The coastal waters of Canada embrace a wide range of physical environments and ecosystems from the warm, sediment-rich waters of the Bay of Fundy to the nutrient-limited cold waters of the high Arctic. This range of biophysical characteristics impacts natural attenuation and weathering processes for oil stranded on shorelines. This study was conducted to: 1) identify and quantify the primary regional parameters that control shoreline oil translocation (removal) processes and pathways and 2) define the effectiveness and environmental consequences of current and potential oiled shoreline treatment strategies and tactics. A specific knowledge gap, here and elsewhere in the world, has been in understanding how the distribution and character of fine-grained sediments affect stranded oil attenuation. Fine-grained sediments (<1mm) can play a critical role in natural or induced (that is, shoreline treatment) oil dispersal. Shoreline sediment samples were collected and analyzed from representative locations on Arctic, Atlantic, and Pacific Ocean beaches to provide a broad geographic characterization of mineral fines at the regional level. This knowledge is the basis for an “Oiled Shoreline Response Program (SRP) Decision Support Tool” to aid spill scientists, students, environmental resource managers, spill responders and the public in understanding the response methods and the ramifications and consequences of their shoreline treatment options without the need to digest technical papers, large reports, or data bases. This MPRI SRP Decision Support Tool is intended to be a dynamic, interactive, multi-layered, geographically and seasonally-based model for shoreline oil spill response decision analyses. A goal of this interactive model is to move away from the traditional static format of learning from explanations in text reports and publications to an interactive tool that encourages its users to explore and fully understand the significance of the different environmental factors outlined in publications and data bases. Recent advances in web technology make this possible. The development of user interface platforms such as React, libraries such as D3, and notebook forms like Observable has created a palette of technologies that together make web application patterns such as Documodels a much more streamlined development process. The power of this medium is to convey a complex subject and to enable a user to grasp keen insights and so understand the consequences of intervention decisions.
The Multi-Partner Research Initiative (MPRI) under Canada's Oceans Protection Plan (OPP) was created to establish an integrated international research program to advance oil spill research in Canada to enhance Canada's level of preparedness and response capability. The focus of the MPRI is to build a research network that brings together the best scientific expertise in oil spill research across Canada and internationally. A key objective for the MPRI is to advance scientific knowledge that addresses major gaps in oil spill response and remediation strategies and support the development, validation, and regulatory approval of Alternative Response Measures (ARMs) on the operational scale.
A specific knowledge gap, in Canada and elsewhere in the world, has been in the understanding of stranded oil attenuation with respect to fine-grained sediment interaction and the potentially critical role of this process in natural or induced (that is, shoreline treatment) oil dispersal. Oil spilled at sea and on coasts enters dynamic environments in which the oil moves rapidly between these very different physical- and eco-systems. The three primary natural oil translocation pathways in the marine environment following an oil spill are: oil to shorelines, shorelines to nearshore waters, surface waters to the marine water column. Understanding the pathways and processes by which oil is translocated off a shoreline is critical to a Shoreline Response Program (SRP) (IPIECA-IOPG, 2020; Owens and Santner, 2020) in deciding whether intervention is appropriate to promote natural attenuation, reduce environmental risk, and accelerate recovery of the environment.
A goal of the MPRI Oil Translocation theme project is to provide the scientific basis and tools for considering oil-sediment interaction as a viable ARM. Providing an SRP Decision Support Tool that transfers the state-of-knowledge and conceptual oil translocation models into a computer-based interactive application will enable scientists, students, environmental resource managers, spill responders, and the public to better understand the effects and consequences of the options that exist to accelerate the weathering of spilled oil, and therefore environmental recovery, following an oil spill that could affect shorelines.
The basis for a shoreline oil attenuation model is to first understand the shoreline character and coastal process of the study area. This project focuses on the coastal waters of Canada, which embrace a wide range of physical environments and ecosystems from the warm, sediment-rich waters of the Bay of Fundy to the nutrient-limited cold waters of the high Arctic. This range of biophysical characteristics impacts natural attenuation and weathering pathways and processes for oil stranded on shorelines. In order to build this foundation for the conceptual modeling, and to support the development of the SRP Decision Tool, a data base of the geographic variability of the relevant coastal process parameters was created using a geographic information system (GIS) approach at distinct spatial and temporal scales. Data available from multiple sources (online and in published reports) are compiled for many of the range of parameters identified later in this discussion (Figure 3). Additionally, shoreline sediment samples were collected in 2018 and 2019 from representative locations on Arctic, Atlantic, and Pacific Ocean beaches and analyzed to provide a broad geographic characterization of mineral fines at the regional level.
A list of parameters for inclusion in the GIS database was developed that included:
Physical environment and processes (by month, season or synoptic): waves, tide type and range, shore ice, temperature, winds, river discharge, daylight
Physical coastal character: substrate materials, backshore slope, coastal typology (Dürr et al. 2011).
This GIS database was designed to store data and to be accessed at three scales so that the characterization of coastal regions can be accomplished at distinct spatial scales, depending on the variables used for regionalization (e.g., Laurelle et al. 2013; Lemmen et al. 2016) and the level of detail that is required by the user.
First Order Scale - Three (3) Ocean Regions
Four themes were reviewed at a macro-scale to provide characteristics for three first-order regions of the Canadian coastal setting (Table 1). Necessarily very generalized, this overview scale nonetheless provides a basis for understanding fundamental or strategic differences between the three ocean coasts of Canada.
Second Order Scale – Twenty-Six (26) Geographic Coastal SubRegions
At the next level of detail, a second order coastal scale includes more of the topics that define coastal environments, including geological character, backshore relief, beach character, fetch and wave exposure, mean tidal range, and sediment availability (Figure 1). An analysis of these traits yields twenty-six subregions: thirteen (13) on the Arctic Ocean coasts, seven (7) on the Atlantic Ocean, and six (6) on the Pacific Ocean (Owens 1994).
Third Order Scale – Climate Atlas of Canada
The third order of detail is based on the Climate Atlas of Canada climate model (version 2 BSSAQv2 (accessed July 10, 2019). The Climate Atlas is a user-friendly, interactive cell-based system that was developed to provide public access to information on climate change and its national, regional and, local impacts. On screen, the user can select to display data by Province, on a “large grid” scale, which is on the order of 100-km size (Figure 2, left) or on a “small grid”, which is on the order of 25–30 km in size. This system is similar to the interactive documodel described below as clicking on a grid presents a side-bar box (Figure 2, right) that presents specific data and infromation for that grid on the selected topic as well as sliders that can alter the time period for the data that is graphically displayed in the box.
Some of the GIS-based data sets, such as the NRCAN CanCoast 2.0 data on mean wave height and geological character (Manson et al. 2019), are directly transferrable to the coastal grid cells for all of Canada. Other data may be provided for individual grid cells; for example, tidal range data have to be managed for each individual cell as this is location-based and not a GIS-based data set. In addition, this data management may require some generalization as not every cell has predicted or historical tide data.
CONCEPTUAL MPRI SIMULATION MODEL FOR SRP DECISION SUPPORT
The development of a decision tool or model requires understanding and estimating the natural fate and behavior of oil on shorelines and the consequences of the broad range of shoreline treatment options. This topic involves consideration of a wide variety of physical, chemical, and biological parameters and processes (Figure 3) as well as translocation pathways from a shoreline and shoreline response technologies.
The problem of integrating these numerous interrelated but quite diverse inputs was addressed during an expert-group workshop. The objective of the workshop was to create a plan for the development of a computer-based tool for decision-making during oil spill incidents directed to oiled shoreline treatment. The workshop was facilitated by a model designer, who had no prior experience or understanding of oil spills (GD), with the goal of unpacking, simplifying, and organizing solutions around this complex issue. The outcome is a model that is presented in a 5-step overview (Figure 4 and Table 2) and for which the conceptual structure of the decision support tool is summarized in Figure 5.
Although this concept is in development, we anticipate that the form of the final tool will emerge as a Documodel which, as the name suggests, is a cross between a document and a model (Observable 2018). As an example, a news article or technical paper can include interactive elements such as sliders and checkboxes that control the information that the viewer sees; this action literally changes parameters of the model underneath the prose and generates a new view (see for example, Pearce 2081, Victor 2011). This type of interactive document can tell a much richer story to the reader than a static report. The information is layered so that greater levels of detail or explanations can be accessed as the user wishes. We believe that this capability will provide users with access to the multiple layers of science-based information and data that support the elements of the decision tool.
The model input is planned as a user-friendly visual interface that requires the user to select the key elements of a scenario; whether for planning or in real time. The modeling process will involve an audit trail to show the assumptions involved and how the model output was created. As the first step in the model, the user defines the scenario by selecting input parameters for five topics: (1) date of the spill; (2) oil type, by categories relevant to the Canadian marine environment; (3) shoreline character/substrate material; (4) oil loading; and (5) geographic location. Two additional pre-loaded input components are the “Canadian Coastal Geographic Variability GIS” database described above and the “Oil Spill Historical Database” created by Concordia University.
Intermediate Calculations and the “Scratchpad”
A series of intermediate calculations follow using this input selection:
1) met-ocean conditions are derived from the Canadian Coastal Geographic Variability GIS database for the location and time (season), with a default of the current date at the time the model is activated;
2) the oil is “aged” from the time of the spill to the time it reaches shore using the met-ocean conditions and the US National Oceanographic and Atmospheric Administration (NOAA) Automated Data Inquiry for Oil Spills (ADIOS) model; and
3) penetration, retention and persistence of the weathered oil, both surface and subsurface, are developed based on existing data (this data will be updated with the results from the OT Theme laboratory and meso-scale tests).
All of the inputs and the intermediate calculations are stored in the memory (the “Scratchpad”).
Oil Translocation Simulator
The Natural Translocation Simulator is a combination of intuitive models, or insight, and engineering models that provide the natural attenuation baseline for the scenario defined by the user-selected input parameters. Some of these intuitive models may be based on probabilistic pathways and at this stage a level of confidence may be built in based on Monte Carlo simulations. Treatment logic, which includes a set of disqualification (or viability) rules, is derived from the ECCC marine shoreline response field guide (ECCC 2016).
Separately, the “Similarity Chooser” searches keywords in the Oil Spill Historical Database to find similar real-world situations that correspond to the selected scenario; this linked information constitutes part of the Documodel output.
The Treatment Options Matrix in the Documodel uses the natural attenuation baseline generated by the Natural Translocation Simulator and is populated from the Treatment Logic (or viability) output. This treatment matrix is the key output to display the initial input parameters and the consequences and ramifications of changing one or more of these parameters. This matrix is perhaps the most challenging element as this will be designed to be easily understandable and compelling in the sense that the user would be encouraged to seek further information. The data or knowledge that underpins the output will be revealed by clicking on an object and, as most of this will be text or tabular, a visualization or animation process will provide a story-book approach as a better way to understand the basis for the science-based decision process.
Examples of Supporting Materials
Two diagrams (Figures 6 and 7) illustrate, albeit in a static format, the character and nature of some of the SRP Decision Support Tool conceptual simulations, example outputs, and user query interface.
One element of the natural translocation simulator is the “Oil Attenuation” interactive graphic (Figure 6) in which each of the two axes, “Percent Oil Stranded Onshore” and “Time”, can be changed by moving a slider on the screen. The y axis reflects different oil concentration targets to achieve a No Further Treatment (NFT) objective. As the horizontal NFT line is moved down and the treatment criteria become stricter, the amount of oil remaining on the shoreline after completion of treatment is reduced and may approach 0% in the case of zero tolerance criteria. The vertical line at right on the Time (Δ t) axis refers to the simulated natural attenuation to approximately 0% oiling (envisioned as days, weeks, months, years, decades). The vertical line at left on this axis is the desired accelerated attenuation time line that would result from intervention and treatment activities. The nearer this vertical line is to the y axis then the shorter is the length of time that shore zone resources would be at risk. The resulting delta is the target reduction time in exposure to the oil on the shoreline that would be addressed by the treatment decision. The solid curve represents natural attenuation through time and the dashed curve the reduction in oil as a result of an intervention or treatment option. Figure 7 from the demonstration version depicts the curves associated with a summer scenario in northern British Columbia with two treatment options (low pressure ambient washing and physical mechanical removal) for a Moderate Surface Oil Distribution of unweathered heavy oil in the upper intertidal zone of a pebble/cobble beach. In this example, the consequences of the choice of treatment are indicative of the lengths of time before the NFT criteria are achieved.
The interactive graphic Treatment Tradeoff Matrix (Figure 8) allows the user to choose between appropriate treatment options. The consequences of this selection are presented in terms of the associated Waste Generation (volume), Effort (time, personnel, logistics and infrastructure support, etc.), Time (to reach NFT: i.e. the exposure time of shore zone resources to the oil on the shoreline), and potential Environmental Consequences. Each of the four categories can be moved (dragged across the screen) so that the user-selected highest priority can be shifted to the left-most column with other categories decreasing towards the right. As the priorities are changed, the treatment options that would be Appropriate and Recommended, Potentially Appropriate, and Not Appropriate or Recommended would change within that column to reflect, for example, amounts of waste generated. This animated interactive screen would provide graphic indicators, for example, with the types and relative or (calculated) volumes of each waste type associated with each shoreline treatment option and treatment endpoint criterion. Importantly, each response option cell has links to the supporting data and information base to explain and understand the resulting consequences and ramifications of the selection.
The effect of the selection of a specific response method (e.g. low-pressure washing and recovery, or mechanical removal) for the input scenario would be shown on the output screen in terms of the relative consequences of that choice with respect to Waste, Effort, Time, and potential Environmental Consequences. In the example provided in Figure 8, the selection of mechanical removal to achieve the 90% oil reduction NFT scenario would take two weeks to complete and the primary consequence would be associated with the high volume of Waste Generation, whereas the primary consequence of the low pressure wash and recover option to achieve that same result would be more time (3 months) and the very high level of the Logistics Effort involved. The output screen design is intended to be a “headline metric”, which highlights the key messages in a manner similar to that used for the “Frost Days” data in Figure 2.
FEATURES OF THE MPRI SRP DECISION SUPPORT TOOL
The MPRI SRP Decision Support Tool is intended to be a dynamic, interactive, multi-layered, geographic-based seasonal simulation model for shoreline oil spill response decision analyses. The Tool is a hybrid object that is part document and part model (“documodel”) that combines data with visualization. The Tool is multi-layered so that users can build specific custom-designed scenarios. The interface includes access to maps, tables, data, and text at increasing levels of detail. The SRP Decision Support Tool can be tailored to provide visual images and data, including access to real time webcam images, meteorological and oceanographic data, or typical time-specific conditions.
The quantitative analyses will be derived from, and summarize, science-based publications and other data. The documodels are loaded with many science-based assumptions and geographical scenarios to communicate the consequences of decisions for site- and time-specific scenarios (for example, an individual small grid cell location and by a specific month) without a live internet link or with the ability to interrogate synoptic real-time environmental data. For key variables, or where there is uncertainty in particular, these features are provided as a screen slider, check box, or radio button (or other on-screen visual cue) that the user can manipulate or change to immediately see the effect of that change on a range of metrics.
Although this project focuses on the coastal waters of Canada, the Tool is applicable globally. In some respects the Tool is similar to the Arctic Council Circumpolar Oil Spill Viability Response Analysis (COSVRA) for offshore response (EPPR 2017) as the treatment logic includes a set of disqualification rules that identify which treatment options are appropriate and/or viable for different shore types, oil types, and seasonal environmental conditions.
The SRP Decision Support Tool is targeted to stakeholders and the public who may be less comfortable with technical papers, large reports or data bases, as well as to scientists, students and technical specialists involved with oil spill response planning, preparedness or operations. The power of this medium is to convey a complex subject and to enable a user to grasp keen insights and so understand the consequences of intervention decisions. The Tool is expected to be available through the web for easy access, download, or use with cloud access.
The study is part of the Oil Translocation Theme project within the Multi-Partner Research Initiative (MPRI) agreement between Fisheries and Oceans Canada and Concordia University, under the Ocean Protection Plan (OPP). An “Oiled Shoreline Response Support Tool” workshop to develop the conceptual model was supported by TransMountain and involved the co-authors CA, GD, EO and ET, as well as Gary Sergy and Blair Humphrey whose participation and contributions are gratefully acknowledged.