2.1. Study Site: A Flat Sandbar of Taipei City
Flooding is considered to be the most immediate risk from climate change to Taipei City that is developed near a river mouth, characterized by flood plain and low-lying topography in a coastal zone [17
]. An illustrative case is the Shetzu Peninsula, which is a low-lying and flat sandbar located in an estuarine region at the junction of the Keelung and Tamsui Rivers of Taipei City, Taiwan. Although Taipei is a highly developed city and has completed flood-control projects, it is still very vulnerable to flooding when extreme climate events [17
], e.g., sea-level rise, occur (Figure 1
). Figure 2
shows the fort map’s contour of Taiwan, which was published in 1904, and aerial photographs taken in 2017 (Google Map) demonstrate the geomorphic changes in the Shetzu Peninsula. In 1963, Typhoon Gloria generated heavy rainfall, which induced a surge of river water, flooding the Shetzu Peninsula. In light of this disaster, the national government began addressing flood risks in Taipei. In 1970, the Taipei Region Flood Control Project was completed, designating the Shetzu Peninsula as a detention basin to reduce flood risks in other parts of Taipei. As the Shetzu Peninsula developed rapidly and the population grew, the government subsequently installed flood-control infrastructure, such as embankments and pumping stations. However, the functionality of the flood detention basin and the embankment in Shetzu Peninsula was the lowest protection height in Taipei. In addition, construction was banned to restrict population growth and land use. Although Shetzu Peninsula has been restricted to low-density development, approximately 10,000 residents, whose families have lived there for generations, have continued to reside in the area. For more than four decades, these residents have lived in settlements and houses that have been banned from major repairs and alterations.
After the embankment was built, Shetzu Peninsula seemed to exhibit the capacity to cope with the magnitude of the flood disembarked by Typhoon Gloria (1963). A close examination of the residents’ environmental and socio-cultural perceptions revealed that the local residents acquired knowledge about the impacts and adaptation to environmental threats based on flood-risk perceptions [20
]: people who have experienced related impacts exhibit more knowledge and skills compared to those who have not [21
]. However, Shetzu Peninsula residents currently rely entirely on the hard infrastructure for flood control and have grown accustomed to it. Therefore, residents have difficulty imagining flood situations that may happen in the future, and this causes the impact of climate risk to increase [22
]. It was argued that flood damages caused by extreme climate and weather events without any adaptive measures could progressively increase, causing damages to the infrastructure and the residents [23
]. The Shetzu Peninsula is a flat sandbar in the villages of Taipei City, which is located in an estuarine region that is exposed to flooding; and where additional risks are imminent with climate change and remain unmeasurable.
2.2. Communicating Flood Risks
A science-based risk analysis with psychological risk dimensions may not be considered in technical risk assessments for cultural subgroups of society [24
]. To help the local residents better understand and respond to the flood risks under climate change, this study used flood potential mapping and a rise in sea level scenario through an interactive visualization model to communicate flood risks with the local residents in the retail sector (Figure 3
). It is noted that risk map may fail to represent the information base that is needed by stakeholders for selecting the suitable objectives and designing the appropriate actions for flood risk management [25
]. This model focuses on raising awareness throughout presenting information relevant to the interest and location of the end-user [26
]. It underlines the concept of place attachment that is already well established within sociology and geography, describing the identity created around a settlement or place [14
The interactive visualization model uses Geodesign techniques to address the design and planning method that tightly couples with the creation of design proposals and the simulation of the impacts pertaining to geographic contexts [28
]. It provides a design framework and supporting technology for professionals to manipulate geographical information, emphasizing environmental evidence-based data, design, and decisions. In a typical Geodesign procedure, the opinions and collaborations of the public, designers, decision-makers, and related stakeholders are necessary during the project planning and design stage [29
]. The study used the three-dimensional modeling software program designed by Esri CityEngine to create an interactive visualization model. This model is equipped with a procedural modeling approach and a strong online model-sharing service for smoother communication and interaction with stakeholders, such as the local residents and policy-makers [30
2.3. Identifying Key Industrial Sector
As mentioned above, the retail sector is expected to be the most affected by flooding. The study further conducted Location Quotient (LQ) analysis to verify if the key industrial sector of Shetzu Peninsula is referred to as a retail sector, by focusing on regional sectoral composition and concentrations of cluster activity [32
]. The LQ technique is based on economic base theory and is the oldest, simplest, and most widely used technique for regional economic analysis [34
]. It estimates the share of an economic activity in a spatial unit compared to the corresponding share of this activity in the national or specific denominator region’s economy [35
], which, in terms of a variable, in this study, is the employment. The focus was put on the analysis of the economic base in an area to identify key industrial sector and focus groups of the study for formulating appropriate measures to be adapted to local conditions (problems, needs, and potential of the area) [36
] in regard to climate change.
The LQ analysis identified the employment concentration by utilizing the data from the government census of the period: 2006–2015 [37
]. This approach addresses industrial sectors that differ significantly in respect to exposed assets and susceptibility to floods [39
]. In addition, LQ analysis determines the highest concentration for sectoral composition for an area, identifying a local measure of geographical concentration of industries by comparing the local share of jobs in a specific industry and in the City. It was noted that (Table 1
Peripheral places, such as Shetzu Peninsula (compared to Shihlin District and Taipei City) would have smaller LQ deviation since the LQ indices represent the level of sectoral dominance in the regional economies and are to issue economic diversity, size and economic scale. That is, the industry in question may not show large location quotients because the relative importance of the sector is muted with respect to overall economic activity. Therefore, the study neglected to incorporate agglomerating influences and the sectoral distribution in the analysis, simply using LQ to correlate the regional convergence of retail industry within the sectoral distribution and with the relevant allocation of employment at regional level to define whether retail sector is the basic industry for Shetzu Peninsula.
The study took into account that the employment population disaggregated into sectors of manufacturing, electricity, gas and water, construction, retail, communication, finance and insurance, and service sectors. Data that were conducted every five years were gathered from the Industry, Commerce and Service Census of the Directorate General of Budget, and Accounting and Statistics (DGBAS) of Taiwan, and the census numbers from 2006 to 2011 were interpolated and used to extract the employments for 2005, 2007–2010, and 2012–2015, respectively. The census divides the employment data into two-digit sectoral division, where the different sector division categories were used in accordance to the 2001 and 2006 data divisions. It is noted that: (1) data from the industry and commerce census were compared to aggregated data from the Ministry of the Interiors of Taiwan; (2) data included annual demographic data (municipal level and national level), employment, industry and commerce census data in the eight sectors according to the two digit category; and (3) the aggregated ratio for the gap years of each census data (every five years) was used to conduct the LQ analysis. The LQ formula is:
Ei,r is the number of employee jobs in industry i region r;
Er is the number of employee jobs in region r;
Ei is the number of employee jobs in industry i; and
E is the number of employee jobs in Shetzu, Taipei City.
The study focused on the three villages of Fuzhou, Fuan, and Yonglun that are located in Shetzu Peninsula. These areas are identical to the redevelopment areas proposed by Taipei City. The study computed proportionally the employment population from the extrapolated employment in the industry and commerce and services census of 2006 and 2011 from the growth ratio obtained from the annual population census from the Ministry of Interiors; and the results were used to explore the employment population for the other gap years between the two censuses. For the three villages of Fuzhou, Fuan, and Yonglun, the sectoral employment of each neighborhood was computed proportionally with the total employment extrapolated for the Shihlin District. After obtaining the extrapolated employment for the years of 2005–2015, the LQ analysis was conducted.
This approach enabled the observation and analysis of any LQ changes in the three villages and the opportunity to compare the results to that of Shetzu Peninsula and Shihlin District. The LQ approach represents the level of sectoral dominance in the regional economies and is sensitive to issues of economic diversity, size and economic scale. The study neglected to incorporate agglomerating influences and the sectoral distribution in the LQ analysis to correlate the regional convergence of retail industry within the sectoral distribution with the relevant allocation of employment at local level. This helped to understand the changes in the industrial sectors of the studied site and to identify the key industry of the Shetzu Peninsula in the retail sector.
2.4. Semi-Structured Interviews
Supported by the interactive visualization model, the consumption intensity map, and the LQ analysis that identified key industrial sector, a semi-structured interview process was conducted using focus groups. The study underlined that studies that use more than one method require fewer participants, as do studies that use multiple (very in-depth) interviews with the same participant [42
]. For this reason, only 15 respondents were included in the study. As such, the semi-structured interview process continued until the respondent either repeated the same answers or was not able or willing to answer [44
It was underscored that the risk events occur in a societal context and that the residents who provide a vital workforce would acquire an understanding of flood risks and have adopted acceptable levels of these risks. It was also argued that the flood-resilient residents are able to tell a “story” about the environmental impacts, and that their response collaborates with researchers [45
]. The interview process underlines another aspect that complicates how people perceive, evaluate, and act on climate change risks, namely the concept of social amplification of risk framework [46
]. That is, social and cultural risk amplification processes modify perceptions of risk in terms of risk communication [48
]. It was noted that risks are communicated through different signals, such as images, and these signals can amplify or attenuate the perception of risks and their manageability [49
]. That is, the semi-structured interview was used to determine the respondents’ flood-risk perceptions that are related to their flood resilience thinking under climate change. The interview addresses residents’ reactive awareness, their proactive actions in response to climate change, and their inextricable blend of face and feeling, reason and gut reaction, and cognition and intuition [50
]. The interview format helped to provide insights into retailers’ resilience thinking in a systematic manner and enabled us to have an open dialogue with residents and explore how risk events interact with psychological, social, and cultural processes [51
The study underlined that flood risks refer to the local residents’ knowledge, skills, and networks (indices) that organisms accumulate during their developmental processes, which reflect their adaptability to the changing environments; the changing of social relations and the context lead to a change in the meaning of places [52
]. The indices of knowledge, skills, and networks are built upon a psychological methodological perspective, and reflect an uncertainty related to the input data and provide emphasis on the qualitative character of the assessment results in terms of potential impacts [13
]. The study argued that the indices refer to residents’ preferences, values, attitudes, behaviors, and decision models, which the individuals exhibit differently depending on the environments and their perception of these environments. These differences cannot be evaluated using physical geographical data. Therefore, flood resilience in the retail sector was explored based on human adaptability by using the three indices: knowledge, skills, and networks. The study underlined that human adaptability positively influences individual responses to the impacts of climate change. Consequently, the indices were calculated using positive points.
It is noted that human adaptability, which influences risk perceptions, is the key to coping and adapting to changes toward social resilience [52
]. Human adaptability reflects the influence of various impacts on individuals and is strongly related to individual knowledge and skills, socioeconomic structures, and institutional policies [52
]. The learning of new knowledge and skills to reduce vulnerabilities and manage climate change risks also help achieve sustainability [54
]. Many studies have also started to focus on the role of social networks, and the frequency with which people talk to or are influenced by close friends and family on the issue of climate change [55
The study conducted semi-structured interview schedules with respondents who were key retailers of the village of Fuzhou. The interview was open-ended, allowing new ideas to be brought up during the interview by the respondents. That is, the interview design allowed us to have open dialogues with the respondents, noting that risk events interact with psychological, social, and cultural processes, which helped to guide the interviewing process towards an understanding of respondents’ resilience thinking in a systematic manner [51
]. The respondents were selected to represent age, gender, occupation, and leadership. Interviews took place in their stores. All of the interviews were recorded and transcribed. Interview questions with specific themes included exploring memories and responses to the flood scenario of both extreme precipitation and sea level rise, focusing on their awareness of the potential flood risk in the future climate.
Respondents’ opinions were then arranged according to the aforementioned indices: knowledge, skills, and networks (referred to human adaptability), and the keywords of building, road, land use, and flooding (referred to risk information) based on the interactive visualization model (Figure 3
). The categorical values of the respondents’ flood-resilience thinking can, thus, be obtained. Supported by those indices and keywords, the study developed and used an interview guide focused on respondents’ human adaptability and risk information as follows:
Human adaptability: Have you ever gained any knowledge, used any skills, or shared your knowledge or skills in your neighboring networks? (if referred to one index, note value “1”).
Risk information: Are you aware of the protection of buildings, roads, and land use against flooding (if referred to one keyword, note value “1”).
This interview guide addressed an informal grouping of topics and questions on how to adapt to floods in terms of risk communication under climate change that the interviewer can ask in different ways for different respondents.