Environmental Risk Assessment for Coastal Engineering Planning in Coastal Cities

中国环境学会  2011年 06月21日


  Luoping Zhang1, Weiwei Wang1 and Paolo F. Ricci2
  1 Environmental Science Research Center, Xiamen University, Xiamen, Fujian 361005, P. R. China
  2Environmental Health Sciences, University of Massachusetts, Amherst, MA, 01003, USA
   
  Abstract


  Most strategic environmental assessments for Coastal Engineering Planning (CEP) involve routine environmental impact assessments, often without environmental analysis and assessment of non-routine and transient events. Current environmental risk assessment (ERA) practices focus on either a limited set of factors or a single project. This study sets up a scientifically sound and practical framework of ERA for CEP, accounting for precautionary principles, ecological security, and community-based principles, thus providing scientific support for decision-making processes. The innovations of this work include: a general ERA framework and approach for CEP, and a retrospective risk analysis that is the basis for prospective risk assessment for CEP. 
  Keywords: Environmental risk assessment; Coastal engineering planning; China
   
  Introduction
   
  Environmental Risk Assessment (ERA) allows a causal analysis of environmental outcomes from routine to non-routine events (such as low probability events characterized by large magnitude or severity, or both). Environmental risk management and assessment is the process of identifying, evaluating, selecting, and implementing the best action, out of a set of possibly choices, to reduce risk to human health and ecosystems (ECO FRAM, 1999; Sergeant, 2000). Although different countries use different approaches and models for different types of ERA, the fundamental risk assessment methodology is common to all (Ricci, 2006; Power & McCarty, 2002). However, there is little ERA-based study to support coastal engineering planning (CEP). Although some researchers proposed theories and case studies of Strategic Environmental Assessments (SEA) for coastal development, especially for CEP, those focus on routine environmental impact assessment (EIA) instead of catastrophic hazards and their risk assessment. Moreover, current environmental risk assessment (ERA) practices focus on either a limited set of factors or a single project. There is no a systematic framework and approach of ERA for CEP.
   
  Risk assessment is both art and science. It is based on a process for establishing the likelihood and magnitude of adverse effects to human health and to the environment from specific chemical, biological, and physical agents (Bridges, 2003).  Although there are many definitions of risk ranging from frequentistic to Bayesian and fuzzy (Pearl, 2000; Gelman et al., 1995; Ricci, 2006), an often used practical measure is the combination of the probability of an event occurring and the magnitude of the outcome – the expected value of the magnitude of the outcomes over all events:   
  Expected Risk = Σ (pri*Mi)
   
  Magnitude (e.g., 100 prompt deaths) is measured on the X-axis (generally the positive half for accidents). Probability is measured as cumulative probability on the Y-axis. If the triplet probability, magnitude, and severity of the results from an accident are modeled, then the formula is extended to account for these, and possibly other dimensions of the accidents and its outcomes. At a higher level of description, Constini and Servida (1992) considered a complete quantitative risk analysis and assessment procedures, and suggested that it should include these steps: hazard identification, incident frequency and consequence estimating, risk calculation and risk mitigation. The methods that are available to accomplish these tasks for ERA include: (1) Expert Consulting: Expert Judgment, Delphic and other methods; (2) Quantitative methods: Audit Table, Matrix Method, and Ratio Method; (3) Probability and Statistical Methods; (4) Decision-theoretic analysis: AHP (Analytic Hierarchal Processes), FTA (Fault Tree Analysis) and ETA (Event Tree Analysis). These can be applied to either an area or a facility or even to a single event. However, there are few studies on ERA for coastal engineering and planning (Balas et al., 2004). Existing problems in studying on ERA include: uncertainties, lack of knowledge, and ambiguity (Hong & Apostolakis, 1993). To our knowledge, ERA for CEP shows no literature about these conditions.
   
  The importance of ERA for CEP is that environmental risk affects environmental health, security and sustainable development because it accounts for the complexity, accumulation, long-term, and regionality of CEP. The objectives of the paper are to set up a framework and approach of ERA for CEP. It will follow precautionary principles, ecological security, and community based principles.
   
  1. Characteristics of coastal zone and its planning
   
  The general characteristics of coastal zones are that they involve abundant resources and biodiversity, many vulnerabilities, high population densities, and large hazards (and thus risks) from natural and man-made sources. Compared to other planning of projects, CEP has unique characteristics: it is generally consists of an integration of objectives and factors, including the impacts from several engineering projects which are complex, with possibly multiple risk factors that often are cumulative and objectives. Therefore, errors in formulation of the decision-analytic aspects of decision-making can lead to higher risks and costs that were not foreseen in the initial analysis.
   
  2. Framework of ERA for CEP
   
  The framework of ERA for CEP is different from other ERA’s framework because CEP is a territorial, accounts for multiple projects and complex integrative plans. The framework is show in Figure 1.
   
  The most important and different aspect of our contribution to the ERA framework we set up for CEP is a scientifically sound retrospective risk assessment (RRA) process because it is formal and reproducible, and accounts for future information. Due to multiple projects and complex plans, CEP always contains more uncertainties and cumulative effects ranging from routine to transient risk impacts, all of which are hard to be predicted and assessed in the ERA. The RRA procedure asks to investigate: risk identification, probabilities, causal factors, outcomes, and the characterization of the consequence based on the rare accidents and routine events that have happened in the past time in entire study area. It thus provides reliable basis for the prospective risk assessment and risk management.
   
  More specifically, Figure 2 depicts the overall probabilistic basis for our analysis. Hypothetical inverse cumulative distribution functions for two Systems, A and B. The uncertainty about each curve is represented by upper and lower confidence limits on each curve, for each of the two systems. These curves are plausible within the relevant range of the data used to fit them, and thus their X- and Y- intercepts are not known. The X-axis measures magnitude, the Y-axis measures cumulative probability. In Figure 2 we portray two different systems, A and B, the inverse cumulative distributions for each, and the variability about each inverse distribution. Our analyses are based on probabilistic risk assessment method; the acronym PRA stands, particularly in reliability or failure analysis of technological systems, for probabilistic risk analysis. In our context, an event, such as a spill, can cause prompt deaths, injuries, delayed deaths and injuries, as well as property and other ecological damage. More precisely, that event (e.g., a marine accident) can cause a range of adverse outcomes: the spill may cause from zero damage, if nobody is around to, let us say, a 1000 birds killed. Thus, we are dealing with a random variable, the number of prompt bird deaths is measured on the abscissa and the magnitude is the number of birds killed by the spill: it ranges from 0 to a possibly very large number. As an example pertinent to our case study, a summary statistic for this random variable is the expected value of the distribution of those bird deaths, accounting for each probability of the discrete number of possible deaths from a single spill of certain magnitude, says 1000 liters of crude oil.

  The magnitude of the consequences being analyzed by each distribution must be homogeneous: they can be prompt deaths, delayed deaths, injuries requiring hospitalization, property damage and so on, but cannot be mixed (Grimmett & Stirzaker, 1982). The cumulative frequency is interpreted as the probability of an event being either less than or equal to a specific value or other quantity. Thus, the cumulative frequency of the events characterized by a magnitude between 1,001 and 10,000 is less than or equal to 0.933. On this rationale, it is also possible to calculate cumulative frequencies greater than a specific magnitude. Such frequency is calculated as (1-Fi), that is (1-0.933) = 0.067, which is the complement of 0.933. The complement of the cumulative distribution is often called the risk curve in technological risk assessments. It therefore follows that several different cumulative distribution functions, some discrete and some continuous information must be used to capture the full impact of accidents.
   
  To understand the logic of an accident we have used fault-tree and event-tree analysis, developing the frequencies associated with the gates of these trees from the available data or analogy to similar events. The results from combining fault-trees and event-trees are used to generate cumulative (or, as it is more usual, complements of the cumulative distribution) of the consequences (Bradford and Cooke, 2001) and thus provide the complete probabilistic methodology that underlies our case study of Xiamen Bay, Fujian Province, PRC.
   
  3. Case study
   
  3.1 Profile of study area and Coastal Engineering Planning (CEP)  
  The study area is in Xiamen Bay, located at the southeast of China, including the Western Bay, Tong’an Bay, Estuary and immediate seawater areas (Figure 3).

  Xiamen Bay has abundant natural, marine, and navigation resources, such as deep-line coast in Western Bay, Tong’an Bay and Estuary, sand beach in Eastern Bay and Southern Bay, a National Conservation Area with several endangered species including Chinese White Dolphin in Western Bay and Tong’an Bay, lancelet in Eastern Bay and Dadeng Bay, egret in Western Bay and Estuary, and mangroves in Western Bay and Estuary. The sensitive and vulnerable objectives covered most area of Xiamen Bay.
   
  The projects of CEP were showed in Figure 3. It included nine phases of project, and each phase contained several engineering projects of reclamation in the sea. The purposes of the Projects were for harbor, road and real estate construction.

  3.2 Retrospective Risk Assessment (RRA)  
  The RRA was conducted after CEP review, data collection and field investigation. The risk identification, probability, characterization and consequence were assessed by the rare accidents and routine events that have happened in the past decades of Xiamen Bay, and information from public involvement. The RRA evaluated the risk of oil spill, eutrophication, typhoon and tidal storm by statistical analyzing the data.  
  Since 1950, there were 62 reclamation projects in Xiamen Bay, over 70% happened in Western Bay and Tong’an Bay. It caused adverse impacts such as hydro-dynamic changing, seawater quality decreasing, and adverse impacts on ecology, especially in semi-enclosed bays. 
  In Xiamen Bay, the number of all types of ships entering and leaving ports is over 0.23 million per year, and is increasing. According to the statistical analysis of data from Xiamen Maritime Bureau, there were 133 ship accidents in study area in history, which cost 24.9million USD (nominal). For oil spill, there were 6 accidents happened in study area from 1995 to 2005. The average volume of spilling oil is 305.5 ton per accident, average lost is 0.76 million USD (nominal). The probability of oil spill in Xiamen Bay has now reached 0.55 per year.  
  From 1969 to 2003, 171 typhoons affected Xiamen. The periodicity of these typhoons (Figure 4) is being studied for another paper; nonetheless, averaging is appropriate because of an apparent stationarity in the time series of typhoons. Typhoons generally occur in the period July to September. Moreover, the tidal storm caused by typhoon no. 9914 in 1999 which hit Xiamen directly caused 72 deaths and 5 billion USD of damages. It also changed the configuration and morphology of the southeast coast of Xiamen.
  According to the data from Xiamen Environmental Monitoring Station, from 1987 to 1995, there was no large red tide in study area, but the trend is increasing since 1995. In the last decades, there were 30 red tides in study area; these red tides usually occur from May to August.
   
  3.3 Prospective Risk Assessment (PRA)
  The results of prospective risk assessment, based on the RRA and combination of probabilistic analysis of rare events as well as routine events via probabilistic methods, showed that:
   
  1.        Study area is at high risk of ship accident, the average probability of ship accident is 1.32×10-4 per year per ship. The trend of ship accidents in Xiamen Bay is increasing.
  2.        The docks in Western Bay and Estuary Bay are at high risk of oil spill, the probability is 2.7×10-2 per dock per year.
  3.        The probability of oil spill accident in Xiamen Bay is 0.55 per year, the average leaking volume of oil is 305.5 tons. Each oil spill accident cost 0.76 million USD (nominal), which is much higher than the ship non-spill accident (0.082 million USD, nominal).
  4.        The highest risky areas of ship accidents and oil spills are Western Bay and Estuary Bay due to rapid shipping increasing.
  5.        Environmental engineering structure and the shores of Xiamen are threatened by storm tide, which have caused serious damage in the past.
  6.        The trend of red tides is increasing, and is mainly affecting the Western Bay.
   
  From a probabilistic hazard analysis, for 9 selected engineering projects in the Bays of Xiamen, the maximum and minimum increases in risk are given in Table 1, based on the 0.05 and 0.95 lower and upper probability limits, relative to the class of events considered. 
  Table 1, Maxima and minima changes in the probability of damage, per project

Projects

1

2

3

4

5

6

7

8

9

Pmax×10-3

0.355

1.87

1.90

2.48

2.31

2.31

2.33

2.34

2.36

Area

Estuary Bay

Western Bay

Western Bay

Western Bay

Western Bay

Western Bay

Western Bay

Western Bay

Western Bay

Pmin×10-3

0.038

0.014

0.017

0.017

0.021

0.014

0.041

0.055

0.052

Area

Tong’an Bay

Tong’an Bay

Tong’an Bay

Tong’an Bay

Tong’an Bay

Tong’an Bay

Estuary Bay

Estuary Bay

Estuary Bay

   
  The conclusions in the case study are as follows:
   
  1.      In Western Bay, project 1 is feasible; considering the project 4 will be implemented in the mouth of Western Bay, it will cause the significant changes in the hydrodynamic conditions in Western Bay and  increase the risk of oil spills, it will not be advisable. Projects 2 and project 5 should be considered and studied in greater depth.
  2.      Tong’an Bay will be the next hot spot for development in Xiamen Bay. The feasible projects are projects 1, 2, and 6. Other projects should be given more detailed attention and study before any further decision is taken.
  3.      Project 1 in Estuary Bay will cause the highest risk and thus should be studied further before any decision is taken to implement it.


  4. Conclusions
   
  Presently, most of frameworks for ERA focus on single projects or certain chemicals. It is obvious that for coastal zones, multiple interfaces, and very diverse physical, ecological and chemical process, are  extremely complex. The impacts of CEP are accumulative and far-reaching; existing frameworks for ERA are limited to carry out for CEP. To augment the capabilities of these processes, the system of ERA for CEP has been set up through a probabilistic framework in this study.
   
  This paper introduced RRA into the basic framework of the ERA for CEP. Based on the study on cumulative impacts of CEP and calculation of probabilities and consequences, RRA can guide the operation of risk identification and prospective risk assessment for CEP. The quantitative results of RRA can also be used for additional prospective quantitative analysis and risk management.
   
  References
   
  Balas C. E., Balas L. & Williams A. T. (2004). Risk Assessment of Revetments by Monte Carlo Simulation. Proceedings of the Institution of Civil Engineers-Maritime Engineering, 157(2): 61-70.
  Bedford T. & Cooke R. (2001). Probabilistic Risk Analysis: Foundations and methods. Cambridge University Press, Cambridge.
  Bridges J. (2003). Human Health and Environmental Risk Assessment: the Need for a More Harmonised and Integrated Approach. Chemosphere, 52: 1347-1351.
  Contini S. & Servida A. (1992). Risk Analysis in Environmental Impact Studies. Environmental Impact Assessment, 7: 79-103.
  ECO FRAM (1999). Ecological Committee on FIFRA Risk Assessment Methods Aquatic Report. Washington, DC, USA.
  Gelman A., Carlin J. B., Stern H. S. & Rubin D. B. (1995). Bayesian Data Analysis. Chapman & Hall, London.
  Grimmett G. R. & Stirzaker D. R. (1982). Probability and Random Processes. Oxford University Press, Oxford.
  Hong Y. & Apostolakis G. (1993). Conditional Influence Diagrams in Risk Management. Risk Analysis, 13:623-627.
  Pearl J. (2000). Causality: Models, reasoning and inference. Cambridge University Press, NY.
  Power M. & McCarty L. S. (2002). Trends in the Development of Ecological Risk Assessment and Management Frameworks. Human and Ecological Risk Assessment, 8(1): 7-18.
  Ricci P. F. (2006). Environmental and Health Risk Assessment and Management: Principles and Practices.  Environmental Pollution, vol. 9, Springer, Dordrecht, The Netherlands.
  Sergeant A. (2000). Management Objectives for Ecological Risk Assessment Developments at US EPA. Environmental Science & Policy, 3: 295-298.
   

 
污染防治与管理更多>>
循环经济与绿色产业发展 更多>>
低碳经济与可持续发展更多>>
中国面临的主要环境问题及对策 更多>>