CAD. Figure 4.4 illustrates the trial-and-error process of the algorithm to achieve the lowest transportation cost for waste disposal from the source to the landfill. It is evident from the analysis that utilizing the model would yield savings of 337,400 CAD. The model's high level of customization is evident in its ability to adjust capacities for different facilities, preventing traffic congestion in smaller plants. Furthermore, the model accommodates diverse treatment rates and relevant costs across facilities, contributing to a more cost-effective waste allocation strategy. The model's adaptability is a key highlight, enabling rapid adjustments in response to emergencies, such as the malfunction of a treatment plant, ensuring an optimized flow of waste transportation. 51 Table 4.7 The identified path from source to landfill with volume based on the Bella Bella oil spill incident in British Columbia ID Path Volume (tonne) Description 26 Zs1l1 7.28 Volume of waste from source 1 to landfill 1 58 Zs1r1 6.4 Volume of the waste from source 1 to receiving facility 1 59 Zr1l1 6.4 Volume of waste from receiving facility 1 to landfill 1 Figure 4.4 Genetic Algorithm minimization cost graph The model's capacity to consider multiple generation nodes reflects its applicability to real-world scenarios where waste transfer involves intricate networks. The reported daily transfer of 60 tonnes of non-marine oily waste in British Columbia underscores the 52 significance of such a comprehensive approach. Moving beyond immediate applications, the study offers valuable insights for decision-makers regarding future facility expansions and strategic placements. It serves as an evaluative tool, providing feedback on the current performance of facilities and proposing a holistic optimization of the waste management system. The model's versatility extends to different geographical areas, making it a valuable tool for calculating distances based on the coordinates of new facilities or potential oil spill sources. 53 CHAPTER 5 CONCLUSION AND RECOMMENDATIONS 5.1 Conclusion In conclusion, this research represents a significant step forward in the field of Oily waste management, offering an in-depth exploration of estimation models and allocation frameworks. The study systematically examined three prominent models – Artificial Neural Networks (ANN), Support Vector Regression (SVR), and Random Forest (RF) – and found that RF stands out as the most adept in accurately predicting volumes of oily waste. Incorporating Bayesian optimization in hyperparameter tuning enhanced the precision and adaptability of the estimation model, rendering it a potent tool for practical applications in diverse scenarios. Utilizing the Genetic Algorithm (GA), the waste allocation framework emerged as a robust solution to the intricate challenges of transporting varied types of oily waste to treatment and disposal facilities. Its flexibility enables decision-makers to tailor the framework to specific needs, accounting for facility capacities, waste compatibility, and realtime conditions. This research showcased its practical efficacy by applying the framework to the Bella Bella oil spill incident in British Columbia, highlighting its pivotal role in facilitating informed decision-making during emergencies. Beyond its theoretical contributions, this research holds practical implications for waste management practitioners, environmental authorities, and response teams. The rigorous evaluation and validation of the proposed models and frameworks establish a solid foundation for enhancing waste management practices in offshore environments. As the offshore industry evolves, integrating artificial intelligence and optimization techniques becomes imperative for developing sustainable and efficient waste management strategies. 54 The models and frameworks introduced in this study address current challenges and set the stage for future innovations in the field, emphasizing the continual need for advancements in off-shore waste management. The success of the waste allocation framework lies in its adaptability, providing decision-makers with the ability to customize the model based on facility capacities, waste compatibility, and real-world constraints. Furthermore, applying the Genetic Algorithm enables quick adjustments to the network in response to emergencies, ensuring that the optimized flow of waste transportation can be identified promptly. The model's capacity to handle multiple generation nodes and provide insights into facility performance and potential expansions further strengthens its utility in practical waste management scenarios. In summary, this research underscores the pivotal role of advanced modelling and optimization techniques in addressing the complexities of off-shore oily waste management. The presented models and frameworks advance our understanding of waste estimation and allocation and offer tangible tools for improving decision-making in the face of environmental emergencies. As the industry progresses, integrating these innovations will be essential in shaping sustainable practices and minimizing the ecological impact of off-shore activities. The summary of advantages and disadvantages of the incorporated models in this study are presented as follows: Strengths: • Precision in Waste Estimation: Bayesian optimization enhances the precision of waste estimation models, ensuring accurate predictions. • Optimized Waste Allocation: Genetic Algorithm-based waste allocation framework optimizes transportation routes and minimizes the costs. 55 • Model Adaptability: The waste allocation framework is highly customizable, allowing adjustments based on real-time conditions and facility capacities. • Practical Efficacy: Application of the framework in the Bella Bella oil spill incident demonstrates its practical efficacy in real-world emergencies. • Flexibility for Decision-Makers: Decision-makers can tailor the waste allocation framework to specific needs, considering waste compatibility and facility capacities. • Quick Network Modifications: Genetic Algorithm enables adjustments to the waste transportation network in response to emergencies or facility malfunctions. • Handling Multiple Generation Nodes: The model can consider multiple generation nodes, addressing the complexity of waste transfers to various facilities. • Insights for Facility Expansion: Provides insights into the performance of existing facilities, aiding in strategic decisions for future expansions. • Feedback Mechanism: The model acts as a feedback mechanism for the performance of present facilities, allowing for continuous improvement. • Applicability in Different Areas: The model's ability to calculate distances by entering geographical coordinates makes it applicable in diverse geographical locations. Weaknesses: • Limited Public Data: Lack of publicly available data for actual incidents, such as the Bella Bella oil spill incident, limits the accuracy of model validation. • Complexity of NP-Hard Problems: Waste allocation problems are NP-hard, making manual handling impractical due to the extensive number of decision variables. • Dependency on Initial Solutions (GA): The performance of the Genetic Algorithm is highly dependent on the randomness of the initial generated solutions. 56 • Incomplete Information for Bella Bella Oil Spill Incident: Limited information on the volume of the Bella Bella oil spill incident waste hinders precise modelling and validation. • Challenges in Modeling Waste-Waste Compatibility: Incorporating waste-waste compatibility in the models is challenging due to the heterogeneous nature of oily wastes. • Specificity of Truck and Facility Requirements: Shipping each type of waste requires specific trucks and treatment facilities, adding complexity to the waste allocation problem. • Continuous Model Improvement: The model's performance depends on frequent updates and adjustments based on evolving waste management practices. • Resource-Intensive Bayesian Optimization: Bayesian optimization, while effective, can be resource-intensive, requiring significant computational power. • Inherent Uncertainties in Oil Spill Incidents: The unpredictable nature of oil spill incidents introduces uncertainties that may affect the accuracy of waste estimation. • Need for Real-Time Data Integration: Real-time data integration is crucial for the models, and its absence may impact the adaptability and accuracy of waste allocation. 5.2 Recommendations Given the strengths and weaknesses outlined in the previous section, broadening the study by integrating various approaches or components will contribute to a more thorough evaluation of off-shore oily waste estimation. The suggested viewpoints are summarized below: 57 • Integration of Real-Time Data: Developing methods for real-time data integration to enhance the adaptability and accuracy of waste management models. • Enhanced Validation Processes: Implementing comprehensive validation processes, including creating realistic scenarios, to overcome data limitations for incidentspecific models. • Exploration of Advanced Optimization Algorithms: Investigating the application of advanced optimization algorithms beyond Genetic Algorithms for waste allocation to improve efficiency. • Incorporation of Waste-Waste Compatibility Models: Developing models that explicitly consider waste-waste compatibility, addressing the challenges posed by the heterogeneous nature of oily wastes. • Dynamic Facility Capacity Updates: Implementing mechanisms for dynamic updates of facility capacities, allowing for real-time adjustments in waste allocation frameworks. • Human-Computer Interaction: Conducting studies on human-computer interaction to enhance the user-friendliness of waste allocation frameworks for decision-makers. • Integration of Machine Learning for Incident Volume Estimation: Investigating the use of machine learning techniques for more accurate estimation of incident volumes, considering the complexities of real-world incidents. • Environmental Impact Assessment: This includes an environmental impact assessment component in future studies to evaluate the sustainability of waste management practices proposed by optimization models. 58 REFERENCES Aguilera F, Méndez J, Pásaro E, Laffon B (2010) Review on the effects of exposure to spilled oils on human health. Journal of Applied Toxicology, 30(4), 291-301. doi: 10.1002/jat.1521 Alumur S, Kara B (2007) A new model for the hazardous waste location-routing problem. Computers & Operations Research, 34(5), 1406-1423. Azevedo A, Oliveira A, Fortunato AB, et al. (2014) A cross-scale numerical modelling system for management support of oil spill accidents. Mar Pollut Bull 80:132–147. doi: 10.1016/j.marpolbul.2014.01.028 Barron MG (2012) Ecological Impacts of the Deepwater Horizon Oil Spill: Implications for Immunotoxicity. Toxicol Pathol 40:315–320. doi: 10.1177/0192623311428474 Beegle-Krause CJ (2005) General NOAA oil modeling environment (GNOME): A new spill trajectory model. In: 2005 International Oil Spill Conference, IOSC 2005. Allen Press, pp 3277–3283 Bejarano AC, Mearns AJ (2015) Improving environmental assessments by integrating Species Sensitivity Distributions into environmental modelling: Examples with two hypothetical oil spills. Mar Pollut Bull 93:172–182. doi: 10.1016/j.marpolbul.2015.01.022 Bergstra J (2012) Random search for hyper-parameter optimization. jmlr.org 13:281–305 Binet S, Leszkowicz A, Brandt H, et al. (2003) Bitumen fumes: Review of work on the potential risk to workers and the present knowledge on its origin. The Science of The Total Environment 300(1-3):37-49. DOI:10.1016/S0048-9697(02)00279-6 59 Beyer J, Trannum HC, Bakke T, et al. (2016) Environmental effects of the Deepwater Horizon oil spill: A review. Mar. Pollut. Bull. 110:28–51 Breiman L (2001) Random forests. Mach Learn 45:5–32. doi: 10.1023/A:1010933404324 Chen Z, An C, Chen X, et al. (2021) Inexact inventory-theory-based optimization of oily waste management system in shoreline spill response. Sci Total Environ 777:146078. doi: 10.1016/j.scitotenv.2021.146078 Cutler DR, Edwards TC, Beard KH, et al. (2007) Random forests for classification in ecology. Ecology 88:2783–2792. doi: 10.1890/07-0539.1 Dave D, Ghaly AE (2011) Remediation technologies for marine oil spills: A critical review and comparative analysis. Am. J. Environ. Sci. 7:424–440 Diamantidis NA, Karlis D, Giakoumakis EA (2000) Unsupervised stratification of crossvalidation for accuracy estimation. Artif Intell 116:1–16. doi: 10.1016/S00043702(99)00094-6 Fatehi M, Asadi HH (2017) Data integration modeling applied to drill hole planning through semi-supervised learning: A case study from the Dalli Cu-Au porphyry deposit in the central Iran. J African Earth Sci 128:147–160. doi: 10.1016/J.JAFREARSCI.2016.09.007 Fingas M (2016) Oil Spill Science and Technology: Second Edition. Gulf Proffesional Publishing Ghanbari F, Kamalan H, Sarraf A (2021) An evolutionary machine learning approach for municipal solid waste generation estimation utilizing socioeconomic components. Arab J Geosci 14:1–16. doi: 10.1007/s12517-020-06348-w 60 Government of Canada (2022) Marine Spills Contingency Plan https://www.ccggcc.gc.ca/publications/environmental-environnementale/marine-pollutiondeversements-en-mer/plan-overview-eng.html. Accessed 10 Oct 2022 Goyal MK, Bharti B, Quilty J, et al. (2014) Modeling of daily pan evaporation in subtropical climates using ANN, LS-SVR, Fuzzy Logic, and ANFIS. Expert Syst Appl 41:5267–5276. doi: 10.1016/j.eswa.2014.02.047 Guo W (2017) Development of a statistical oil spill model for risk assessment. Environ Pollut 230:945–953. doi: 10.1016/j.envpol.2017.07.051 Guo WJ, Wang YX (2009) A numerical oil spill model based on a hybrid method. Mar Pollut Bull 58:726–734. doi: 10.1016/j.marpolbul.2008.12.015 Hoang DT (2008) Metaheuristics for Np-Hard Combinatorial Optimization Problems. System IMO (2001) Guidelines for the Development of Shipboard Marine Pollution Emergency Plans. International Maritime Organization. IPIECA, IOGP (2014) Oily Waste Management for Offshore Petroleum Operations. International Association of Oil & Gas Producers. ITOPF (2021) Statistics - ITOPF. In: Int. Tanker Owners Pollut. Fed. https://www.itopf.org/knowledge-resources/data-statistics/statistics/. Accessed 2 Oct 2022 Kourou K, Exarchos TP, Exarchos KP, et al. (2015) Machine learning applications in cancer prognosis and prediction. Comput. Struct. Biotechnol. J. 13:8–17 61 Lehr W, Wesley D, Simecek-Beatty D, et al. (2000) Algorithm and interface modifications of the NOAA oil spill behavior model. In: Environment Canada Arctic and Marine Oil Spill Program Technical Seminar (AMOP) Proceedings. pp 525–539 Li P, Chen B, Li Z, Jing L (2014) An agent-based simulation-optimization coupling approach for device allocation and operation control in response to offshore oil spills. In: International Conference on Marine and Freshwater Environments, iMFE 2014 Lin HY, Hsu PY, Yeh YT (2006) Application of the AHP in data warehouse system selection decisions for SMEs in Taiwan. Int J Manag Enterp Dev 3:599–617. doi: 10.1504/IJMED.2006.010356 Lottes P, Stachniss C (2017) Semi-supervised online visual crop and weed classification in precision farming exploiting plant arrangement. In: IEEE International Conference on Intelligent Robots and Systems. Institute of Electrical and Electronics Engineers Inc., pp 5155–5161 Major DN, Wang H (2012) How public health impact is addressed: a retrospective view on three different oil spills. Toxicol Environ Chem. 2012;94:442–467. doi: 10.1080/02772248.2012.654633. Marta-Almeida M, Ruiz-Villarreal M, Pereira J, et al. (2013) Efficient tools for marine operational forecast and oil spill tracking. Mar Pollut Bull 71:139–151. doi: 10.1016/j.marpolbul.2013.03.022 Metcalf S (2014) Developing Location-based Oil Spill Waste Management Plans. Int Oil Spill Conf Proc 2014:1633–1646. doi: 10.7901/2169-3358-2014.1.1633 62 Mohammadiun S, Hu G, Alavi Gharahbagh A, et al. (2021) Optimization of integrated fuzzy decision tree and regression models for selection of oil spill response method in the Arctic. Knowledge-Based Syst 213:106676. doi: 10.1016/j.knosys.2020.106676 Peterson CH, Rice SD, Short JW, et al. (2003) Long-Term Ecosystem Response to the Exxon Valdez Oil Spill. Science (80) 302:2082–2086 Sadri S, Research DB-WR (2012) Nonparametric methods for drought severity estimation at ungauged sites. Wiley Online Libr 48:12505. doi: 10.1029/2011WR011323 Sain SR, Vapnik VN (1996) The Nature of Statistical Learning Theory. Technometrics 38:409. doi: 10.2307/1271324 Sameen MI, Pradhan B, Lee S (2019) Self-Learning Random Forests Model for Mapping Groundwater Yield in Data-Scarce Areas. Nat Resour Res 28:757–775. doi: 10.1007/S11053-018-9416-1/FIGURES/10 Smola AJ, Schölkopf B (2004) A tutorial on support vector regression. Stat Comput 14:199–222. doi: 10.1023/B:STCO.0000035301.49549.88 Snoek J, Larochelle H, Adams RP (2012) Practical Bayesian optimization of machine learning algorithms. In: Advances in Neural Information Processing Systems. pp 2951–2959 Thissen U, Van Brakel R, De Weijer AP, et al. (2003) Using support vector machines for time series prediction. Chemom Intell Lab Syst 69:35–49. doi: 10.1016/S01697439(03)00111-4 Tkalich P (2006) A CFD solution of oil spill problems. In: Environmental Modelling and Software. Pergamon, pp 271–282 63 Svetnik V, Liaw A , Tong C, et al. (2003) Random Forest: A Classification and Regression Tool for Compound Classification and QSAR Modeling. J Chem Inf Comput Sci 43:1947–1958. doi: 10.1021/CI034160G Wang WC, Chau KW, Cheng CT, Qiu L (2009) A comparison of performance of several artificial intelligence methods for forecasting monthly discharge time series. J Hydrol 374:294–306. doi: 10.1016/J.JHYDROL.2009.06.019 Xu L, Li J, Brenning A (2014) A comparative study of different classification techniques for marine oil spill identification using RADARSAT-1 imagery. Remote Sens Environ 141:14–23. doi: 10.1016/j.rse.2013.10.012 Zhang Z, Friedrich K (2003) Artificial neural networks applied to polymer composites: A review. Compos Sci Technol 63:2029–2044. doi: 10.1016/S0266-3538(03)00106-4 Zhao M, Chow TWS, Tang P, et al. (2017) Route selection for cabling considering cost minimization and earthquake survivability via a semi-supervised probabilistic model. IEEE Trans Ind Informatics 13:502–511. doi: 10.1109/TII.2016.2593664 64