A Comprehensive Method for Evaluating Marine Disaster Risk Reduction Capacity in China

A region’s capacity for marine disaster risk reduction is characterized by the resources that can be mobilized. These resources include pre-disaster defense, disaster monitoring, warning, emergency response, post-disaster restoration, and reconstruction. It is a very important index to effectively evaluate the regional capacity to overcome marine disasters. At present, there is no unified model and method for comprehensively evaluating the regional marine disaster reduction capacity. This study proposes a novel evaluation index system for a county-level administrative region using expert opinions, questionnaires, and analytic hierarchy process methods. Based on the comprehensive evaluation in three pilot areas, the current situation of regional marine disaster reduction capacity is analyzed, which would contribute to the effective management of marine disaster risks in the future. The results and experiences are of great value to future disaster reduction capacity assessment promotion and practice in all coastal counties of China.


Introduction
Different kinds of marine disasters are reported in China [1]. The frequency of disasters caused by super-typhoons, storm surges, and red tides increased significantly since 2000 [2]. With the rapid development of marine economy, maritime disasters often result in casualties and considerable economic losses. It is worth pointing out that the annual total economic loss of various marine disasters is still as much as 10 billion renminbi (RMB), for which the total economic loss reached its peak in 2005, about 33.2 billion RMB [3]. To reduce the risk of marine disasters, the marine disaster risk reduction (DRR) capacity needs to be improved. From Ten Years of International Disaster Reduction [4] to the Hyogo Framework for Action [5], and then to the Sendai Framework for Reducing Disaster Risk [6], the cognition of reducing disaster risk and DRR capacity is increasing continuously. It can be commenced from the initial proposition of developing and strengthening the disaster prevention and reduction capacity, followed by promoting the establishment of a disaster reduction system in the community, improving the disaster resistance and preparedness capacity, and eventually enhancing the disaster reconversion and reduction capacity.
At present, there is certain research on evaluating the regional DRR capacity, and several evaluation methods and processes were developed and applied. For example, the United Nations Development Program (UNDP) [7] evaluates the national DRR capacity in five technical areas (i.e., ownership, The objectives of this study are (1) to build an index system of regional marine DRR capacity for administration areas at the county level, and (2) to develop a quantitative method for assessing the marine DRR capacity and apply it to the pilot assessment. The results of this study should have important theoretical and practical implications for guiding government agencies to improve their DRR capacity building. They may also provide a good foundation for future assessments of marine disaster mitigation capacity across the whole country.

Study Area and Data Source
In a comprehensive evaluation for the regional marine DRR capacity (2015-2016) project [25], 11 coastal counties located in Weifang city of Shandong province, Taizhou city of Zhejiang province, and Huizhou city of Guangdong province were chosen to implement a pilot evaluation (Figure 1). In this study, an index system was established using multilateral methods. Experts with various professional backgrounds, coming from different central and local industrial sectors (related to marine DRR capacity), were invited by the Marine Disaster Reduction Center of State Oceanic Administration to participate in designing a reliable index system. The experts' professional backgrounds involved maritime, engineering, emergency, publicity and education, comprehensive disaster reduction, etc.
Sustainability 2020, 12, x FOR PEER REVIEW 3 of 22 The objectives of this study are (1) to build an index system of regional marine DRR capacity for administration areas at the county level, and (2) to develop a quantitative method for assessing the marine DRR capacity and apply it to the pilot assessment. The results of this study should have important theoretical and practical implications for guiding government agencies to improve their DRR capacity building. They may also provide a good foundation for future assessments of marine disaster mitigation capacity across the whole country.

Study Area and Data Source
In a comprehensive evaluation for the regional marine DRR capacity (2015-2016) project [25], 11 coastal counties located in Weifang city of Shandong province, Taizhou city of Zhejiang province, and Huizhou city of Guangdong province were chosen to implement a pilot evaluation (Figure 1). In this study, an index system was established using multilateral methods. Experts with various professional backgrounds, coming from different central and local industrial sectors (related to marine DRR capacity), were invited by the Marine Disaster Reduction Center of State Oceanic Administration to participate in designing a reliable index system. The experts' professional backgrounds involved maritime, engineering, emergency, publicity and education, comprehensive disaster reduction, etc.   After extensive discussions, a three-grade index system (including six first-grade indices, 21 second-grade indices, and 136 third-grade indices) was finally adopted.
The raw data for calculating the weight values in the index system were derived from the experts' experience. Firstly, an expert questionnaire for county-level marine DRR capacity was designed, and each questionnaire included 30 marking tables (one for the first-grade indices, six for the second-grade indices, and 23 for the third-grade indices). Then, the questionnaires were distributed among 86 experts in 18 units, and eventually 69 valid questionnaires were collected with 2070 valid marking tables. To fully benefit from the experts' experience in all fields, the questionnaires were distributed among experts in hydrology, meteorology, emergency, and comprehensive DRR fields, in addition to the experts from 16 units in the working group. The affiliations of additional experts included China Institute of Water Resources and Hydropower Research, National Climate Center, Institute of Geographical Sciences and Natural Resource Research, Chinese Academy of Sciences, etc.

Development of an Evaluation Index System
Disaster risk assessment is an integral part of disaster management. When designing the DRR capacity assessment indicator system, the location of DRR capacity assessment in disaster management and the relationship with other disaster management measures need to be clarified before, during, and after the disasters, before fully considering their purpose, content, and application areas.

Management Phase and Capacity Assessment
Disaster mitigation capacity assessment and capacity building research should be based on a more complete disaster risk analysis. By analyzing the types, intensity/frequency, and spatial differentiation of natural disaster risks, combined with the capacity assessment results, the capacity building gap analysis can be conducted. After that, the needs analysis is set, the regional DRR targets are set at a certain stage, and finally specific capacity building projects can be implemented.
The management contents include disaster preparedness, forecasting, and early warning. Management in disaster mainly includes the emergency response, while that during the post-disaster stage mainly includes restoration and reconstruction. Disaster publicity and education run through all stages of disaster management. From the disaster prevention and mitigation perspective, disaster reduction measures are often divided into engineering measures and non-engineering measures. The former measures emphasize the construction of hardware facilities, while the latter measures include various soft measures outside the construction of hardware facilities. Therefore, when designing indicators, the current study fully considered the disaster management cycle and the means of disaster prevention and mitigation.

Capacity Needs and Priority Domains
The demand for disaster risk reduction capacity is often related to the magnitude of disaster risk. A greater risk suggests a greater need for DRR capacity building. However, human beings' pursuit of risk mitigation could be endless. Safety standards gradually rise with socioeconomic development. Hence, regional DRR capacity building targets should be rationally set in accordance with the regional risk levels and the region's safe standards. To reduce the risk of disasters and improve the level of safety, DRR capacity building requires the support of DRR resources and means.

Assessment Object and Construction Subject
The main bodies involved in the assessment of disaster mitigation capacity are diverse. From the administrative perspective, there are national, provincial, city, county, and township (town) levels. From the social perspective, in addition to government departments, communities (villages), families, units, enterprises, and individuals are the main bodies of DRR. The capacity building of marine DRR should form an efficient multi-agent capacity building system. Moreover, the differences in capacity building objectives of different entities and departments should be considered. To effectively allocate resources, the relationship with DRR capacities of relevant departments (such as meteorology, water conservancy, earthquake, civil affairs, and agriculture) should be coordinated and handled.

Regional Commonality and Specificity of Indicators
To build a regional DRR capacity indicator system, we should consider the commonality and the regional individuality of indicators. For example, sea ice disasters generally occur in northern China. From a historical perspective, there is a need for sea ice disaster reduction. Sea ice-prone areas naturally form strong sea ice disaster mitigation capabilities, while southern coastal areas do not need the capacity to cope with sea ice disasters. Another example is the frequent tropical storm surges in southern China, which enabled the southern region to gain rich experience in dealing with tropical cyclones, while northern China has a relatively weak coping capacity.
The indicator system constructed here considered the different marine disaster risks when applying it in different pilot areas. The differences in regional indicators were considered. Through pilot research, the indicator system can be applied to the national assessment of marine disaster mitigation capacity; thus, the indicator system should be universal in China's coastal areas.
Each item in the established evaluation indices should be well documented and easily quantified, and each one can truly reflect the main characteristics and actual situation of regional marine DRR capacity in China. Meanwhile, the evaluation indices should be connected with the current statistical indices of Civil Affairs and Marine Management Departments in China, and they should be consistent to facilitate their work, measurement, and calculation. Maneuverability is also expressed so that decision-makers can further improve the regional marine DRR capacity using simple and reliable results. Since the study is still in its infancy, and statistical data for regional marine DRR capacities are scattered and incomplete, the available data were used to the utmost extent to extract the relevant information.
Based on scientific aspects, accessibility, typicality, and practicability principles for a comprehensive evaluation index of DRR capacity were established, as well as combining the literature's results with experts' and scholars' conclusions [26][27][28], thereby achieving a classified, hierarchical, and comprehensive evaluation index system for marine DRR capacity. It was developed through the selection and screening of various methods after the relevant units solicited opinions and suggestions from various experts and managers (who act on maritime disaster management at the primary level) ( Table 1).
Disaster warning and forecasting capacity includes (1) monitoring capacity (i.e., the capacity of monitoring the occurrence, development process, and dynamic changes of various related factors in marine disasters), (2) warning and forecasting capacity (i.e., the capacity of forecasting and warning the time, location, intensity, and range of impacts of possible marine disasters), and (3) information publishing capacity (i.e., the capacity of government departments, sea-related enterprises and personnel, and the public receiving pre-alarm information through television, radio, internet, telephone, fax, newspapers, etc.).
Disaster preparedness capacity includes (1) disaster avoidance capacity (i.e., the capacity of an area to avoid the loss of people and assets from storms, waves, tsunamis, and other marine disasters, generally including disaster prevention awareness and disaster avoidance facilities), (2) material supporting capacity for disaster reduction and relief (i.e., covering the size of the disaster relief material reserve, the number of disaster relief materials, and the amount of relief material procurement funds), and (3) supporting capacity of social economic base (i.e., an important criterion for measuring the pre-disaster response capacity and post-disaster recovery capacity of a region).

Establishing the Evaluation Model
Herein, we made an attempt to consider that the regional marine DRR capacity is the comprehensive ability of the region in the whole process of disaster management including disaster prevention and reduction, emergency management, rescue and recovery, and post-disaster restoration and reconstruction. Therefore, when a comprehensive evaluation method was developed for regional marine DRR capacity, the importance of various capacities and the marine DRR performance of various regions were combined to eventually conduct a comprehensive evaluation.
When conducting a comprehensive evaluation, the corresponding score C i ' was determined according to certain criteria. Then, the results of comprehensive evaluation of regional marine DRR capacity for each region E was calculated by multiplying the assignment value of each index and the corresponding weights (W i ), and summing them up. A comprehensive evaluation model for regional marine DRR capacity is given by where E denotes the comprehensive evaluation value for regional marine DRR capacity, C i is a coefficient for quantitative scoring of each evaluation index, and W i represents the weight of each evaluation index, where i = 1, 2, . . . , n j . Here, j represents the third grade which is the lowest grade in the evaluation index system, and n j denotes the number of third-grade indices.
The main characteristic of the analytic hierarchy process (AHP) decision-making method [29,30] is its simplicity of decision-making with respect to complex problems with multiple objective functions and multiple criteria. The AHP builds a hierarchy (ranking) of decision items using comparisons between each pair of items expressed as a matrix. The AHP method is particularly appropriate for situations where decision results cannot be directly and accurately measured [31,32].
The basic steps for conducting a comprehensive evaluation of regional marine DRR capacity based on the AHP method are described below.

Establishing a Hierarchical Structure Model
The efficient factors are decomposed into multiple grades from top to bottom according to their attributes. Thus, factors in the same grade may be affected by a factor at a higher grade, while factors in the same grade are dominant or may be affected by factors in the next grade. The top grade is taken as the target grade, typically involving only one factor. The bottom grade is usually the scheme or object grade. The middle grades may involve one or several grades, and they frequently represent the criteria or grades for index. When there are lots of criteria (e.g., more than nine criteria), the grades should be further decomposed into secondary grades. Table 1. The integrated evaluation index system for marine disaster risk reduction (DRR) capacity.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Engineering defense capacity A1 Public engineering defense capacity B1 The ratio of seawall meeting the protection requirement C1 The ratio of standard seawall C2 The proportion of protected people to seawall C3 The reciprocal of the vulnerable point density of seawall C4 The pump station density of seawall C5 The drainage capacity of drainage facilities C6 The management level of seawall C7 The frequency of patrolling seawall during disasters C8 The ratio of high standard seawall C9 The length of 200-year return period of seawall C10 The length of 100-year return period of seawall C11 The length of 50-year return period of seawall C12 The reciprocal of the density of pipes into the sea C13 The defense capacity of important disaster-bearing bodies B2 The ratio of the area of culture zone protected by the standard seawall C14 The ratio of the gross product of culture zone protected by the standard seawall C15 The proportion of protected fishing boats C16 The proportion of fishing ports above level 2 C17 The number of fishing boats which can withstand wind above grade 7 C18 The proportion of ports with ice-breaking capacity (only suitable for northern ports) C19 Whether to carry out post-evaluation work regarding the protection capacity of ports C20 The proportion of owned oil platforms with sea wave and sea ice warning guarantee information C21 The proportion of platforms with emergency plan (considering the sea ice and sea waves in north and south divisions separately) C22 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
The proportion of bulwarks with a designing return period of tidal level and waves above 50 years C23 Whether the petrochemical zone formulated a marine disaster emergency plan C24 The design of tidal level return period of nuclear power plants C25 The design of wave height return period of nuclear power plants C26 The design of maximum prevention tsunami waves of nuclear power plants C27 The target achievement rate of wind resistance capacity C28 The target achievement rate of the anti-ice capacity of bridge pier C29 The ratio of boats which can withstand wind above grade 8 C30 The ratio of large coastal projects carrying out a special evaluation of marine disaster risk C31 The setting ratio of indicators (warning signs) C32 The proportion of professional life-saving equipment C33 The setting proportion of professional rescue teams C34 The set number of emergency plans C35 The number of emergency exercises C36 Natural protection capacity B3 The area of protected forest (e.g., mangrove forest, Chinese tamarisk forest) per unit shoreline C37 The area of salt marsh per unit shoreline C38 The area of coral reef per unit shoreline C39 The proportion of inhabited island C40 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Disaster warning and forecasting capacity A2 Monitoring capacity B4 The sea wave observation point density C41 The water level observation point density C42 The water temperature observation point density C43 The sea condition video monitoring capacity C44 The usage capacity of satellite monitoring systems C45 The mobile monitoring capacity C46 The aerial monitoring capacity C47 The tsunami monitoring capacity C48 The team size of observers C49 Warning and forecasting capacity B5 The number of marine forecasting organizations C50 The team size of marine forecasters C51 The marine disaster warning capacity C52 The marine environmental forecasting capacity C53 The computing power of high-performance computers C54 The operational numerical value forecasting capacity C55 The receiving capacity of pre-warning products C56 The information publishing capacity B6 The publishing channels C57 The maximum publishing frequency C58 The publishing system of forecasting information C59 The short message platform C60 The number of outdoor display screens in the evaluation region C61 The number of fishing boats in the evaluation area equipped with safety environment guarantee service systems during fishery production C62 The number of mobile phone applications on disaster warning and forecasting in the evaluation region C63 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Disaster emergency response capacity A3 Emergency command capacity B7 Whether a marine disaster emergency command organization is established C64 The department of marine disaster emergency command organization C65 The administrative position of the general command of the marine disaster emergency command organization C66 The type of emergency command communication system C67 The type of emergency communication system under extreme disaster conditions C68 Emergency response plan system B8 The ratio of villages (communities) preparing the marine disaster emergency response plan in coastal towns (streets) C69 The name and number of technical emergency response plans or operation manuals C70 The number of emergency exercises for marine disasters last year C71 The support capacity for assisting decision B9 Whether there is an expert technical supporting group for assisting decisions C72 Whether to establish or use the supporting platform for assisting decisions C73 Whether to establish or use a consultation system; if yes, please fill in the communication mode of the consultation system C74 Whether there is a team of disaster information staff; if yes, please check the administrative division level of the established disaster information staff C75 Whether to carry out risk evaluation work for marine disasters; if yes, please fill in the evaluated disaster type and scale C76 Whether to establish a database of geographic information systems of the disaster-bearing system C77 Whether to carry out verification work of warning tide level; if yes, please fill in the number of shorelines verified the warning tide level in the evaluation area C78 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Aiding and rescuing capacity B10 Whether there is a professional marine rescue team C79 The number of professional rescue teams C80 The expenditure of professional rescue teams last year C81 Whether there are rescue helicopters and ships C82 The number of rescue helicopters per unit length of coastline C83 The number of rescue ships per unit length of coastline C84 Post-disaster recovery and repair capacity A4 Post-disaster emergency repair guarantee capacity B11 Post-disaster electric power guarantee capacity C85 Post-disaster water supply guarantee capacity C86 The government compensation capacity C93 The mutual protection capacity of fishing boats C94 The mutual protection capacity of cultivation C95 The insurance rate C96 The hardware construction of marine disaster reduction publicity and education B17 Whether bulletin boards for disaster prevention and reduction publicity are set up in the neighborhood committee, community, and shelter C97 Whether bulletin boards for disaster prevention and reduction publicity are set up in other places (such as roads, squares, shopping malls, outdoor advertising columns) C98 Whether a high-tone loudspeaker (alarm) is placed in the evaluated coastal community (village) C99 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Whether a warning tide level mark is set up in the evaluation area C100 Publicity and education capacity A5 Whether a marine disaster warning sign is set up in the evaluation area C101 The number of disaster prevention and reduction publicity and education bases for popularizing related knowledge of marine disaster prevention and reduction C102 Publicity and education capacity of marine disaster prevention and reduction B18 The capital amount the government used for the publicity and education of disaster prevention and reduction every year C103 The capital amount the government used for the publicity and education of marine disaster prevention and reduction every year C104 Whether there is a publicity and education plan on marine disaster prevention and reduction in the evaluation area C105 Whether a uniform publicity plan on marine disaster prevention and reduction is set up in the evaluation area every year C106 The number of specialized organizations working on the publicity and education of marine disaster prevention and reduction knowledge in the evaluation area C107 The number of bulletin boards in the neighborhood committee, community, and shelter used for publicizing disaster prevention and reduction C108 The number of publicity activities for marine disaster prevention and reduction held in the evaluation area every year C109 The number of people attending marine disaster prevention and reduction training in the evaluation area every year C110 The number of communities (villages) holding at least one publicity activity about marine disaster prevention and reduction every year C111 The situation of compiling or using marine disaster prevention and mitigation publicity materials in the evaluation area C112 The people working on the publicity and education of marine disaster prevention and reduction in the evaluation area C113 Whether the marine disaster prevention and reduction knowledge is included in the middle school and primary school curriculum C114 Table 1. Cont.

First-Grade Indices Second-Grade Indices Third-Grade Indices
Whether television and other media broadcast popular scientific knowledge of marine disaster prevention in the evaluation area C115 The number of televisions and other media broadcasting popular scientific knowledge of marine disaster prevention in the evaluation area C116 Disaster preparedness capacity A6 The disaster avoidance capacity B19 The number of emergency shelters per unit area C117 The proportion of the emergency capacity C118 The proportion of shelters suitable for marine disaster C119 The number of havens per unit coastline C120 The number of shelter anchor grounds per unit coastline C121 The number of facilities that can bear level 12 wind waves C122 The number of sheltered boats C123 The material supporting capacity for disaster reduction and relief B20 The number of disaster relief material storage points C124 The area of disaster relief material storage points C125 The types of reserve materials C126 The quantity of reserve materials C127 Whether there is a disaster preparedness agreement with local supermarkets C128 The disaster relief material purchasing fund every year C129 The supporting capacity of social economic base B21 Fiscal revenue C130 The traffic network density C131 The renminbi (RMB) deposit remaining C132 Mobile phone ownership per head C133 The proportion of fixed-line telephones C134 The number of healthy technical workers per thousand people C135 The number of medical beds owned per thousand people C136

Constructing Judgment Matrix
In the second grade of the hierarchical structure model, because of the existence of several factors at the same grade which are subjected to or may be affected by each factor in the upper grade, a judgment matrix is constructed by the paired comparison method extended to the bottom grade.
In this study, six first-grade indices were adopted to establish the judgment matrix, and each index was scaled and assigned according to its relative importance to the target grade using a 1-9 scaling method, and a 6 × 6 judgment matrix was constructed. For the second-grade indices, six judgment matrices were established using the same method and principle, and the dimensions of each matrix could be determined from the number of second-grade indices at the lower grade. In addition, for each second-grade index, 23 judgment matrices were established by comparing the relative importance of each third-grade index at the next grade. Since the second-grade indices for defense capacity of typical disaster-bearing bodies and the publicity and education management capacity of marine disaster prevention and reduction contained 24 and 14 matrices, respectively, the two judgment matrices were broken down into 11 × 11 and 12 × 12 judgment matrices and 6 × 6 and 8 × 8 judgment matrices in order to avoid the effect of superabundant matrix elements in the judgment of experts.
When comparing the importance of i elements and j elements relative to an element in the upper layer, a quantitative relative weight a ij is used. When assuming that n elements participate in the comparison, then A = (a ij ) n × n is called a pairwise comparison matrix. For the criteria of the judgment matrix, the nature and importance of the elements being compared were assigned a value from 1-9 as follows: a ij = 1, i element and j element have the same importance as the upper level of factors; a ij = 3, i element is slightly more important than j element; a ij = 5, i element is more important than j element; a ij = 7, i element is much more important than j element; a ij = 9, i element is extremely more important than j element; a ij = 2n, n = 1, 2, 3, 4, the importance of i element and j element is between a ij = 2n − 1 and aij = 2n + 1; a ij = 1/a ji , n = 1, 2, . . . , 9, if and only if a ji = n. The characteristics of the pairwise comparison matrix were as follows: a ij > 0, aij = 1, a ij = 1/a ji (note that, when i = j, a ij = 1).
The six first-level indicators were used to establish a judgment matrix. That is, the relative importance of each indicator relative to the target layer (integrated marine disaster reduction capacity), according to the 1-9 scale method, was used to create a 6 × 6 evaluation matrix PairMatrix m,1,1 , as shown in Table 2. The six first-grade indicators were expressed as follows: A1: Engineering defense capacity; A2: Disaster warning and forecasting capacity; A3: Disaster emergency response capacity; A4: Post-disaster recovery and repair capacity; A5: Publicity and education capacity; A6: Disaster preparedness capacity. For the second-grade indices, six judgment matrices were established using the same method and principle, and the dimensions of each matrix were determined by the number of second-grade indices at the lower grade. For example, for the first-grade indicator A1 (engineering defense capacity), the second-grade indicators of the next level (Table 3) were as follows: B1: Public engineering defense capacity; B2: The defense capacity of important disaster-bearing bodies; B3: Natural protection capacity.
The pairwise comparison matrices generated under the different hierarchical relationships were all recorded as PairMatrix m,i,j (Equation (2)).
where m = 1, 2, . . . , k, where k is the number of participating experts, and i is the number of hierarchical structures. In this study, i = 1, 2, 3, 4, 5, 6. The scheme layer was j = 1, 7, . . . , n i , where n i is the number of evaluation indices of the i layer. x k,i,j represents the relative importance values of the k expert for index a and index b of the i + 1 layer in the j sub-hierarchy of the i level with respect to its superior (i level) index. Assuming that the filled judgment matrix was PairMatrix m, i,j = x k,i,j (a, b) n i × n i , the elements of each matrix were simplified to x i,j . The judgment matrix had the following properties:

Calculating Weight Vectors in Hierarchical Single Arrangement, as well as Conducting the Consistency Test
Compared with other methods for determining the index weight coefficient, the most significant advantage of the AHP method is performing a consistency test to ensure the experts' thoughts and logical coherence. For each comparative matrix pair, their maximum eigenvalue and corresponding eigenvector are calculated, and a consistency test is conducted using the consistency index, random consistency index, and consistency ratio. If the test is passed, the normalized eigenvector is the weight vector; otherwise, it is required to reconstruct the comparative matrix pairs.

Calculating the Weight Vector in Total Arrangement, in Addition to Conducting the Consistency Test
The combined weight vector for the bottom grade to the target grade is calculated, and consistency checking is conducted according to the mathematical relationships. If this test is passed, a decision can be made according to the results of the combined weight vector. Otherwise, the model should be regenerated or the comparative matrix pairs with a large consistency ratio should be reconstructed. Among the 2070 marking tables collected in this work, 1752 marking tables (about 84.6%) passed the consistency check. The box plots of the weights of the first-and second-grade indicators were completed, as shown in Figure 2.
where m = 1, 2, …, k, where k is the number of participating experts, and i is the number of hierarchical structures. In this study, i = 1, 2, 3, 4, 5, 6. The scheme layer was j = 1, 7, …, ni, where ni is the number of evaluation indices of the i layer. xk,i,j represents the relative importance values of the k expert for index a and index b of the i + 1 layer in the j sub-hierarchy of the i level with respect to its superior (i level) index. Assuming that the filled judgment matrix was PairMatrixm, i,j = xk,i,j (a, b) ni × ni, the elements of each matrix were simplified to xi,j. The judgment matrix had the following properties: xi,j > 0; xj,i = 1/xi,j; xi,i = 1.

Calculating Weight Vectors in Hierarchical Single Arrangement, as well as Conducting the Consistency Test
Compared with other methods for determining the index weight coefficient, the most significant advantage of the AHP method is performing a consistency test to ensure the experts' thoughts and logical coherence. For each comparative matrix pair, their maximum eigenvalue and corresponding eigenvector are calculated, and a consistency test is conducted using the consistency index, random consistency index, and consistency ratio. If the test is passed, the normalized eigenvector is the weight vector; otherwise, it is required to reconstruct the comparative matrix pairs.

Calculating the Weight Vector in Total Arrangement, in Addition to Conducting the Consistency Test
The combined weight vector for the bottom grade to the target grade is calculated, and consistency checking is conducted according to the mathematical relationships. If this test is passed, a decision can be made according to the results of the combined weight vector. Otherwise, the model should be regenerated or the comparative matrix pairs with a large consistency ratio should be reconstructed. Among the 2070 marking tables collected in this work, 1752 marking tables (about 84.6%) passed the consistency check. The box plots of the weights of the first-and second-grade indicators were completed, as shown in Figure 2.

Determining the Basis for Grade Rating
Different methods were used in different studies for determining the basis of each index. In this paper, the relative importance of each index was considered, which required ranking the index grade

Determining the Basis for Grade Rating
Different methods were used in different studies for determining the basis of each index. In this paper, the relative importance of each index was considered, which required ranking the index grade in different regions. Meanwhile, the collected data were easily translated into evaluation scores to assess the DRR capacity for each region. The grade was divided into poor, medium, and excellent grades using the interval method, taking the grade rating threshold offered by pilot units as a reference. Then, the value of importance of ranking was assigned to each index in order to determine the score. The value of importance of ranking was assigned according to the score of the corresponding grade. After that, the actual data were converted into comparable score values. A standardized index was used to normalize all values of the third-grade index in the range 0-1. Based on histogram analysis, the evaluation indices for the first-grade index for regional marine DRR capacity were divided into poor, medium, and excellent grades with scores of 0.45, 0.45-0.60, and 0.60, respectively.

Evaluation Results for Pilot Regions
According to the proposed comprehensive evaluation index for regional marine DRR capacity, the degree of importance of each index was analyzed, the index weight was determined, and then the regional marine DRR capacity was calculated. The calculated index was comparatively analyzed with the actual situation, and the reliability, rationality, and feasibility of the index system were verified, in addition to the established model and the proposed methodology for evaluating regional marine DRR capacity.

Evaluation Results for Pilot Regions in Guangdong Province
Two pilot regions, Dayawan and Huidong, were selected for evaluation and analysis in Guangdong province (Figure 3). In general, in terms of the evaluation results for the regional marine DRR capacity, Dayawan (0.6139) obtained an excellent grade and was slightly superior to Huidong (0.5651), which obtained a medium grade. Furthermore, in the Dayawan region, the disaster emergency response capacity and disaster preparedness capacity obtained an excellent grade, while the engineering defense capacity, disaster warning capacity, post-disaster recovery and repair capacity, and publicity and education capacity obtained a medium grade. On the other hand, in the Huidong region, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the publicity and education capacity was ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor.
Sustainability 2020, 12, x FOR PEER REVIEW 15 of 22 reference. Then, the value of importance of ranking was assigned to each index in order to determine the score. The value of importance of ranking was assigned according to the score of the corresponding grade. After that, the actual data were converted into comparable score values. A standardized index was used to normalize all values of the third-grade index in the range 0-1. Based on histogram analysis, the evaluation indices for the first-grade index for regional marine DRR capacity were divided into poor, medium, and excellent grades with scores of 0.45, 0.45-0.60, and 0.60, respectively.

Evaluation Results for Pilot Regions
According to the proposed comprehensive evaluation index for regional marine DRR capacity, the degree of importance of each index was analyzed, the index weight was determined, and then the regional marine DRR capacity was calculated. The calculated index was comparatively analyzed with the actual situation, and the reliability, rationality, and feasibility of the index system were verified, in addition to the established model and the proposed methodology for evaluating regional marine DRR capacity.

Evaluation Results for Pilot Regions in Guangdong Province
Two pilot regions, Dayawan and Huidong, were selected for evaluation and analysis in Guangdong province (Figure 3). In general, in terms of the evaluation results for the regional marine DRR capacity, Dayawan (0.6139) obtained an excellent grade and was slightly superior to Huidong (0.5651), which obtained a medium grade. Furthermore, in the Dayawan region, the disaster emergency response capacity and disaster preparedness capacity obtained an excellent grade, while the engineering defense capacity, disaster warning capacity, post-disaster recovery and repair capacity, and publicity and education capacity obtained a medium grade. On the other hand, in the Huidong region, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the publicity and education capacity was ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor.

Evaluation Results for Pilot Regions in Shandong Province
Three pilot regions in Shandong province, including Binhai district, Changyi district, and Shouguang city, were selected for evaluating the achieved results (Figure 4). In the figure, A1 stands for engineering defense capacity, A2 stands for disaster warning and forecasting capacity, A3 stands for disaster emergency response capacity, A4 stands for post-disaster recovery and repair capacity, A5 stands for publicity and education capacity, and A6 stands for disaster preparedness capacity. The meanings of A1-A5 are the same in Figures 2 and 3.

Evaluation Results for Pilot Regions in Shandong Province
Three pilot regions in Shandong province, including Binhai district, Changyi district, and Shouguang city, were selected for evaluating the achieved results ( Figure 4).
In general, in terms of a comprehensive evaluation index for regional marine DRR capacity index, the index values for the three pilot regions were all near 0.6, and their difference was negligible. Shouguang city (0.6356) and Binhai district (0.6172) were ranked as excellent, while Changyi district was ranked as medium. In Binhai district, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium grade, and the post-disaster recovery and repair capacity was ranked as poor. Similarly, in Changyi district, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium grade, and the post-disaster recovery and repair capacity was ranked as poor. However, in Shouguang city, all capacities obtained an excellent grade, except for post-disaster recovery and repair capacity, which was poor.
Sustainability 2020, 12, x FOR PEER REVIEW 16 of 22 In general, in terms of a comprehensive evaluation index for regional marine DRR capacity index, the index values for the three pilot regions were all near 0.6, and their difference was negligible. Shouguang city (0.6356) and Binhai district (0.6172) were ranked as excellent, while Changyi district was ranked as medium. In Binhai district, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium grade, and the postdisaster recovery and repair capacity was ranked as poor. Similarly, in Changyi district, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium grade, and the post-disaster recovery and repair capacity was ranked as poor. However, in Shouguang city, all capacities obtained an excellent grade, except for post-disaster recovery and repair capacity, which was poor.

Evaluation Results for Pilot Regions in Zhejiang Province
Six pilot regions in Zhejiang province were selected and analyzed (Jiaojiang district, Linhai city, Luqiao district, Sanmen county, Wenling city, and Yuhuan county).
In Jiaojiang district, the engineering defense capacity, disaster warning capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the post-disaster recovery and repair capacity was ranked as poor. In Linhai city, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster warning capacity, disaster emergency response capacity, and publicity and education capacity were ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor. In Luqiao district, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster emergency response capacity and publicity and education capacity were ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor. In Sanmen County, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster emergency response capacity and publicity and education capacity were ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor. In Wenling city, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor. In Yuhuan County, the engineering defense capacity, disaster warning capacity, and

Evaluation Results for Pilot Regions in Zhejiang Province
Six pilot regions in Zhejiang province were selected and analyzed (Jiaojiang district, Linhai city, Luqiao district, Sanmen county, Wenling city, and Yuhuan county).
In Jiaojiang district, the engineering defense capacity, disaster warning capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the post-disaster recovery and repair capacity was ranked as poor. In Linhai city, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster warning capacity, disaster emergency response capacity, and publicity and education capacity were ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor. In Luqiao district, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster emergency response capacity and publicity and education capacity were ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor. In Sanmen County, the engineering defense capacity and disaster preparedness capacity were ranked as excellent, whereas the disaster emergency response capacity and publicity and education capacity were ranked as medium, and the disaster warning capacity and post-disaster recovery and repair capacity were ranked as poor. In Wenling city, the engineering defense capacity, disaster emergency response capacity, and disaster preparedness capacity were all ranked as excellent, whereas the disaster warning capacity and publicity and education capacity were ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor. In Yuhuan County, the engineering defense capacity, disaster warning capacity, and disaster emergency response capacity were all ranked as excellent, whereas the publicity and education capacity was ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor.
Sustainability 2020, 12, x FOR PEER REVIEW 17 of 22 disaster emergency response capacity were all ranked as excellent, whereas the publicity and education capacity was ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor.

Comparing the Evaluation Results
After comparatively analyzing the regional marine DRR capacity for 11 pilot regions, it was revealed that the values of comprehensive evaluation indices for regional marine DRR capacity were relatively close together ( Figure 6). Both the maximum and the minimum values belonged to Zhejiang province, where the Jiaojiang district achieved the maximum value, and Sanmen County had the minimum value. Furthermore, the evaluation results of regional marine DRR capacity for Guangdong Dayawan, Shandong Binhai, Shandong Shouguang, Zhejiang Jiaojiang, Zhejiang Wenling, and Zhejiang Yuhuan all obtained an excellent grade. The evaluation results of regional marine DRR capacity for Guangdong Huidong, Shandong Changyi, Zhejiang Linhai, Zhejiang Luqiao, and Zhejiang Sanmen all obtained a medium grade. Figure 6. Regional comparison of assessment results of comprehensive regional marine disaster risk reduction (DRR) capacity in pilot area. In the figure, R1 stands for Guangdong Dayawan, R2 stands for Guangdong Huidong, R3 stands for Shandong Binhai, R4 stands for Shandong Changyi, R5 stands for Shandong Shouguang, R6 stands for Zhejiang Jiaojiang, R7 stands for Zhejiang Linhai, R8 stands for Zhejiang Luqiao, R9 stands for Zhejiang Sanmen, R10 stands for Zhejiang Wenling, and R11 stands for Zhejiang Yuhuan. The meanings of R1-R11 are the same in Figure 7.
The regional comparison of assessment results of various marine DRR capacities is shown in Figure 7. It is disclosed that the evaluation values of engineering defense capacities for the 11 pilot regions were basically the same. The evaluation values of engineering defense capacity for the other 10 pilot regions were all higher than 0.6, except for Guangdong Dayawan (0.5589), which was below 0.6. Additionally, the evaluation value of engineering defense capacity for Shandong Binhai (0.7115) was higher than 0.7, and evaluation values were equal to 0.6928, 0.6850, 0.6791, 0.6496, 0.6483, 0.6458, 0.6293, 0.6285, and 0.6163 for Zhejiang Yuhuan, Zhejiang Wenling, Zhejiang Linhai, Zhejiang

Comparing the Evaluation Results
After comparatively analyzing the regional marine DRR capacity for 11 pilot regions, it was revealed that the values of comprehensive evaluation indices for regional marine DRR capacity were relatively close together ( Figure 6). Both the maximum and the minimum values belonged to Zhejiang province, where the Jiaojiang district achieved the maximum value, and Sanmen County had the minimum value. Furthermore, the evaluation results of regional marine DRR capacity for Guangdong Dayawan, Shandong Binhai, Shandong Shouguang, Zhejiang Jiaojiang, Zhejiang Wenling, and Zhejiang Yuhuan all obtained an excellent grade. The evaluation results of regional marine DRR capacity for Guangdong Huidong, Shandong Changyi, Zhejiang Linhai, Zhejiang Luqiao, and Zhejiang Sanmen all obtained a medium grade. disaster emergency response capacity were all ranked as excellent, whereas the publicity and education capacity was ranked as medium, and the post-disaster recovery and repair capacity was ranked as poor.

Comparing the Evaluation Results
After comparatively analyzing the regional marine DRR capacity for 11 pilot regions, it was revealed that the values of comprehensive evaluation indices for regional marine DRR capacity were relatively close together ( Figure 6). Both the maximum and the minimum values belonged to Zhejiang province, where the Jiaojiang district achieved the maximum value, and Sanmen County had the minimum value. Furthermore, the evaluation results of regional marine DRR capacity for Guangdong Dayawan, Shandong Binhai, Shandong Shouguang, Zhejiang Jiaojiang, Zhejiang Wenling, and Zhejiang Yuhuan all obtained an excellent grade. The evaluation results of regional marine DRR capacity for Guangdong Huidong, Shandong Changyi, Zhejiang Linhai, Zhejiang Luqiao, and Zhejiang Sanmen all obtained a medium grade. 6. Regional comparison of assessment results of comprehensive regional marine disaster risk reduction (DRR) capacity in pilot area. In the figure, R1 stands for Guangdong Dayawan, R2 stands for Guangdong Huidong, R3 stands for Shandong Binhai, R4 stands for Shandong Changyi, R5 stands for Shandong Shouguang, R6 stands for Zhejiang Jiaojiang, R7 stands for Zhejiang Linhai, R8 stands for Zhejiang Luqiao, R9 stands for Zhejiang Sanmen, R10 stands for Zhejiang Wenling, and R11 stands for Zhejiang Yuhuan. The meanings of R1-R11 are the same in Figure 7.
The regional comparison of assessment results of various marine DRR capacities is shown in Figure 7. It is disclosed that the evaluation values of engineering defense capacities for the 11 pilot regions were basically the same. The evaluation values of engineering defense capacity for the other 10 pilot regions were all higher than 0.6, except for Guangdong Dayawan (0.5589), which was below 0.6. Additionally, the evaluation value of engineering defense capacity for Shandong Binhai (0.7115) was higher than 0.7, and evaluation values were equal to 0.6928, 0.6850, 0.6791, 0.6496, 0.6483, 0.6458, 0.6293, 0.6285, and 0.6163 for Zhejiang Yuhuan, Zhejiang Wenling, Zhejiang Linhai, Zhejiang Figure 6. Regional comparison of assessment results of comprehensive regional marine disaster risk reduction (DRR) capacity in pilot area. In the figure, R1 stands for Guangdong Dayawan, R2 stands for Guangdong Huidong, R3 stands for Shandong Binhai, R4 stands for Shandong Changyi, R5 stands for Shandong Shouguang, R6 stands for Zhejiang Jiaojiang, R7 stands for Zhejiang Linhai, R8 stands for Zhejiang Luqiao, R9 stands for Zhejiang Sanmen, R10 stands for Zhejiang Wenling, and R11 stands for Zhejiang Yuhuan. The meanings of R1-R11 are the same in Figure 7.
The regional comparison of assessment results of various marine DRR capacities is shown in Figure 7. It is disclosed that the evaluation values of engineering defense capacities for the 11 pilot regions were basically the same. The evaluation values of engineering defense capacity for the other

Discussions
The key point in the assessment process was whether the set of index weights was reasonable or not. Therefore, communication and on-site training were fully conducted with the help of experts for decision-making, and to ensure that the index weights given by experts passed the significance test. In the future, appropriate sampling methods should be used to select working group members consisting of experts. When a quantitative weight is synthesized for expert opinions, the contribution of experts should be determined according to the experts' research field for each index.
In the future, a relationship between DRR capacity assessment and disaster management should be created, and a normal evaluation mechanism should also be established. Technical guidance for evaluating marine DRR capacity should be issued in a timely manner by the Marine Disaster Reduction Center of State Oceanic Administration, and the evaluation database for DRR capacity and

Discussions
The key point in the assessment process was whether the set of index weights was reasonable or not. Therefore, communication and on-site training were fully conducted with the help of experts for decision-making, and to ensure that the index weights given by experts passed the significance test. In the future, appropriate sampling methods should be used to select working group members consisting of experts. When a quantitative weight is synthesized for expert opinions, the contribution of experts should be determined according to the experts' research field for each index.
In the future, a relationship between DRR capacity assessment and disaster management should be created, and a normal evaluation mechanism should also be established. Technical guidance for evaluating marine DRR capacity should be issued in a timely manner by the Marine Disaster Reduction Center of State Oceanic Administration, and the evaluation database for DRR capacity and the development of reliable and efficient software should be carried out to promote the broad application of the evaluation results.
The benefits of comparing marine DRR capacities among the three Chinese provinces (i.e., Guangdong province, Shandong province, Zhejiang province) still need further study. The current situation of marine disaster mitigation capacity in the study areas will be further analyzed, and the evaluation results will be compared and analyzed. The specific measures and recommendations will be proposed for the water conservancy department, marine department, civil affairs department, and hygiene department.
Under the macro guidance of the national DRR plan [33], a regional DRR plan should be formulated in accordance with the background of disaster zoning. It is also necessary to conduct land-use planning for future developments in coastal areas in order to avoid new investments where the hazard level is high. There is an urgent need to establish coordination mechanisms between departments, to closely coordinate and integrate resources, and share information. According to the needs of prevention and management of marine natural disasters, we must pay close attention to the formulation and revision of regulations and national standards. It is necessary to establish prevention-oriented work ideas based on the characteristics of marine disasters, and to accelerate the construction of working mechanisms such as marine disaster prediction and warning, information reporting, emergency response, reconstruction, and investigation. There is an urgent need to strengthen the publicity and education of marine emergency management in order to maximize the ability of society to respond to marine disasters. It is also necessary to carry out extensive marine disaster emergency training in all walks of life.

Conclusions
Based on the current situation of marine DRR capacity building in China, a three-level index system was proposed. An expert questionnaire and the AHP method were applied to quantify the weights of the three-level indicators. The key conclusions are as follows: (1) The relationships between regional marine DRR capacity and various aspects of disaster management before, during, and after disasters were developed systematically. Disaster cycle, risk response, object and subject of capacity assessment, regional commonality, and specificity of indicators were considered comprehensively. By integrating the DRR experience of experts in various fields (such as ocean, hydrology, meteorology, civil affairs, emergency, education, and comprehensive DRR), we established a three-level indicator system for regional marine DRR capacity assessment. It includes six first-level indicators, 21 second-level indicators, and 136 third-level indicators. (2) The Evaluation Index Weights Expert Questionnaire was designed. The questionnaires were handed out to various departments and university experts involved in marine DRR capacities. After calculating the average of the weights provided by these experts, the final weights were found to be reasonable. Reasonable integration of the weights of experts in relevant fields enriched the knowledge and accumulated valuable experience for supporting the regional marine DRR capacity assessment and capacity building. (3) The evaluation of the three-level indicator system and the index weight results was performed in 11 pilot counties in Shandong, Zhejiang, and Guangdong Province. Through capacity assessment, gap analysis and demand analysis were further refined. It was possible to finally set regional DRR targets within a certain period. The results have important theoretical and practical implications for guiding regional DRR capacity building. They contribute a good foundation for future regional assessment of regional marine DRR capacity. (4) When implementing the constructed indicator system in different pilot areas, one should account for the different marine disaster risks that need to be dealt with in different regions. The differences in regional indicators should be considered. Through pilot research, the indicator system could eventually be applied to the national assessment of marine disaster mitigation capacity; thus, the indicator system should be versatile in China's coastal areas. The evaluation indicated that there were very few indicators with regional characteristics. This had almost no impact on the evaluation results; thus, a unified indicator system could be adopted throughout the country. (5) In pilot assessments, the analysis should cover the hazards, hazard-affected environments, hazard-affected bodies, and response capacity of the marine disaster system. The availability of data on the grassroots management departments should be considered. Representative evaluation indicators need to be further screened and condensed. It is necessary to get a good balance among the integrity of the indicator system and data availability, the representation of indicators and number of indicators, and the key indicators and alternative indicators.