PROJECT MARINER: ENHANCING HNS PLANNING AND PREPAREDNESS VIA INNOVATIVE TRAINING TOOLS

One product developed under MARINER by Public Health England (PHE) offers bespoke training via an open source downloadable software tool, delivering region- specific desk top exercises simulating maritime incidents at local, regional or international levels.

The tool incorporates a database of HNS and environmental information, which users populate for their own region, together with a modelling interface and a library of exercise materials providing scenarios, feedback and debrief documentation. Exercises are generated based upon the HNS type chosen, the scale of the incident, its location and the prevailing seasonal conditions. E-learning modules around cross border alerting and co-operation complete the overall training package.

The software produces maps, datasheets and modelling simulations, all aligned with injects encompassing each phase of the incident management cycle and incorporating options for cross border alerting and response. The system is to be piloted across several EU regions with a finalised version planned for release in late 2017.

Using this system it is intended that planning, preparedness and response arrangements can be routinely tested with realistic incident simulations, incorporating the specific features and limitations of users regions to enhance response capabilities and highlight critical local factors.

Hazardous and Noxious Substances (HNS) can be defined as “Any substance other than oil, which, if introduced into the marine environment is likely to create hazards to human health, to harm living resources and marine life, to damage amenities or to interfere with other legitimate uses of the sea” (IMO, 2000).

Global estimates indicate up to 50,000 different HNS carried by sea (IMO, 2009), with quantities in excess of 200 million tonnes carried annually (IMO, 2016). Quantities are also increasing, as the post 2013 generation of container ships (New Panamax and Triple E ships) at 12,000 to 18,000 twenty foot equivalent units (TEU), have much larger capacity compared to previous Post Panamax vessels at around 4,000 to 8,000 TEUs (BBC, 2013). The container ship MSC Napoli which ran aground off the coast of Devon, for example, carried 2318 containers at 4,419 TEU, a number of which lost overboard, washed ashore resulting in the need for a multiagency response during the 2007 incident (ARCOPOL 2012).

Data identify over 100 incidents reported globally between 1998 and 2013, with a cumulative volume released of 1,560,000 m³ (IMO, 2016). Furthermore, with projections of increased shipping of chemicals and an expanding range of HNS being transported, some increase in the number of incidents involving HNS may be expected.

Large maritime incidents involving HNS, whilst rare, have the potential for major impact upon both the environment and human health. In the case of oil spills, the major concerns tend to be related to environmental and ecological impact with response and recovery techniques well defined for such impacts. However, when spills involve HNS, risks to human health can become a significant aspect of response and recovery, e.g. MV Rena (CEDRE, 2016). Furthermore, with such a range of chemicals classified as HNS, approaches to managing such incidents are far less well defined than oil spills. Thus it is imperative to develop robust planning and response mechanisms which can be readily employed to protect both the marine environment and human receptors.

Much work has been undertaken in this field but outcomes are fragmented. In view of this, a 2 year project (“MARINER”) was initiated in 2016 aimed at collating global best practise into innovative tools for planning and response. Covering the European Atlantic Area and funded by the EU research programme DG ECHO, the project involves collaboration of a range of agencies from UK, France, Spain and Portugal. The project builds on HNS research initiated by projects such as ARCOPOL, which developed a range of tools and guides for HNS shoreline response and preparedness e.g. tools and approaches to prioritise HNS based upon risks to public health and the marine environment (Harold, 2014; Neuparth, 2011).

MARINER is divided into several tasks (Figure 1). Tasks B and D focus on the capture and compilation of existing knowledge and response protocols. Task C aims to improve the interoperability and the operational use of models for forecasting HNS fate, transport and impact. Task E focusses on training using innovative tools and approaches. Within Task E PHE’s Centre for Radiation, Chemical and Environmental Hazards Wales (CRCEW) has developed a tool to simulate maritime incidents offering users the option to create bespoke training exercises.

Exercising plans is a key component of preparedness. Whilst the most effective exercises involve actual deployment of personnel and resources in the field, such events are often too costly and logistically difficult to organise on a routine basis. As such, organisations will often use “desk-top” exercises to provide a more economical and practical approach for routine training. Although well defined, desk-top techniques can often be quite generic and lack an appreciation of local idiosyncrasies. Furthermore, even developing bespoke exercises at desk-top level can involve considerable use of resources.

In view of this it was proposed to develop a software tool capable of incorporating local and regional specific characteristics to produce bespoke exercises and supporting resources, whilst following generic themes to capture the regional nuances critical to provide realistic test conditions. The tool will form part of a virtual learning environment where practitioners can access interactive, web-based training materials that provide the necessary underpinning knowledge for these scenarios. Knowledge can then be tested and applied via interactive case-studies

Whilst a range of web-based exercise packages are available such as the Federal Emergency Management Agency Crisis Management package (US Homeland Security, 2015), none are specifically aimed at maritime response or contain options that can reflect local conditions and features. Thus it was felt that the proposed tool could offer a unique and innovative approach to training for planners and responders.

In order to develop the tool, several key design parameters were set:

• Applicable to all EU Atlantic regions, by incorporating local and regional coastal and met-ocean characteristics,

• Capable of applying to a range of different HNS types and incident sizes,

• Reflecting seasonal variations, which can have a marked effect on shoreline impact,

• Providing a variety of editable scenarios to test local and regional response plans on a site specific basis and also to include wider cross border impacts, where applicable.

The tool has been developed as a web-based, free to access application, housed within a secure, managed server. In this way the tool is accessible across the widest range of hardware (desk tops, tablets, smart phones) with less need for operating system updates.

The tool comprises a database of preloaded information on HNS, a library of exercise materials, and interfaces for mapping and modelling. The user also populates the database with their own regional receptor information prior to use (Table 1, columns B and C).

Receptor information combines environmental, health and socio-economic data relevant to the region, uploaded as a series of GIS mapping layers by the user. Environmental data comprise regional shoreline characteristics, as defined within existing methods (NOAA, 2002). Health and socio-economic indicators comprise coastal population estimates, ports, commercial fisheries and bathing waters. These specific receptor categories were chosen to be reflective of the critical considerations when assessing the types of response to a pollution incident at a particular location. For example, in areas of rocky shore, the type of response selected will differ compared to a similar spill impacting mud flats. Likewise, priorities for human health, public safety and socio-economic response will be guided by the key parameters loaded into the tool.

Preloaded information on HNS covers the key behavioural categories that could be involved in an incident, based upon recognised classifications; Gas, Floater, Dissolver and Sinker (CEDRE, 2004). This approach was chosen with an element of pragmatism to illustrate how widespread an impact would be from a specific HNS category and which environmental compartments would be most affected. Each behaviour class is represented by a surrogate chemical in order to produce fate and transport models and to provide indicative chemical datasheets within exercise outputs. If however a user wanted run an exercise for a specific chemical then the existing datasheets could be edited to reflect this.

A series of preloaded exercise scenarios, selected to be representative of the more common maritime pollution incidents, form the final element of the exercise tool database. The exercise chosen encapsulates all aspects of the incident management process, commencing with alerting and detection, risk assessment and data interpretation, response and recovery and risk communication.

Specialist software developers from Cardiff Metropolitan University were contracted to produce the application based upon the concepts provided by CRCEW. The tool was built using Joomla architecture (Figure 2), an open source CMS build using PHP as programming language, MySQL as a database and Apache Tomcat as a web server (Yang, 2008).

The Extension layer contents incorporate preloaded and user input materials together with menus and unique access settings for each user. The Application layer consists of JInstallation, JAdministration, JSite and XML-RPC, providing installation and administration functions. The Framework layer consists of libraries that are required to run the application, and integrate third party extensions. Plug-ins, including Google Maps, extend functionality.

A customised extension named ExGenerator was developed to meet the application requirements, built based on the MVC pattern (Supaartagorn, 2016) and using Joomla Framework, with requests to a controller sent by a graphical user interface (GUI). ExGenerator handles the input using Google Maps API integration, JQuery and Bootstrap design (Figure 3).

Fate and transport modelling provides an integral element of the tool outputs. To achieve this, the tool automatically requests simulations and results for chosen HNS over a pre-defined period from an external server operated by Action Modulers. Action Modulers are specialist model developers based in Portugal. They have many years’ experience of developing fate and transport models for oil and HNS in the marine environment. They are the lead partner on Task C of the MARINER Project and have been involved in many other international maritime pollution response projects including ARCOPOL.

The numerical model MOHID that solves hydrodynamics and spill simulations fate is integrated in a web-service that is both used for the Mariner project platform and for the tool described in this article (Figure-4). The web service is able to manage multiple data layers, including met-ocean forecasting data, management scenario definition via User Interface (MOHID Studio), all saved in database for persistence. All the data layers are accessible via web services specifically, OGC WMS (maps) and REST API (time series, chemical and oil spill model interaction) which facilitate the interoperability with multiple websites and tools. A specific web platform with modelling results was established for the MARINER

The numerical model MOHID includes a lagrangian transport module, where simulated pollutants are represented by a cloud of discrete particles (or super-particles) (Fernandes, 2016). MOHID is a public-domain open-source system, developed following a modular structure combined with object oriented programing (Neves, 2013).

In the chemical spill version, spilled mass is transported in 3D space and time. The horizontal movement is controlled by currents, wave-induced velocity (Stokes Drift), wind-drift velocity in the surface layer (for floating substances), spreading, and horizontal turbulence. The vertical movement is estimated in accordance with vertical advection from currents, rising velocity, sinking velocity, and turbulent dispersion. The model estimates the distribution of chemical (as mass and concentrations) in all environmental compartments, tracking each phase separately and incorporating chemical degradation, being estimated as a constant rate of “decay” specific to each environmental compartment.

Figure 4 illustrates the process for simulation requests with Action Modulers web services (model platform). ExGenerator sends requests to the model platform using Ajax (via REST API). The platform reads the user defined options (spill substance type and location, scenario to run), selects the scenario with met-ocean forcing and runs the spill simulation for the defined period. After the simulation is finished requests can be made to return simulation results in JSON encoded data (http://www.json.org). JSON is designed to be a data exchange language user readable and easy to parse. JSON is directly supported inside JavaScript (Klarr, 2007) and is best suited for such applications (Nurseitov, 2009). The tool uses JavaScript in ExGenerator to parse the data and generate ZIP files.

In its development phase, the tool has been piloted for the Bristol Channel, a large tidal estuary which separates South Wales and South West England (UK) and adjoins the Atlantic. This area was selected as being one of the main shipping areas of the UK Atlantic coast, having a cross-border dimension and being particularly challenging for responders with regard to its met-ocean characteristics and shoreline variability. The area comprises approximately 160 miles of coastline and includes several major tourist beaches, commercial ports, coastal cities and a range of infrastructure. In addition CRCEW provides representation to the specialist group (Environment Group) that advises on response in this region.

Environment Groups provide regional advice and assistance to responders during an incident around environmental and health protection under the requirements of the National Contingency Plan developed by the UK Maritime Coastguard Agency (MCGA, 2015). The groups are comprised of scientific and medical representatives from environment, health and fisheries agencies, and also liaise with local authority emergency planners, port managers and representatives of the third sector such as wildlife protection charities. As part of their responsibilities under emergency planning and preparedness, these groups prepare sensitivity and response plans for their regional coastlines and exercise these regularly. The tool has been designed to help support and facilitate the development of such plans and exercises.

Piloting has comprised a 3- stage process involving population of the database with regional data, beta testing of the populated tool for functionality and a full live trial.

Stage 1 considered the availability of data sources and ease of generating these as compatible GIS layers. Coastal data were obtained from surveys undertaken as part of the normal works of environment and health agencies and available via free to access portals such as the UK government mapping portal “Magic” (Natural England, 2016). HNS data were collated following meetings with regional ports and review of traffic statistics. High resolution (0.5 km) oceanographic and tidal data for the Bristol Channel and regional meteorological data (NOAA, 2016) was collated to represent typical spring and neap tide seasonal conditions.

Collation and provision of coastline data was relatively straightforward, however obtaining specific data regarding HNS was difficult as ports were reluctant to disclose sensitive and client confidential information. As a result the tool was used with the surrogate HNS representing the key behavioural categories and reflecting all potential exposure routes following a release. In the case of met-ocean information, data was obtained by Action Modulers to define main seasonal parameters for the region (e.g. air temperature, wind velocity and direction, wave height, river inflow etc.) prior to its availability for modelling. These parameters were used to force MOHID implementation in Bristol Channel for each season scenario.

Stage 2 focussed on the outputs in terms of their timeliness, realism and coverage of the management cycle. This involved in-house operational testing as well as dissemination of procedures and exercise materials to MARINER project partners.

Once populated, the tool can operate in real time, producing scenarios and injects as the exercise proceeds or can be used to prepare materials in advance of the event. Exercises can be generated as paper hard copies or can be presented directly on screen.

The tool has a straightforward, user-friendly interface (figure 6). The user sets the location of the incident by entering latitude and longitude co-ordinates or by selecting location on the base map. The facilitator then selects which class of HNS is required and scale of the incident, the season in which the incident will occur, and the type of scenario.

In run mode, the tool will populate the exercise materials with the selected parameters; prepare regional mapping and request fate and transport modelling from the external server. The modelling is not instantaneous although, depending on the parameters selected and the time period covered by the model, the tool runtime should be in the order of minutes. The tool will notify the facilitator when ready, after which the exercise can begin.

The exercise provides an initial scene setting introduction including description of the incident and regional mapping. Injects then follow as prompted by the facilitator, updating events and providing defined questions to be responded to by players (Figure 7). Questions are open and do not define any specific actions but instead are aimed at prompting players to use the procedures and approaches defined within their own regional plans.

Exercises are designed within a standard template so injects and descriptive / supporting materials are populated by user inputs and generated as a defined sequence. Again there is an option for users to produce their own bespoke scenarios should local considerations warrant this e.g. an area where off shore discharges are common. This however must be structured to fit the predefined report sequence. The whole exercise process is designed to be completed during a half to one day training session, including introductions and briefing, and a final feedback and debrief.

Stage 3 is to be completed but will involve a full desk top exercise to be run live with responders from England and Wales in June 2017. Feedback will be used to finalise the tool for wider use. Additional pilots to appraise the tool and its applicability to other regions are planned to include Galicia in Northern Spain and Brest on the French Atlantic coast.

Applying the tool to a wider audience will require the provision of met-ocean data for fate and transport modelling. Currently the tool can link to high resolution data for several EU Atlantic areas, however beyond these, modelling would be limited to low resolution forecasting using open source services such as Copernicus (Copernicus, 2016). Furthermore it would be necessary for seasonal simulations to be developed for a region ahead of the tool being used, thus some form of initial notification process is needed. These are aspects that will require additional development, with perhaps the development of a process for users to submit bathymetric data, scenario parameters (e.g. air temperature, wind and waves for seasons), as part of a pre-notification of use. In the longer term it is hoped that the tool could help to develop a wider international community database, incorporating global high resolution data, available to all users and stimulating the opportunity for routine multinational training and exercise programmes.

Likewise HNS data cover the basic behaviour classes but do not cover possible permutations or mixtures of chemicals, which can often be the case for incidents involving container ships where many different products are transported together e.g. MSC Flaminia in 2012 (PHE, 2014). As stated previously, there is an element of pragmatism to this, offering an appropriate range of behaviours to test all potential impact scenarios, whilst minimising the need for in depth studies of HNS mixture behaviours. It should be noted however that in real incidents several behaviour classes could be encountered simultaneously, albeit likely to be in far smaller quantities compared to bulk chemical incidents.

An automated software application has been developed to generate bespoke maritime exercises. Using this system it is intended that planning and response arrangements can be routinely tested with realistic incident simulations, incorporating the specific features and limitations of the test regions to enhance capabilities and highlight critical local factors.

The process has involved conceptual design, development of methodology, software design and preparation of exercise materials. The beta version of the tool is being piloted for the Bristol Channel a coastal region on the UK Atlantic coast.

Exercises contribute to preparedness, particularly the human elements around capacity and resilience. This in turn reduces the impact of incidents, thereby protecting public health and the environment. It is believed that the tool provides a unique and innovative approach to such training and one which can provide bespoke realistic scenarios, applicable to the global community and without the requirement for extensive resources for preparation.

While the exercises themselves follow a generic format, the ability to change key elements such as location, HNS type, scale and temporal conditions, allows the generation of many different scenarios and thus enables the tool to be used repeatedly. Similarly, these factors allow facilitators to target exercises for example to known high risk areas such as busy shipping lanes and to simulate effects under realistic seasonal conditions.

It is further anticipated that the tool will also facilitate testing of cross border impacts and the application of multinational approaches. Cross-border co-operation is also a defined output of the MARINER project and for this the tool is being supplemented with a range of E-learning materials developed around International Health Regulations (WHO, 2005).

Future work will be focussed upon finalising the tool based upon the UK pilot tests and wider EU testing, and to further develop modelling options. The finalised tool will also be accompanied by a framework document for users together with a tutorial video. E-learning materials will also provide the underpinning theory and opportunity to apply and test knowledge through interactive, web-based modules.