Safety Arguments Research Articles

Radionuclide transport and retardation in uplifting granitic rocks: Part 2 – Modelling coupled processes in uplift scenarios

Uplifting fractured granitic rocks occur in substantial areas of countries such as Japan. Some of these areas might be considered when siting a deep geological repository for radioactive wastes. A repository site would be selected in such an area only if it is possible to make a safety case, accounting for the changing conditions during uplift. The safety case must include robust arguments that chemical processes in the rocks around the repository will contribute sufficiently to minimise radiological doses to biosphere receptors. Numerical modelling is an important aspect of making these arguments. To provide confidence in the safety arguments, numerical models need to be sufficiently realistic, but also parameterised conservatively (pessimistically). However, model development is challenging because uplift involves many complex couplings between groundwater flow, chemical reactions between water and rock, and changing rock properties. The couplings would affect radionuclide mobilisation and retardation, by influencing diffusive radionuclide fluxes between groundwater flowing in fractures and effectively immobile porewater in the rock matrix (rock matrix diffusion, RMD) and radionuclide partitioning between water and solid phases, via: (i) mineral precipitation/dissolution; (ii) mineral alteration; and (iii) sorption/desorption. It is difficult to represent all this complexity in numerical models while showing that they are parameterised conservatively. Here we present a modelling approach, illustrated by simulation cases for some exemplar radioelements, to identify realistically conservative process conceptualisations and model parameterisations.

Religious dress in the healthcare setting: Unpacking legal arguments and balancing individual rights

This article explores the complex debate surrounding healthcare providers seeking to express their religious beliefs through the wearing of religious attire considering critical legal, ethical, and practical aspects. Drawing insights from the jurisprudence of the European Court of Human Rights and European Court of Justice, it sheds light on the implications for healthcare settings. The study highlights the sensitivity of the healthcare context, where specific arguments surface, such as the debated significance of religious neutrality given the vulnerability of patients and concerns about health and safety. Difficulties may arise when patients feel discomfort when treated by visibly religious healthcare providers. However, patients’ preferences regarding the religious neutrality of healthcare providers cannot be generalised, especially when considering the needs of certain minoritised communities. Furthermore, questions can be raised about the admissibility of potentially accommodating prejudice through prohibitions on religious dress. The article further examines hygiene and safety arguments used by healthcare institutions to restrict religious attire, arguing that while health and safety standards are crucial, scientific support for outright bans is limited, especially considering the possibility of reasonable adjustments. The central theme is the necessity for balanced considerations while upholding religious freedom and prioritising patient care, with particular emphasis going on the idea that the needs of patients cannot be generalised.

Confidence assessment in safety argument structure - Quantitative vs. qualitative approaches

Some safety standards (e.g., ISO 26262 in automotive industry) propose the use of argument structures to justify that the high-level safety properties of a system have been ensured. The goal structuring notation (GSN) is a graphical tool used to represent these argument structures. However, this approach does not address the uncertainties that may affect the validity of the arguments. Thus, some authors proposed to complement GSN patterns with a quantitative confidence assessment procedure. In this paper, we first present a refined procedure that expresses the relation between premises (pieces of evidence) and the conclusion (top-goal to be demonstrated) using logical expressions. Then using Dempster-Shafer theory, we quantify uncertainty on each expression to build an explicit mathematical formula for propagating uncertainty to the conclusion. Inputs for the propagation model are collected from experts and transformed into numerical values using an improved elicitation model. Afterwards, we introduce a purely qualitative alternative to the quantitative procedure based on the theory of qualitative capacities. Finally, we adapt the propagation and elicitation models to this framework.

Design and Assurance of Safety-Critical Systems with Artificial Intelligence in FPGAs: The Safety ArtISt Method and a Case Study of an FPGA-Based Autonomous Vehicle Braking Control System

With the advancements in utilizing Artificial Intelligence (AI) in embedded safety-critical systems based on Field-Programmable Gate Arrays (FPGAs), assuring that these systems meet their safety requirements is of paramount importance for their revenue service. Based on this context, this paper has two main objectives. The first of them is to present the Safety ArtISt method, developed by the authors to guide the lifecycle of AI-based safety-critical systems, and emphasize its FPGA-oriented tasks and recommended practice towards safety assurance. The second one is to illustrate the application of Safety ArtISt with an FPGA-based braking control system for autonomous vehicles relying on explainable AI generated with High-Level Synthesis. The results indicate that Safety ArtISt played four main roles in the safety lifecycle of AI-based systems for FPGAs. Firstly, it provided guidance in identifying the safety-critical role of activities such as sensitivity analyses for numeric representation and FPGA dimensioning to achieve safety. Furthermore, it allowed building qualitative and quantitative safety arguments from analyses and physical experimentation with actual FPGAs. It also allowed the early detection of safety issues—thus reducing project costs—and, ultimately, it uncovered relevant challenges not discussed in detail when designing safety-critical, explainable AI for FPGAs.

Reliability Assessment and Safety Arguments for Machine Learning Components in System Assurance

The increasing use of Machine Learning (ML) components embedded in autonomous systems—so-called Learning-Enabled Systems (LESs)—has resulted in the pressing need to assure their functional safety. As for traditional functional safety, the emerging consensus within both, industry and academia, is to use assurance cases for this purpose. Typically assurance cases support claims of reliability in support of safety, and can be viewed as a structured way of organising arguments and evidence generated from safety analysis and reliability modelling activities. While such assurance activities are traditionally guided by consensus-based standards developed from vast engineering experience, LESs pose new challenges in safety-critical application due to the characteristics and design of ML models. In this article, we first present an overall assurance framework for LESs with an emphasis on quantitative aspects, e.g., breaking down system-level safety targets to component-level requirements and supporting claims stated in reliability metrics. We then introduce a novel model-agnostic Reliability Assessment Model (RAM) for ML classifiers that utilises the operational profile and robustness verification evidence. We discuss the model assumptions and the inherent challenges of assessing ML reliability uncovered by our RAM and propose solutions to practical use. Probabilistic safety argument templates at the lower ML component-level are also developed based on the RAM. Finally, to evaluate and demonstrate our methods, we not only conduct experiments on synthetic/benchmark datasets but also scope our methods with case studies on simulated Autonomous Underwater Vehicles and physical Unmanned Ground Vehicles.

Operational Design Domain-Driven Coverage for the Safety Argumentation of Automated Vehicles

Automated Vehicles aim to increase road safety as automated driving systems (ADS) take over the human driving task in the operational design domain (ODD), introducing severe challenges for safety validation. Pure driving over many kilometers to gather enough evidence for a safety argument is not feasible. Scenario-based testing is an approach to overcome this, but challenges like parameter discretization still prevail, hindering safety assurance. This work proposes contributions towards a traceable and efficient safety argumentation for ADS built upon ODD coverage. First, ODD coverage is thoroughly quantified across all scenario levels, assuming distribution functions’ availability for the scenario parameters. Secondly, a sampling method for n-dimensional scenario parameter distributions is proposed. The provided algorithms adapt an initial k-means clustering using pre-defined boundary conditions requiring significantly fewer scenarios. Furthermore, a risk metric for urban intersections is presented for scenario evaluation. The risk metric consists of two parts, scene prediction of traffic participants (TPs) and risk assessment. The scene prediction uses a manoeuvre-based motion model with a data-driven approach towards trajectory prediction, increasing the validity. For the risk assessment, a probabilistic risk prediction for the TPs is performed for each scenario scene. The risk metric shows a reasonable tradeoff between sensitivity and specificity, outperforming time-to-collision. These contributions are exemplarily applied at an intersection using a simplified setup for generating TPs and ego vehicle trajectories. The results indicate that an increased safety argumentation is enabled using the proposed methods alongside a coverage process, facilitating further research.

Formal Development of Safe Automated Driving Using Differential Dynamic Logic

The challenges in providing convincing arguments for safe and correct behavior of automated driving (AD) systems have so far hindered their widespread commercial deployment. Conventional development approaches such as testing and simulation are limited by non-exhaustive analysis, and can thus not guarantee safety in all possible scenarios. Formal methods can provide mathematical proofs that could be used to produce rigorous evidence to support the safety argument. This paper investigates the use of differential dynamic logic and the deductive verification tool KeYmaera X in the development of an AD feature. Specifically, this paper demonstrates how formal models and safety proofs of different design variants of a Decision & Control module can be used in the safety argument of an in-lane AD feature. In doing so, the assumptions and invariant conditions necessary to guarantee safety are identified, and the paper shows how such an analysis helps during the development process in requirement refinement and formulation of the operational design domain. Furthermore, it is shown how the performance of the different models is formally analyzed exhaustively, in <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">all</i> their allowed behaviors.

Graphical safety assurance case using Goal Structuring Notation (GSN) — challenges, opportunities and a framework for autonomous trains

The development of fully autonomous vehicles is an ambition that took seed in the automotive industry a few years ago and is now growing in the railways considering their benefits. The main objective of autonomous train is to perform its operations and assure its mission with an acceptable safety level in all possible operational conditions. Such an objective needs to be supported by a safety demonstration. In order to authorize the operations of railway systems, they must be proven safe. This requires a technical and operational safety assessment, and also a safety assurance process during the system’s whole life-cycle. The goal of such activities is to ensure that designed systems comply with railway safety standards and regulations. Both safety arguments and evidences are required to demonstrate that this compliance is achieved. These sets of evidence are documented in a so-called safety case. Recently, graphical safety cases, such as Goal Structuring Notation (GSN)-based safety case, have become an interesting alternative to narrative reports and plain texts. The graphical structure and visual properties improve the presentation and comprehension of the safety arguments. In this paper, we firstly review the use of the GSN for building graphical safety case for different transportation systems, with a focus on the railway domain. Then, we discuss the opportunities and challenges of considering such an approach in railway and we propose a high-level framework for building the GSN-based safety assurance case for the autonomous trains.

The Role of Explainability in Assuring Safety of Machine Learning in Healthcare

Established approaches to assuring safety-critical systems and software are difficult to apply to systems employing ML where there is no clear, pre-defined specification against which to assess validity. This problem is exacerbated by the “opaque” nature of ML where the learnt model is not amenable to human scrutiny. Explainable AI (XAI) methods have been proposed to tackle this issue by producing human-interpretable representations of ML models which can help users to gain confidence and build trust in the ML system. However, little work explicitly investigates the role of explainability for safety assurance in the context of ML development. This paper identifies ways in which XAI methods can contribute to safety assurance of ML-based systems. It then uses a concrete ML-based clinical decision support system, concerning weaning of patients from mechanical ventilation, to demonstrate how XAI methods can be employed to produce evidence to support safety assurance. The results are also represented in a safety argument to show where, and in what way, XAI methods can contribute to a safety case. Overall, we conclude that XAI methods have a valuable role in safety assurance of ML-based systems in healthcare but that they are not sufficient in themselves to assure safety.

Dependable learning-enabled multiagent systems

We are concerned with the construction, formal verification, and safety assurance of dependable multiagent systems. For the case where the system (agents and their environment) can be explicitly modelled, we develop formal verification methods over several logic languages, such as temporal epistemic logic and strategy logic, to reason about the knowledge and strategy of the agents. For the case where the system cannot be explicitly modelled, we study multiagent deep reinforcement learning, aiming to develop efficient and scalable learning methods for cooperative multiagent tasks. In addition to these, we develop (both formal and simulation-based) verification methods for the neural network based perception agent that is trained with supervised learning, considering its safety and robustness against attacks from an adversarial agent, and other approaches (such as explainable AI, reliability assessment, and safety argument) for the analysis and assurance of the learning components. Our ultimate objective is to combine formal methods, machine learning, and reliability engineering to not only develop dependable learning-enabled multiagent systems but also provide rigorous methods for the verification and assurance of such systems.

Cross-country cultural and economic freedom influences on the relationship between economic policy uncertainty and ADR mispricing

Extending the findings by Grossmann and Ngo (2020) that ADR mispricing is higher upon heightened U.S. economic policy uncertainty (EPU), we document that culture and economic freedom differences moderate the relationship between EPU and ADR mispricing. In line with a combined flight to safety and familiarity argument, an increase in U.S. EPU is associated with higher mispricing for ADRs from countries with cultures and economic freedom similar to the U.S. The mispricing of these ADRs is also impacted by domestic EPU considering cultural differences. This is especially the case for cultures perceived to have lower risk aversion or home bias and might be explained by the familiarity argument as well as domestic investors’ flight toward domestic stocks that have, according to the product market spill over hypothesis, more of a global exposure (e.g., cross-listed stocks). The main findings are not altered by the Covid-19 period or other robustness checks.

Playing to the gallery: the invocation of human rights, legal actors, and outside audiences in debates on roadside drug testing in the Australian Capital Territory

ABSTRACT Roadside drug testing (‘RDT’) was introduced in the Australian Capital Territory (‘ACT’) through a private member’s Bill that was fiercely debated in the ACT parliament. It was the first RDT scheme in Australia subject to human rights scrutiny, and thus parliamentarians had to weigh road safety arguments against human rights considerations in determining whether the proposed measures were proportionate. This paper focuses on the debate over RDT in the ACT parliament, how human rights and safety concerns figure, and how legal actors and ‘outside audiences’ are invoked. The debate was marked by frequent appeals to family members present in the public gallery and legal actors such as the Chief Police Officer, the Human Rights Commissioner, and socio-legal experts outside the chamber. Drawing from cultural legal studies scholarship on legal insiders and outside audiences in legal proceedings, this paper analyses how these stakeholders figure in human rights debates on RDT, and with what effects. The paper argues that the invocation of outside audiences reflects what Crawley and Tranter term ‘a need to bring the human back’ but cautions that this humanising process only brings some humans in, and therefore connects human rights to only certain human lives in certain ways.

The Missing Link: Developing a Safety Case for Perception Components in Automated Driving

&lt;div class="section abstract"&gt;&lt;div class="htmlview paragraph"&gt;Safety assurance is a central concern for the development and societal acceptance of automated driving (AD) systems. Perception is a key aspect of AD that relies heavily on Machine Learning (ML). Despite the known challenges with the safety assurance of ML-based components, proposals have recently emerged for unit-level safety cases addressing these components. Unfortunately, AD safety cases express safety requirements at the system level and these efforts are missing the critical linking argument needed to integrate safety requirements at the system level with component performance requirements at the unit level. In this paper, we propose the Integration Safety Case for Perception (&lt;b&gt;ISCaP&lt;/b&gt;), a generic template for such a linking safety argument specifically tailored for perception components. The template takes a deductive and formal approach to define strong traceability between levels. We demonstrate the applicability of ISCaP with a detailed case study and discuss its use as a tool to support incremental development of perception components.&lt;/div&gt;&lt;/div&gt;

Treatment Approaches for First Episode and Early-Phase Schizophrenia in Adolescents and Young Adults: A Delphi Consensus Report from Europe.

PurposeAlthough first-episode psychosis (FEP) in youth, particularly early-onset schizophrenia (EOS), is managed similarly to adult-onset schizophrenia, few antipsychotics are approved for people aged 13–18 years. We aimed to explore areas of uncertainty in EOS management and provide evidence-based recommendations to mental health specialists. We used the Delphi methodology to gain knowledge in areas lacking evidence-based strategies. This standardized methodology consists of the development of a questionnaire by content experts, which is then submitted to a broader panel of professionals (panelists) to survey their level of agreement on the topics proposed.Materials and MethodsThe developed questionnaire covered patient management from diagnosis to maintenance treatment and was administered to a broader panel of specialists across Europe. Based on an analysis of responses received in this first round, the items that needed further insight were submitted to the panel for a second round and then reanalysed.ResultsAn initial set of 90 items was developed; in round I, consensus was reached for 83/90 items (92%), while it was reached for 7/11 (64%) of the items sent out for rerating in round II. Feedback for rounds I and II was obtained from 54/92 and 48/54 approached experts, respectively. There was broad agreement on diagnostic standards, multimodal approaches and focus on adverse events, but uncertainty in terms of pharmacological strategies (including clozapine) in case of failure and antipsychotic dosing in younger patients.ConclusionDespite knowledge about diagnostic clues and integrated management of EOS, this study highlights the lack of standardization in treating EOS, with safety arguments having a major role in the decision-making process. Targeted clinical trials and systematic dissemination across Europe of current scientific evidence on the value of early intervention services is hoped to contribute to standardized and improved quality care for patients with early-phase psychosis and schizophrenia.

Safety Diagnostics and Degraded Operational Modes

Unmanned ground combat vehicles (UGCV) promise numerous ultimate military, civilian, and space applications. The focus of this research mainly deals with how a non-geometric hazard scenario may potentially lead to a mission ending situation and a non geometric hazard can be considered as any terrain feature or object that is adjudged to be non-traversable by virtue of its physical properties. The novelty of this research lies in how a UGCV in a military applications can overcome the challenges of traversing through ever changing natural obstacles when compared to the reinforcing obstacles which are encountered by autonomous vehicle in a conventional structured scenario (see Figure 1).Figure 1. Traversal across an off-road terrain facing uneven terrain and obstacles.An operational design domain (ODD) can help specify potential unsafe situations and restrict the vehicle’s operation within them. A complete ODD will guarantee that the safety arguments can be dealt in a streamlined manner and in scenarios of restricting the ODD, the overall availability of the system gets reduced. With a semantic ODD structure for the UGVC, the standard development of the safety process for an automotive system is achieved and with the already created safety protocols, the overall situation space is reduced during critical situations.DISTRIBUTION A. Approved for public release; distribution unlimited. OPSEC5045The process of identifying reduced operational domain (ROD) starts with the quantificationand analysis of maneuvers and further leads to the identification of critical situations. Criticalareas are conditions of the environment that cause unintentional behaviour which eventuallyleads to an accident. A state machine constituting nominal behaviour will form the basis for acomponent fault tree (CFT) which is used to identify conditions that cause critical situations.Subsequently, the corresponding risk of an identified critical situation is assessed as a part ofthis process.The safety diagnostics mainly depends on a probabilistic model-based controller examining adynamic environment in which the stochastic evolution depends on the input of observationsand the current behaviour of the UGCV. In order to model the real-life performance of safetycritical systems realistically and accurately, Markov chain and Bayesian filters are highlyuseful. The finite state essence of the discrete controller may possibly lead to incorrectbehaviour of the complete system if an unforeseen situation occurs and for which there is alack of any predefined contingency. For this purpose, it becomes important to have a sense ofa complete set of admissible scenarios and also to develop a structured decision-making processfor each of the previously mentioned scenarios. State machine and failure propagation treescan help in determining the failure probabilities which keep updating based on changes incircumstances and this would help define if a mission should continue. As a part of thisresearch, Markov decision process which forms the basis for decision making process isemployed to identify and compare a set of state sequences and this in turn would help inrealizing better maneuverability of the vehicle.The assignment of ROD can be accomplished with the complete risk assessment of the criticalsituations. By assigning the ROD, the aim of increasing the overall availability of the systemwhich degraded from the nominal driving behaviour is fulfilled and this would allow for thesafe operation of UGCV.

Cement as a component of the multi-barrier system: radionuclide retention in cementitious environments

Abstract. Safety concepts regarding nuclear waste disposal in underground repositories generally rely on a combination of engineered and geological barriers, which minimize the potential release of radionuclides from the containment-providing rock zone or even their transport into the biosphere. Cementitious materials are used for conditioning of certain nuclear waste types, as components of waste containers and overpacks, as well as being constituents of structural materials at the interface between backfilling and host rock in some repository concepts. For instance, the preferred option for the disposal of high-level waste (HLW) in Belgium is based on the supercontainer design, which consists of a carbon steel overpack surrounded by a thick concrete buffer (Bel et al., 2006). In the event of formation water interacting with cementitious materials, pore water solutions characterized by (highly) alkaline pH conditions will form. These boundary conditions define the chemical response of the radionuclides, but also influence the behaviour of neighbouring components of the multi-barrier system, e.g. bentonitic or argillaceous backfilling and host rock. Hardened cement paste or Sorel cement are considered to be main sorbing materials present in the near field of repositories for low- and intermediate-level waste (L/ILW). Hence, interactions of radionuclides with cementitious materials represent a very important mechanism retarding their mobility and potential migration from the near field (Wieland, 2014; Ochs et al., 2016). While the quantitative description of the sorption processes (usually in terms of sorption coefficients, i.e. Kd values) is a key input in the safety analysis of nuclear waste repositories, detailed mechanistic analysis and understanding of sorption phenomena provide additional scientific arguments and important process understanding, and thus enhance both the quality of safety arguments and the overall confidence in the safety assessment process. Research at KIT-INE dedicated to the interaction of cementitious materials with radionuclides is conducted in the context of different repository concepts, including clay (low- and high-ionic strength conditions), crystalline rock or rock salt. Experimental and theoretical studies are performed within the framework of national (GRAZ, BMWi) and international (CEBAMA and EURAD-CORI, EU Horizon 2020 Programme) projects, extending to third-party projects with several waste management organizations in Europe, e.g. SKB (Sweden), ONDRAF-NIRAS (Belgium) or BGE (Germany). The combination of classical experimental (wet chemistry) methods, advanced spectroscopic techniques and theoretical calculations provides both an accurate quantitative evaluation and a fundamental understanding of the sorption processes. Examples of recent studies at KIT-INE on radionuclide behaviour in cementitious systems in the context of both L/ILW and HLW will be presented in this contribution to explain methodologies, scientific approaches and results. The present state of knowledge as well as main remaining uncertainties affecting the retention processes of radionuclides in cementitious environments under different conditions will be critically discussed, also in view of current international research activities and repository projects.

Renormalizability of Alternative Theories of Gravity: Differences between Power Counting and Entropy Argument

It is well known that General Relativity cannot be considered under the standard of a perturbatively renormalizable quantum field theory, but asymptotic safety is taken into account as a possibility for the formulation of gravity as a non-perturbative renormalizable theory. Recently, the entropy argument has however stepped into the discussion claiming for a “no-go” to the asymptotic safety argument. In this paper, we present simple counter-examples, considering alternative theories of gravity, to the entropy argument as further indications, among others, on the possible flows in the assumptions on which the latter is based. We considered different theories, namely curvature-based extensions of General Relativity as f(R), f(G), extensions of teleparallel gravity as f(T), and Horava–Lifshitz gravity, working out the explicit spherically symmetric solutions in order to make a comparison between power counting and the entropy argument. Even in these cases, inconsistencies were found.

Cultivating conscience: Moral neurohabilitation of adolescents and young adults with conduct and/or antisocial personality disorders.

Individuals diagnosed with conduct disorder (CD) in childhood and adolescence are at risk for increasingly maladaptive and dangerous behaviors, which unchecked, can lead to antisocial personality disorder (ASPD) in adulthood. Children with CD, especially those with the callous unemotional subgroup qualifier (“limited prosocial emotions”/DSM‐5), present with a more severe pattern of delinquency, aggression, and antisocial behavior, all markings of prodrome ASPD. Given this recognized diagnostic trajectory, with a pathological course playing out tragically at the individual, familial, and societal level, and lack of effective remediation via current standards of care, we posit an alternate treatment approach; case‐specific compulsory moral habilitation aimed at rectifying the undeveloped affective domain of adolescents and young adults suffering from these disorders. We begin with a brief historical overview of response to mental illness, review CD and ASPD diagnostic criteria and treatment limitations, and posit a unique neurohabilitative approach that focuses on the absent affective moral development of these populations. Next, we invoke a public health safety argument to justify case‐specific compulsory moral habilitation, discuss neurotechnologies to be considered in treatment, and conclude with ethical considerations and suggestions for further research.

Criticality Analysis for the Verification and Validation of Automated Vehicles

The process of verification and validation of automated vehicles poses a multi-faceted challenge with far-reaching societal, economical and ethical consequences. In particular, fully automated vehicles at SAE Level 4 and 5 will be expected to operate safely in an arbitrarily complex, infinite-dimensional domain called open context . In order to give structure to the open context, we propose a methodical criticality analysis that maps an infinite-dimensional domain onto a finite and manageable set of artifacts that capture and explain the emergence of critical situations for automated vehicles. We propose a combined approach of expert-based and data-driven methods to identify relevant phenomena and explain the underlying causalities. Leveraging on abstraction, we define a clearly laid out process that converges towards a manageable set of artifacts based on two assumptions on the nature of traffic. A criticality analysis precedes the design phase of an automated vehicle and is therefore located outside the V-model. As the open context is analyzed independently of a concrete realization, it is relevant for any automated vehicle operating within that domain. Therefore, its results can subsequently be used to derive safety principles and mitigation mechanisms for automated driving and to set up a coherent safety argument for the homologation process.

Reconciling Conflicting Interests of Coastal and Riparian States: The Hard Case of Black Sea Straits

There are two basic understandings of the regime of the Black Sea straits: the Black Sea straits as a legal regime and the Black Sea straits as a political regime [1]. The legal assessment of the Black Sea Straits regime requires determining what the existing regulation of the Straits is, how open the Straits are to international navigation, and if closed, whether there are real legal grounds for closing straits while the reference to the Black Sea Straits as a political regime allows for the possibility that straits may be closed for ensuring the security of Turkey and the Black Sea riparian states [1]. It is worth noting that arguments advanced by international legal scientists on the Black Sea straits as legal regime fundamentally differ from each other. Some scientists consider the Montreux Convention to be a major problem in the legal regulation of the Black Sea straits. They consider it necessary for Turkey to recognise the 1982 Convention on the Law of the Sea as a legally binding treaty [2]. Others argue that the main problem in regulating the Black Sea straits is the unilateral regulations adopted by Turkey (1994, 1998 and 2003 Regulations), which, in their view, violate the basic norms of the Montreux Convention, especially the regime of free passage through the straits established by this Convention [3; 4]. Another group of scientists believes that although the regime of the Black Sea straits is significantly restricted by Turkish unilateral regulations, these acts are aimed at protecting the marine environment and safety, and, therefore, the Turkish policy of regulating the Black Sea Straits is legally justified [1; 5]. This article is dedicated to the international legal regulation of navigation in the Black Sea Straits. The aim of the paper is to evaluate the current regime of the Black Sea Straits, the relationship and differences between the regime established by the Montreux Convention and the unilateral acts adopted by Turkey on the regulation of traffic in the Black Sea Straits, and to answer the question whether the urgent need to protect the natural environment and maritime safety entitles Turkey to restrict the regime established by the Montreux Convention. Thus, special attention will be drawn to the Montreux Convention, the rules and recommendations adopted by the International Maritime Organisation and the case law of international courts. In the view of the authors, the environmental and safety arguments put forward by Turkey for restricting navigation through the Black Sea Straits have two conceptual dimensions. First, these arguments are acceptable when it comes either to introducing norms related to the movement of ships to ensure safety of navigation or providing an obligation of notification to the Turkish authorities [6]. Another important thesis advanced by this article is that in each particular case, the regulations adopted by Turkey should be interpreted in the light of the recommendations made by the International Maritime Organisation. The main rationale of this argument is that under the existing regulations, Turkish authorities can still suspend the movement of ships in the straits for various reasons, some of which are quite vague. However, the article showcases that Turkey can, in case of pressing environmental need, when there is an urgent interest in the protection of the natural environment, act with the motive of protecting the natural environment, regardless of whether this action derives from a particular international treaty.