A growing number of countries across the world have been affected by floods due to dense population, inappropriate land use planning and climate change (Cardona, 2004; Gandini et al., 2020; Lin et al., 2019; Myhre et al., 2019). Developing countries however are the most affected. This is due to their poor socio-economic conditions, inadequate financial resources, increased imperviousness, inadequate drainages, poor solid waste management and the construction of houses in flood plains and wetlands (Adelekan, 2010; Asiedu, 2020; Peters-Guarin et al., 2012). In developing countries, most informal settlements are located in sensitive and fragile environments, which makes them susceptible and vulnerable to flood hazards (Abunyewah et al., 2018; Mahabir et al., 2016). According to Roy et al. (2018, p. 283), close to 60% of the informal settlement dwellers in developing countries live in Sub-Saharan Africa and the UN-Habitat (2010) estimates that by 2050, the number of people living in informal settlements will increase to 3 billion. The UN-Habitat (2015) defines an informal settlement as housing areas where people build houses without complying with planning and building regulations, lack basic services and infrastructures and have no security of tenure. There is, therefore, a need to have a critical interest in reducing flood vulnerability in informal settlements, particularly in developing countries because informal settlements accommodate most of the urban dwellers (Duijsens, 2010; Flores et al., 2020; Zerbo et al., 2020).
A review of literature conducted by Membele et al. (2022a) shows that there has been a shift in the way flood vulnerability is considered in developing countries. Flood vulnerability is widely been considered to be integrated because it combines both physical and social vulnerability (Cutter et al., 2003; Ehsan et al., 2022; Paul & Routray, 2010). Integrated flood vulnerability is important because it takes a holistic and interdisciplinary approach crucial in facilitating a complete assessment of flood vulnerability (Barroca et al., 2006). Flood vulnerability is therefore considered as an interrelationship of exposure, sensitivity or susceptibility and adaptive capacity (Chen et al., 2021; Ehsan et al., 2022; Huynh & Stringer, 2018; Roy & Blaschke, 2013; Sahana & Sajjad, 2019). Exposure is defined as the predisposition of a system, community or physical items to impacts of floods due to location (Balica et al., 2012; Ehsan et al., 2022; Hung & Chen, 2013; Sadeghi-Pouya et al., 2017), while sensitivity is defined as the fragility or capacity of a system, individual or community to withstand the impact of flood hazards (Jha & Gundimeda, 2019; Roy & Blaschke, 2013; Yankson et al., 2017). Adaptive capacity is the ability of an individual, system or community to adjust, respond or recover from an adverse impact of floods (Borbor-Cordova et al., 2020; Kienberger, 2012; Roy & Blaschke, 2013). In our view considering flood vulnerability from an integrated perspective facilitates strategic policy formulation and implementation, which are important for sustainable disaster management.
According to Mazumdar and Paul (2018) locating vulnerable people in a community and identifying the reasons for their vulnerability has been a huge challenge for decision and policymakers. Mapping flood vulnerability especially at a local level is crucial because it helps to precisely locate where highly vulnerable people or households are, thereby helping in designing appropriate emergency alternatives and mitigation strategies (Bisht et al., 2018; Hoque et al., 2019; Mazumdar & Paul, 2018; Romanescu et al., 2018).
Therefore, this study was anchored on the ‘place-based’ approach to mapping flood vulnerability (Cutter, 1996; Cutter et al., 2008; Dintwa et al., 2019). According to Dintwa et al. (2019), the feedback mechanism embedded in place-based approaches where an increase or decrease in risk, leads to enhanced or decreased vulnerability, allows it to inform policy and mitigation interventions. Hung and Chen (2013) contend that mapping flood vulnerability is important because it helps to guide decision-makers on how they can prepare and deal with climate change impacts. Furthermore, mapping flood vulnerability enhances the participation of community members (Martin et al., 2018; Membele et al., 2022b; Scheuer et al., 2013; Wilk et al., 2018; Yen et al., 2019). Membele et al. (2022b) argue that local and indigenous knowledge help to foster community participation and the implementation of context-specific adaptation measures. Local and indigenous knowledge is however not the same.
According to Langill (1999), local knowledge is knowledge acquired or possessed by people because of living in a particular community or locality for a considerable period. Indigenous knowledge is a body of knowledge that is uniquely developed and rooted in the culture of a particular area. This knowledge is embedded in people’s way of thinking, skills, technology, culture and social practices and passed on from one generation to the next through repetition or demonstration (Fabiyi & Oloukoi, 2013; Sillitoe, 2007; UNEP, 2008).
However, the use of local and indigenous knowledge in mapping flood vulnerability especially in informal settlements remains underutilised (Dube & Munsaka, 2018; Membele et al., 2021, 2022a), mainly because some practitioners argue that local and indigenous knowledge cannot be scientifically validated. Chanza and De Wit (2016) contend that community members’ situational and experiential knowledge is crucial in mapping flood vulnerability at a local level. Hung and Chen (2013) further argue that the incorporation of local and indigenous knowledge in mapping flood vulnerability especially in developing countries has been a challenge. However, local and indigenous knowledge have been identified to be crucial in dealing with hazards like floods at local levels (Holley et al., 2011; Mavhura et al., 2013; Membele et al., 2021, 2022b; Ziervogel et al., 2016). In particular, the use of indigenous knowledge in helping to protect communities in high-risk areas to build resilience has also been underscored by the Sendai Framework for Disaster Risk Reduction (UNDRR, 2015).
Many strategies have been used to map flood vulnerability in developing countries. The indicator-based Multi-Criterial Decision Making (MCDM) using Analytic Hierarchical Process (AHP) and Geographical Information Systems (GIS) have been widely used in mapping flood vulnerability in developing countries (Abdullah et al., 2021; de Brito & Evers, 2016; Membele et al., 2022a; Rehman et al., 2021). The indicator-based approach is common in mapping flood vulnerability because of its flexibility, trustworthiness, transparency and ability to combine many elements that contribute to making people and places vulnerable to hazards like floods (Balica et al., 2009; Ciurean et al., 2013; Kappes et al., 2012; Nasiri et al., 2016). GIS-based MCDM approaches have been helpful in vulnerability mapping because they have an explicit, rational, spatial and efficient process that leads to justifiable and explainable choices, thus helping to enhance quality decision making (Abdrabo et al., 2020; Ferretti & Pomarico, 2013; Morea & Samanta, 2020).
The AHP developed by Saaty (2007) was widely used in mapping flood vulnerability in developing countries, due to its simplicity, flexibility and ability to structure the decision problem in a hierarchy (Das, 2020; de Brito et al., 2018; Hoque et al., 2018; Msabi & Makonyo, 2021; Roy & Blaschke, 2013). However, Aminu et al. (2014) argue that the AHP considers flood vulnerability elements as separate elements. It is our considered view that the ‘separateness of elements’ seldom happens in real life because flood vulnerability elements namely exposure, sensitivity and adaptive capacity are interwoven (Akukwe & Ogbodo, 2015; Ebi et al., 2006; Hung & Chen, 2013; Roy & Blaschke, 2013; Yuan et al., 2016). Ghorbanzadeh et al. (2018) further contend that the AHP does not consider multiple alternatives at a time.
One of the MCDM approaches that take into account interdependent elements is the Analytic Network Process (de Brito et al., 2018; Ekmekcioğlu et al., 2022; Esfandi et al., 2022; Ghorbanzadeh et al., 2018; Ghosh et al., 2021). However, studies that used ANP in mapping flood vulnerability, particularly in developing countries and informal settlements in particular are rare (Membele et al., 2022a). A few studies (de Brito et al., 2018; Ishtiaque et al., 2019) that used ANP to map flood vulnerability in developing countries, mapped flood vulnerability at a municipal level and sub-district level respectively, but not at a local and fine scale such as informal settlement. Furthermore, a few studies (de Brito et al., 2018; Ishtiaque et al., 2019) that have used the ANP to map flood vulnerability in a developing country context, used experts or decision-makers and not community members. Hoque et al. (2019) contend that mapping flood vulnerability at a local scale such as an informal settlement by using a multi-criteria analysis approach was crucial in providing detailed and accurate flood vulnerability information needed for decision-making. However, it has been argued that many MCDM studies especially in developing countries suffer from a lack of updated spatial data. To overcome this challenge, collecting accurate spatial data using field surveys were gaining traction in data-scarce environments like developing countries (Akukwe & Ogbodo, 2015; Huq et al., 2020; Lian et al., 2017; Muller et al., 2011; Sarkar & Mondal, 2019; Usman Kaoje et al., 2020).
Therefore, this study demonstrates the integration of community members’ local and indigenous knowledge with a GIS-based MCDM using ANP to map flood vulnerability in an informal settlement. In particular, the study was conducted in an informal settlement called Quarry Road West located in Durban, South Africa. The study endeavours to answer the following questions: To what extent can an approach that integrates local, indigenous knowledge and GIS-based MCDM using ANP be used to map flood vulnerability in Quarry Road West informal settlement? How well does the ANP present flood vulnerability in the study area? What areas in the study experience low, moderate and high flood vulnerability? The novelty of this particular study lies in the operationalisation of indicators selected using community members’ local and indigenous knowledge in mapping flood vulnerability in an informal settlement (Membele et al., 2022b). This study is also significant because it represents one of the first experiments that used the ANP to map flood vulnerability through the participation of community members living in an informal settlement