**5. Discussion**

This study used Yunlin County as case study for further discussion to verify the proposed assessment principles on design criteria and risk analyses. It is hoped that the results can be used for future coastal management. The data used in this study were obtained from the databases listed below:


**Figure 4.** GIS layers of the coastal hazard-prone areas in Yunlin County, consisting of (**a**) surge, (**b**) flood, (**c**) erosion, (**d**) ground subsidence, and (**e**) tsunami prone area layers.

The data for the five hazards were all acquired from official sources, in accordance with the stated requirement of data impartiality in this study. After overlaying the resulting five coastal hazard-prone area layers, we found that Kouhu Township was subject to a combination of all five hazard types. This township made up over 66% of the total hazard-prone area, resulting in a hazard factor score of 5. The rest of the villages had scores of 3 or above. The hazard factor score of each village is presented in Figure 5.

**Figure 5.** Hazard grading of the coastal villages in Yunlin County.

Population density, comprehensive income, and land use data were acquired from county government statistics, and were used to estimate vulnerability. Population densities and comprehensive incomes of the villages within the coastal area of Yunlin range from 1 to 4, and 1 to 3, respectively. Land use scores are 2 or 3, with most of the villages constituting productive areas, although a few non-productive areas also exist. The vulnerability grading was derived from the average score of the three indicators for each village (Figure 6).

Risk maps for each village (Figure 7) were constructed by multiplying their respective hazard (Figure 5) and vulnerability (Figure 6) scores. The results show that most villages were classified as Level D (low-intermediate risk). Level C (moderate risk) was the highest risk class reached, and among these villages, only Yongfeng (in the Taixi Township) was adjacent to the shore.

Coastal areas that are adjacent to others with similar natural and anthropogenic environmental characteristics can be incorporated into a single protection area unit, and assigned an appropriate protection level. Table 4 presents the assessment of the coastal defense of Yunlin's coastal areas. The results show an overall coastal protection level of D (low-intermediate risk). The run-up heights

of waves with a 25- and 50-year return period were adopted as the coastal protection structural design criterion.

**Figure 6.** Vulnerability grading of the coastal villages in Yunlin County.

**Figure 7.** Risk classes of the coastal areas in Yunlin County.


**Table 4.** Risk level assessment and suggested design criteria in Yunlin County.

The designed and surveyed height of the coastal defenses is shown in Table 5. The height is referred to mean sea water level. Waves reaching a run-up height equal to the expected 25- and 50-year return period events are also listed in the two columns to the far right for comparison. These data were estimated using DHI MIKE 21 numerical models, including the effects of waves and tides. It is apparent that all the existing coastal defenses in Yunlin met the safety standards.


**Table 5.** Assessment of coastal defenses in Yunlin County.

<sup>1</sup> Due to land reclamation, this seawall was located inland and was not subjected to any marine force. <sup>2</sup> The seawall did not face the sea directly; the effect of waves can be neglected and assessed by the design surge height. <sup>3</sup> It was estimated from the elevation change of the control points that set on the crest of sea wall. The subsidence rates are the average values from 2007 to 2012 except Santiaolun seawall from 2010 to 2012 [44].

However, it should be noted that ground subsidence is still active in the coastal areas of Yunlin, with subsidence rates around coastal defenses in the order of 0.03–0.11 m/year. In addition, global sea level rise rates were estimated at 0.0028–0.0036 m/year between 1993 and 2010 (Church et al., 2013) [47]. Lan et al. (2013) simulated sea-level-rise scenarios in the range of 0.10 to 0.15 m during the time period ranging from 2020 to 2039 for the case of the southwest coast of Taiwan [35]. Thus, the level of protection offered by the breakwater/revetment is likely to be downgraded in the near future, given the influence that sea level rise and ground subsidence have on the long-term stability of structure. The data also shows that the coastal facilities located in Taixi Township are experiencing a higher subsidence rate than the others, and are thus categorized as having the highest risk level. We suggest that periodical

monitoring projects be conducted for further assessment. According to our risk assessment, once the protection level of the existing breakwater/revetment goes below the standard 25-year return period, modification of the defense measures should be conducted immediately.

#### *5.1. Non-Engineering Measures*

Considering Yunlin's relatively high hazard score, both engineering and non-engineering measures should be conducted simultaneously, as part of ICZM. We suggest the following non-engineering measures.

• Delimiting the buffer zone

The buffer zone in Yunlin took into consideration the dominant coastal hazards of storm surges, ground subsidence and tsunamis. Areas with intermittent or even frequent flooding were given special attention. The townships of Taixi and Kouhu have the highest hazard scores and delimiting the buffer zones is suggested as their first priority. However, regulations in the buffer zone may hinder coastal development; two levels of buffer zones were therefore suggested. The first one was delimited based on the more frequent hazards, such as storm surges and ground subsidence, with strict limitations. The second one considers the very rare hazard, such as tsunami, and the regulations there are more flexible. The first level is marked 50 m further landward of the 25-year return period storm surge water level. No further exploitation within this buffer zone should be permitted. The latter one is delimited based on the tsunami potential risk area which was shown in Figure 4e. A restrictions and permissions system should be implemented where the tsunami potential inundation depths are more than 1 m. Any activity that may lead to explosion or toxic pollution should be forbidden. The development of residential communities in highly populated areas is to be regulated. Low density developments are to be encouraged.

• Construction of the hazard maps

A hazard database should be established and maintained, and a risk management and economic analysis should be carried out to develop a coastal protection policy and regime.

• Land-use modification

The overuse of groundwater is the main factor causing ground subsidence in this region. A large amount of groundwater is being extracted and used for aquaculture; therefore, the use of land for new aquaculture applications should be restricted. Local government agencies should be given a portion of grant funds intended to reduce the development of fish farms. For maintaining groundwater balance, river water and rainwater can be used, and wastewater from industrial plants may be treated and then discharged into the rivers or artificial lake to increase the infiltration.

It is suggested that land-use changes into detention basins or is used for building solar power systems. These regions have suffered constant inundation and consequently yielded below-average production benefits. The proposed change of land use is a win-win solution to reduce the impact of inundation and provide new resources.

• Building renovations

Due to the threat of flooding, buildings should be modified to prepare for flood hazards. Some suggested facilities or modifications include waterproof gates, foundations on stilts, and temporary polder dykes. The entire coastal area in Yunlin is affected by ground subsidence, which resulted in low-lying lands. Buildings with low foundations experienced frequent inundation on the lower floors, including the basement and first floor. Hence, limiting the use of low floors to decrease the loss of assets is another option. Following such limitations, reductions in housing and land taxes can be used to encourage residency. The Taixi and Kouhu townships are especially encouraged to deploy these measures. The foundation elevation of any new buildings should be higher than the 25-year design criterion. It can also reduce potential losses induced by tsunamis.

Any changes in land use patterns that fall under scores 4 and 5 in Table 2 should be made tsunami resistant. This would typically require buildings to be made using concrete, and preferably allow vertical evacuation under the worst case scenario high-return tsunami events. Evacuation shelters should be designated in the regions where the tsunami potential inundation depths exceed more than 1 m. The public should be made aware of emergency evacuation routes through the use of mass media and public address systems.

• Improvement of forecasting and warning

The main purposes of the defense system are the improvement of forecasting and warning for storm surge, flood, and tsunami. Disaster prevention training and education to make people aware of coastal hazards and what to do in case of hazards should be conducted on community basis. These exercises and education on precautions and preparedness against hazards will reduce potential losses.

### *5.2. Public Participation*

Responsible authorities are now preparing drafts of coastal protection plans with respect to different regions in Taiwan, and should be announced in 2020 following the Act. During this period, public participations in establishing coastal defense and management policies are to be encouraged. Three-stage of public participations, including opinion polls, public hearings, and consensus conferences, should be carried out before the draft of coastal protection area planning are to be completed. Risk maps generated by this study can be a useful tool to show the public that the threats of flooding are both realistic and imminent, demonstrating possible coastal risks and their effects. At the same time, non-engineering measures proposed in our study can be used to offer alternative measures for mitigation, improving acceptance and understanding by the public. Furthermore, when necessary, factors and weights proposed in this study can be easily adjusted in future studies in accordance with public interests.

#### **6. Conclusions**

Decision of adaptation or mitigation strategies for coastal defenses is made on a national scale in Taiwan. For this, policy makers should be provided with as much information as possible. On the other hand, too much information at one time often causes confusion, leaving it hard to make correct decisions. Furthermore, the dynamic coastal processes are complicated, often with severe socio-economic consequences, making the correct decision is therefore of vital importance. As the authorities are formulating costal protection plans in Taiwan at the moment, coastal risks each region are facing should be clearly identified in the first place. Coastal protection sectors have been announced by CPAMI based on the severity level of coastal hazards, but the risks to coastal residents have yet to be assessed [31]. In this paper, we propose a scheme for a quick but informative risk assessment. It is based on the method of the United Nations Disaster Relief Organization [3], which uses both hazard and vulnerability indicators. It can be readily to be applied since the graded coastal hazard severities are used. In the process of formulating a coastal protection plan, if necessary, the vulnerability factors or weights can be altered according to the public interests. In this way, coastal management criteria will be formulated according to coastal risks. Both residents and authorities can be made aware of the risks of the area, and the officials can draw a prioritized list based on the criteria and make defense plans accordingly. Rational use of resources for adaptation or mitigation measures can be also be achieved.

Yunlin County, with its existing coastal defenses and the present status of land use, was used for the case study to test the proposed assessment principles of coastal management criteria. It was found that the risk class of the coastal areas in Yunlin County can be categorized as low-intermediate. The coastal hazards are comprised mainly of storm surge and ground subsidence. We then proposed implementation of both engineering and non-engineering measures to promote sustainable management of coastal zones, and to reduce the impact of coastal hazards.

**Author Contributions:** W.-P.H. designed the study and developed the methodology. J.-C.H. and C.-S.C. collected the data and analyzed the results. C.-J.Y. primarily made figures and tables of the manuscript.

**Acknowledgments:** This paper is partly aided by a project of the Ministry of Science and Technology, Taiwan, Project No. MOST 106-2621-M-019-003-MY2. Their support is deeply appreciated.

**Conflicts of Interest:** The authors declare no conflict of interest.
