Jàmbá - Journal of Disaster Risk Studies ISSN: (Online) 1996-1421, (Print) 2072-845X Page 1 of 10 Original Research Flood risk perception, disaster preparedness and response in flood-prone urban communities of Rivers State Authors: Rapid urbanisation is contributing to increasing societal vulnerability to disaster. This study Zelda A. Elum1 aimed at exploring the perception on flood risk and ascertaining the determinants of disaster Olanrewaju Lawal2 preparedness among residents in flood-prone urban communities. Descriptive statistics and Affiliations: discriminant regression model were employed on primary data collected from 240 urban 1Department of Agricultural households across five communities at risk of flooding in the study area. The results showed Economics and Extension, Faculty of Agriculture, that most households had low awareness of flood risk and exhibit low levels of adaptive University of Port Harcourt, capacity, having adopted little or no measures to deal with disaster floods. Also, awareness of Port Harcourt, Nigeria flood risk was observed to discriminate the most between the two groups of adopters and nonadopters of flood preventive and management measures (proxy for disaster preparedness), 2Department of Geography followed by flood risk perception, age, location and household size. and Environmental Management, Faculty of Contribution: The study suggests an integrated approach (a combination of preventive, Social Sciences, University of Port Harcourt, Port Harcourt, protective and control measures) by all stakeholders, including government and other relevant Nigeria bodies, increasing public awareness of flood risk and its attending effects for greater responsiveness, supporting communities in regular clearing of drainage areas and strictly Corresponding author: regulating the construction of buildings, particularly in flood prone areas. Zelda Elum, zelda.elum@uniport.edu.ng Keywords: urbanisation; climate change; floods; sustainability; risk perception; vulnerability. Dates: Received: 19 Feb. 2022 Accepted: 08 July 2022 Introduction Published: 29 Sept. 2022 Urbanisation is a growth that defines increasing human activities which have implications for the How to cite this article: economic, social, political and physical geography of an area (Caparros-Midwood, Barr & Dawson Elum, Z.A. & Lawal, O., 2022, 2015). Nigeria’s increasing rate of urbanisation has been linked to environmental degradation ‘Flood risk perception, disaster preparedness and (Duh et al. 2008). Rapid urbanisation aggravates the challenges already posed by climate change, response in flood-prone increasing vulnerability to climate change impacts. Undoubtedly, urbanisation in developing urban communities of Rivers countries is often accompanied by increasing environmental risks and scarcity of resources. Some State’, Jàmbá: Journal of of the environmental risks magnified by urbanisation include flooding. Its increasing occurrence Disaster Risk Studies 14(1), a1303. https://doi. across the globe has been linked to climate change (Abass et al. 2022). Nigeria in particular is org/10.4102/jamba. notably identified as vulnerable to climate change–related disasters such as floods, epidemics and v14i1.1303 droughts on account of the myriads of socio-economic constraints confronting the region and its inhabitants. Copyright: © 2022. The Authors. Licensee: AOSIS. This work Vulnerability describes the susceptability of people or places to damaging impacts arising from is licensed under the exposure to hazard events. In the context of flooding hazards, vulnerability describes how Creative Commons predisposed people or systems are to experiencing floods impacts differently given the variablity Attribution License. in their characteristic features. Vulnerability is influenced by physical, social, economic and environmental factors (Douglas et al. 2008; Lawal & Arokoyu 2015; Tingsanchali 2012), and identiying these factors is a vital pathway to adopting appropriate strategies for coping or mitigating the impacts. Understanding the dynamics of hazards, exposure and vulnerability of communities is important for building their resilience (Etinay, Egbu & Murray 2018). This is because exposure (a component of vulnerability) to hazards like floods is determined by the probability of occurrence of the hazards (e.g. flood risk) and the sensitivity of the people or systems to the impact of the hazard (Brooks 2003), which is dependent on their adaptive capacity Read online: and which consequently influences their resilience. Scan this QR code with your smart phone or According to Lawal and Umeuduji (2017), flood is the most reoccurring natural disaster in Nigeria, mobile device resulting from gradual build-up of rainwater on saturated ground or spontaneously in areas with to read online. inadequate storm water management. Flooding is a natural characteristic of rivers where high http://www.jamba.org.za Open Access Page 2 of 10 Original Research flow of water overflows the natural banks (Baten, Arcos can be affected and to take actions towards reducing the González & Delgado 2018). However, it is one of the major impacts well ahead of the occurrence of disaster events climate change impacts that have caused people to be (United Nations 2016). Often, risk is communicated to displaced from their homes, sometimes causing people to encourage precautionary measures (Netzel et al. 2021) among migrate from rural or agricultural communities to urban or the stakeholders at risk. The models employed for risk nonagricultural settlements (Barrios, Bertinelli & Strobl 2006; communication are very much dependent on the direction of De Brauw, Mueller & Lee 2014). For example, the severe communication, the roles of the communicator and the floods of 2012 that occurred in many parts of Nigeria receiver and the purpose of the communication (Rollason including Rivers State left many homes, communities and et al. 2018). Flood risk communication entails identifying businesses destroyed (Akukwe, Krhoda & Oluoko-Odingo flood-prone areas and letting stakeholders know the causes 2018). As it has been observed in Nigeria and specifically and likelihood of floods occurring as well as the probability Rivers State, more often than not, the cause and magnitude of of damage (Demeritt & Nobert 2014; Rollason et al. 2018). floods is attributed to poor physical planning, poor living Effective risk communication creates greater risk awareness habits of people (e.g. dumping of refuse in drainage areas that informs how the people percieve them as important and putting up shanties without care on water channels) and issues that need disaster preparedness actions or response disregard for city plans, building codes, environmental rules for disaster risk reduction (Barquet & Cumiskey 2018). and regulations in affected communities (Aderogba 2012). Specifically, city expansion due to increasing population in Disaster risk reduction, in the context of flooding, is acquiring Port Harcourt has been perceivably marred with challenges, adequate capacity and knowledge to build sustainable resulting in informal settlements and the building of houses infrastructures so as to reduce people’s exposure and on unapproved sites and on flood plains, all in a bid to have vulnerability to flood hazard (Kwak, Muraoka & Asai 2018). cheap accommodation for the growing urban population There are two broad approaches to disaster risk reduction (Obinna, Owei & Mark 2010). It is on this background that (Intergovernmental Panel on Climate Change [IPCC] 2007), the study sets out to examine vulnerability of urban residents the top-down (based on institutional responses) and bottom- to floods, their awareness and perception of flood disaster up (based on local communities capacity to adapt and risk and to investigate the factors influencing the disaster prepare for disaster). Thus, there is a connection between preparedness and response among the urban dwellers. disaster risk reduction and disaster preparedness. Studies have shown that stakeholders’ adoption and effectiveness of Different methodological approaches have been explored in flood risk management strategies is highly dependent on understanding what influences or informs an individual’s their perception and attitudes towards flood risk (Santoro et risk perception and action. In some studies, the signal al. 2019). Effectiveness in this context is when the benefits of detection theory (a visual perception theory) has been adopting management strategies offset the impacts of the applied on the assumption that the detection or recognition flood in comparison to the outcomes in a nonadoption of a stimulus or event depends on the stimulus intensity and scenario. Empirically, logistic regression has been used to the physical and psychological state of the individual, which investigate the relationship between public risk perception of affects the individual’s ability to discriminate more intense floods and implementation of mitigation measures, using stimulus (or dangerous situations) from the less intense or variables such as age, gender, number of children, years less threatening (Eiser et al. 2012). The physical and living in the same house and highest educational attainment psychological state of the individual is itself shaped by (Netzel et al. 2021). Studies (Baten et al. 2018; Nur & Shrestha numerous factors that can be broadly classified as socio- 2017; Rakib et al. 2017) have documented that children, economic, environmental, cultural and political factors. For women and the elderly are among the most affected during instance, previous or direct experience of flood events is flooding because of their lack of physical capacity to bear assumed to influence one’s risk perception and, in turn, flooding impacts. Therefore, a household with a greater preparedness action, either by aiding the response phase in number of vulnerable persons that buys insurance to guard an organised manner or by causing a low level of personal against disaster loss or keeps a family evacuation plan could preparedness, particularly if there is greater reliance on substantially minimise loss and damage (Hoffmann & publicly built structures and flood defences (Cologna, Bark & Muttarak 2017) from flood hazards. In Hoffmann and Paavola 2017). Risk perception has been studied with Muttarak (2017), diverse factors broadly classified into different variables or the same variables operationalised in sociodemographic (e.g. education, income level, marital different contexts (Netzel et al. 2021). Nonetheless, it is status, children and the aged present in households), assumed that perception of disaster risk involves taking a structural or geographical (e.g. length of residence, natural decision (responses) that is informed by the individual’s environment and hazard risks) and psychosocial factors (e.g. sensitivity to the risk impact or its severity and the level or hazard awareness, knowledge and risk perception) have amount of information about the risk that is made available been identified as significant determinants of disaster to the individual. In other words, risk communication is an preparedness behaviour as well as effective responses. important aspect of assessment of risk perception and adoption of risk management strategies. Risk communication In Port Harcourt, some areas are naturally prone to flooding is about making people aware of disaster risk and how they (Plate 1) while others experience flood as a result of unplanned http://www.jamba.org.za Open Access Page 3 of 10 Original Research development whereby houses are built on valleys, floodplains capacity. Thus, this study becomes apt, as it is necessary to and water channels due to necessity and pressure to supply examine the flood risk perception of urban residents and to affordable housing for the increasing urban population identify local factors shaping their preparedness behaviour (Gerald-Ugwu, Egolum & Emoh 2019). By implication, those for effective response to flooding impacts in Rivers State. The living in these houses built on floodplains are increasingly study is guided by three research questions: (1) what is the vulnerable to flood disaster. Studies on flooding and state of awareness and perception of urban households about urbanisation in Port Harcourt have often focused on causes of disaster flood risk? (2) What is the adaptive capacity of the urban flooding or on the socio-economic and environmental households to cope with floods impact? (3) How do effects. Urban floods can result in significant economic losses households’ risk awareness, perception and other socio- directly and indirectly from loss of property and infrastructure, economic factors relate to their preparedness in dealing with loss of livelihood and poor health (Baten et al. 2018). While the flood event? The study by providing answers to these associated hazards are established, there is limited literature questions will help stakeholders to design appropriate on the disaster preparedness of the households for effective measures for flood prone areas in Rivers State. response mechanism and how it is shaped by their adaptive Materials and methods The study was done in Rivers State (Figure 1). Located on the Atlantic Coast of southern Nigeria, the state covers an area of 10 575 km2 (Federal Republic of Nigeria [FGN] 2011). The state is made up of 23 local government areas (LGAs), with a population of 11.5 million people (FGN 2011). The state capital, Port Harcourt, is popular for its oil and gas industry, and as such, it experiences a high level of commercial activities fuelled by the presence of oil multinationals operating in the state (Obinna et al. 2010). Port Harcourt, by virtue of its status as an oil city, tends to have a high PLATE 1: Photo of flooded areas in Port Harcourt metropolis. population influx (Potts 2015). Port Harcourt lies within the 4°0’0”E 6°0’0”E 8°0’0”E 0 90 180 360 kilometers Service Layer Credits: Sources: Esri, HERE, Garmin, USGS, Intermap, INCREMENT P, NRCan, Esri Japan, METI, Esri China (Hong Kong), Esri Korea, Esri (Thailand), NGCC, (c) OpenStreetMap contributors, and the GIS User Community 6°54’0”E 6°57’0”E 7°0’0”E 7°3’0”E 7°6’0”E 7°9’0”E 0 3.75 7.5 15 kilometers FIGURE 1: Obio-Akpor local government area and its environs. http://www.jamba.org.za Open Access 4°42’0”N 4°45’0”N 4°48’0”N 4°51’0”N 4°54’0”N 4°57’0”N 5°0’0”N 6°0’0”N 8°0’0”N Page 4 of 10 Original Research floodplains of the River Niger (Mmom & Iluyemi 2018). Data TABLE 1: Sample selection plan. was collected through a multistage sampling technique. The Community Sampling frame Sample size at 2% (number of households) proportionate sampling first stage involved a purposive selection of Obio-Akpor Nkpolu 700 14 LGA, which has been identified as highly vulnerable due to Rumuigbo 1300 26 its river floodplain settlements and proximity to rivers and Rumuekpirikom 3800 76 creeks (Akukwe & Ogbodo 2015). In the second stage, five Rumueme 3500 70 flood-prone urban communities were selected. Rumukalagbor 2700 54 Total 12 000 240 In the third stage, using the total number of households in the communities as sampling frame, a 2% proportionate TABLE 2: Definition of variables used in analysis. sampling was employed on a systematic random sampling of Variables Description a total of 240 households. This sample size (2% of total Age Age of respondent measured in years Household size Number of persons living with respondent population) was chosen due to monetary constraints. Households with children less than Number children less than 5 years However, it is above the 100 minimum sample size number 5 years allowable for meaningful research. More so, the sample size Households with dependants older Number of dependants older than than 60 years 60 years given the population size when computed according to Gender Gender of respondent (dummy: male = 1 Yamane (1967 as cited in Tangonyire and Akuriba, 2021) otherwise 0) sample size formula, falls between an acceptable ± 5% to ± Years of residency in the area Number of years living in the area 7.5% margin of error or uncertainty (Conroy 2018). The Education Categorical and ordinal; 0 = none, 1 = primary, 2 = secondary, 3 = tertiary distribution of sample drawn across the communities is Disaster preparedness Dummy; if any flood preventive or presented in Table 1. mitigation actions have been taken = 1 otherwise 0 Experience of past floods Dummy variable; 0 if never affected by The data were collated for the variables listed in Table 2. flood, otherwise 1 These were explored and analysed with descriptive Risk awareness Dummy; 1 if have information on flooding risk otherwise 0 statistics and discriminant analysis respectively. Discriminant Perception of flood risk Categorised as low = 1, moderate = 2, analysis builds a predictive model for group membership. high = 3 In this study, it was used to predict whether a household is Perception on climate change Respondents’ view of climate change as causing flood cause of flooding; dummy. Agree = 1 disaster prepared or not. Discriminant analysis is a otherwise 0 parametric analysis that helps to determine which of the Willingness to purchase insurance Dummy: Yes = 1 otherwise 0 independent variables will discriminate between the groups Perceived level of vulnerability to flood Categorised: low = 1, moderate = 2, high = 3 that makes up the dependent variable. The variables’ contributions to predicting of group membership are entered. The test of equality of group means shows if there determined by the size of their standardised regression are statistically significant group differences with respect to coefficients (Ramayah et al. 2010). The model specification the independent variables. Also, because the model is as follows: contained four dummy variables, a hierarchical discriminant Y = a + b X + b X + …. + b X [Eqn 1] analysis was done to know the effect of the dummy variables as ij 1 ik i 2k n nk they cannot be interpreted like other variables in the linear Where Yij is the discriminate Y score for object k’s discriminant regression model. Therefore, a discriminant analysis was function j. A is the intercept, bi is the coefficients for the first done without the dummy variables and then with the independent variable i and X is the independent variable for variables. Thereafter, the difference in the canonical j object k. Specifically, the stepwise discriminant analysis is correlation was computed, and this indicates a joint used. explanatory effect of the dummy variables as a set. In addition, the structure matrix correlations are used in The cutting score for classifying the observations was ranking the variables in order of importance, as it shows the calculated as in Ramayah et al. (2010): correlation of each explanatory variable with the dependent variable, and they are considered more accurate than the Z = NA ZB+ NBZA standardised coefficients. However, both the standardised cs N + N [Eqn 2]A B and unstandardised coefficients indicate the partial contribution of each significant variable, holding other Where Zcs is optimal cutting score, NA and ZA are the number variables constant. of observations and centroid for group A, respectively, while NB and ZB are the number of observations and A household’s vulnerability perception was categorised centroid for group B respectively. The decision rule is that based on the number of stressors indicated, whereby if a observations with discriminant scores less than the cutting household indicated the presence of 30% or less of the total score are classified into group (0), while those with scores stressors, it was rated as having low vulnerability to that are higher are classified together as belonging to group flooding. Those characterised by 40% – 60% of the stressors (1). At each step in the stepwise discriminant analysis, the were grouped as moderately vulnerable, while households variable that minimises the overall Wilk’s Lambda is that indicated between 70% and 100% of the stressors were http://www.jamba.org.za Open Access Page 5 of 10 Original Research perceived to have high level of vulnerability to flood. TABLE 3: Descriptive characteristics of the urban households. Likewise, the flood risk perception of the households was Variables Total Frequency % Mean Minimum Maximum categorised into three categories based on the number of Age of respondents 198 - - 37.95 17 80 flood impact indicators experienced. This assumes that an Household size 199 - - 4.65 1 20 Households with children 199 104 52.3 - 0 12 individual’s perception is shaped by experience, and as less than 5 years such they will act only if a hazard risk had and will make Households with 198 44 22.2 - 0 4 significant impact. Therefore, a household that had dependants older than 60 years experienced 30% or fewer of the possible impacts was Gender 199 - - - 0 1 classified as having low flood risk perception. Those who Male - 122 61.3 - - - had experienced up to 40% – 60% of the listed impacts were Female - 77 38.7 - - - rated as having moderate perception, while those who had Years of residence in the 198 - - 6.89 1 42 area experienced about 70% – 100% of the impacts were rated as Education 199 - - - 0 3 having high flood risk perception. Similarly, based on No formal education - 3 1.5 - - - the number of responses on adoption of precautionary Primary - 16 8.0 - - - measures, the households were categorised into low Secondary - 72 36.2 - - - adaptive capacity and high adaptive capacity groups, using Tertiary - 108 54.3 - - - a cut-off point of five obtained as the midpoint of the 10 Flood risk awareness 185 - - - 0 1 presented adaptive measures requiring a yes or no response. No - 120 64.86 - - - Residents who indicated a yes to fewer than or equal to 5 Yes - 65 35.14 - - - measures were considered a low adaptive capacity group; Source of information on 65 - - - 0 1flood otherwise, they were considered to have high adaptive Media (radio, TV and - 63 96.92 - - - capacity. social media) Nonmedia (friends and - 3 4.62 - - - family) Ethical considerations Disaster preparedness 199 0 1 No 110 55.3 - - Ethical approval to conduct this study was obtained from the Yes 89 44.7 - - University of Port Harcourt Research Ethics Committee at its Experience of past floods 199 0 1 65th meeting held 02 October 2019 (reference number: No - 45 22.7 - - UPH/CEREMAD/REC/MM65/033). Yes - 153 77.3 - - Perception of flood risk 199 - - - 1 3 Results and discussion Low - 79 39.7 - - -Moderate - 69 34.7 - - - Socio-economic characteristics of the urban High - 51 25.6 - - - households Perception on climate 183 - - - 0 1 change causing flood It is observed in Table 3 that the average age of the residents Agree - 74 40.4 - - - was 38 years, and a slight majority (52.3%) had children Disagree - 109 59.6 - - - who were less than 5 years old in their households. A Willingness to purchase 184 - - - 0 1 insurance majority (61%) of the respondents were male and the No - 126 68.5 - - - respondents had on average lived in the communities for 7 Yes - 58 31.5 - - - years, implying that they may have experienced some of the Perceived level of 199 1 3 past and recent incidences of flooding in the area. Also, a vulnerability to flood majority (98.5%) have had one form of education. Low - 15 7.5 - - - % Moderate - 41 20.6 - - -Furthermore, 55.3 of the residents were not disaster High - 143 71.8 - - - prepared, as they have not taken any precautionary actions, Location 183 - - - 0 1 compared to 44.7% that felt otherwise. Many (77.3%) had Floodplains’ settlement 1164 82.83 - - - experienced floods, yet a majority 74.4% had low to Elevated areas 35 17.17 - - - moderate flood risk perception. In contrast to Hoffmann Employment status 183 - - - 0 1 and Muttarak’s (2017) opinion that education and flood Presently employed - 149 74.87 - - - experience could trigger a training process that has the Presently unemployed - 50 25.13 - - - potential to increase preparedness levels, it is seen from the descriptive results that a majority of the respondents (69%), in which the highest proportion of their respondents had despite their literacy and experience of flooding, are not little awareness of floods, and the result also supports the willing to buy insurance, a form of disaster preparedness notion that flooding in Port Harcourt is mostly caused by measure. This could be partly attributed to the observation human factors (Akukwe 2014). Also, a majority (72%) were that a majority (65%, 74% and 60%) of the respondents were grouped as having high vulnerability to flooding given the not aware of the risks posed by flooding, had low perception level of stressors they perceived as present in their of flood risk and did not agree on climate change being a environment. More so, a higher proportion (75%) of the driver of recent occurring floods, respectively. The result is respondents were presently employed and 25% unemployed similar to the findings of Akukwe and Ogbodo (2015), (including students and the retired). http://www.jamba.org.za Open Access Page 6 of 10 Original Research As shown in Table 4, a majority of the responses occurred (74.14%) were WTP insured. The chi-square test was used to under those who were not willing to purchase (WTP) test if these differences were real or due to chance variation. insurance, and a majority (86.51%) of them had either Since the chi-square statistic is less than 0.002, it could be secondary or tertiary education. A lower number of those concluded that the difference in the respondents’ willingness with secondary education (24.14%) were WTP insured just as to purchase insurance is real and not due to chance. However, a much higher proportion of those with tertiary education given that not all respondents had experienced flood in the past, which could have influenced their willingness to TABLE 4a: Cross-tabulation of differences in education and location among purchase insurance, further cross-classification by previous respondents characterised by willingness to purchase insurance. experience of flood was performed. The chi-square Variable Nonwillingness to purchase Willingness to purchase insurance (0) insurance (1) significance of those who had not experienced flood and were n % n % not WTP insured was 0.132, suggesting but not conclusive of Education chance variation between nonwillingness to purchase No formal education 3 2.38 0 insurance and education. While those who had previous Primary 14 11.11 1 1.72 flood experience and WTP insurance had a significance value Secondary 51 40.48 14 24.14 of 0.001, implying that the relationship observed in the cross- Tertiary 58 46.03 43 74.14 tabulation is real and not a result of chance. Similarly, all Total 126 100.00 58 100.00 symmetric measures were significant and their values greater Location than 0.3, implying a strong relationship. Therefore, there is Nkpolu 33 26.19 17 29.31 need for further education and enlightenment on insurance Rumuigbo 19 15.08 17 29.31 Rumuekpirikom 29 23.02 11 18.97 markets or products to encourage more participation in the Rumueme 23 18.25 2 3.45 insurance market. To further reinforce this result, cross- Rumukalagbor 22 17.46 11 18.97 classification of willingness to pay for insurance, location and Total 126 100.00 58 100.00 flood experience was done. It was observed that a greater Note: Pearson chi-square test = 0.002 for education among respondents characterised by proportion of respondents approximately 29%, 29% and 19% willingness to purchase insurance. Pearson chi-square test = 0.026 for location among respondents characterised by willingness to purchase insurance. from floodplain settlements of Nkpolu, Rumuigbo and Rumuekpirikom, respectively, were more willing to pay TABLE 4b: Cross-tabulation of differences in education among respondents for insurance than those in higher-elevation area of characterised by willingness to purchase insurance. Rumukalagbor. More so, a greater share of respondents who Variable Non-willingness to Willingness to purchase had experienced floods and were also more willing to buy purchase insurance (0) insurance (1) Flood experienced Flood experienced insurance came from the floodplain areas of Rumuigbo (37%), No (0) Yes (1) No (0) Yes (1) Nkpolu (33%) and Rumuekpirikom (20%). An analysis of n % n % n % n % variance with respect to flood experience across the locations Education showed two homogeneous subsets, a significant difference No formal education - - 3 2.80 - - 0 - between one location (Rumukalabor) a high-elevation area Primary 3 16.67 11 10.28 0 - 1 2.17 and the other four (Nkpolu, Rumuigbo, Rumueme and Secondary 1 5.56 50 46.73 3 25.00 11 23.91 Rumuekpirikom) established as flood-prone areas with Tertiary 14 77.78 43 40.19 9 75.00 34 73.91 floodplain settlements (Nwankwoala & Jibril 2019). Total 18 - 107 - 12 - 46 - Note: Flood experienced - Pearson chi-square test, Phi, Cramer’s V and Contingency coefficient = 0.132 for education among respondents characterised by non-willingness to Furthermore, nine local stressors that could influence a purchase insurance. community or an individual’s vulnerability to flooding were Pearson chi-square test = 0.001 for education among respondents characterised by willingness to purchase insurance. identified and presented to the respondents to indicate which stressors were perceivably present in their community or had been experienced. TABLE 4c: Cross-tabulation of differences in location among respondents characterised by willingness to purchase insurance. Variable Non-willingness to purchase Willingness to purchase It is seen from the result in Table 5 that poorly constructed or insurance (0) insurance (1) blocked drainage, where available, unplanned building of Flood experienced Flood experienced structures, frequent rains, inadequate drainage systems, poor No (0) Yes (1) No (0) Yes (1) n % n % n % n % environmental sanitation monitoring by government, Location congestion and lack of awareness of government environmental Nkpolu 1 5.56 2 16.67 32 29.91 15 32.61 management practices in the community are major factors Rumuigbo - - - - 18 16.82 17 36.96 perceived to be predisposing the communities to flooding. As Rumuekpirikom 5 27.78 2 16.67 24 22.43 9 19.57 noted in Appleby-Arnold et al. (2018), when people perceive Rumueme 2 11.11 1 8.33 21 19.63 1 2.17 that government authorities are not working effectively, it Rumukalagbor 10 55.56 7 58.33 12 11.21 4 8.70 creates in them resigned attitudes that can also hamper disaster Total 18 - 12 - 107 - 46 - preparedness. The major impacts of flooding suffered by the Note: Flood experienced - Pearson chi-square test, Phi, Cramer’s V and Contingency coefficient = 0.721 for location among respondents characterised by non-willingness to residents included cutting off of electricity supply to houses, purchase insurance. disruption of household activities and loss of property. It can Pearson chi-square test = 0.012 for location among respondents characterised by willingness to purchase insurance. be deduced from the result that the vulnerability factors http://www.jamba.org.za Open Access Page 7 of 10 Original Research TABLE 5: Assessment of urban households’ vulnerability and disaster TABLE 6: Summary of measures in the discriminant analysis (n = 199). preparedness. Independent Unstandardised Standardised Discriminate F-statistics Vulnerability characteristics Frequency % Rank variables coefficients coefficients loadings (rank) Inadequate drainage system 169 91.85 5th Age (x1) -0.06 -0.61 -0.38 (3rd) 68.01* Poorly constructed drainage 182 98.91 1st Risk awareness (x2) 1.70 0.66 0.62 (1st) 74.61* High frequent rains 173 94.02 3rd Perception on flood 0.66 0.46 0.44 (2nd) 90.92* Blocked drainage areas 173 94.02 3rd risk (x3) Household size (x ) 0.09 0.26 0.21 (5th) 54.88* Unplanned building of structures 174 94.57 2nd 4 Location (x5) 0.79 0.28 0.37 (4th) 46.08*Poor maintenance of environment by designated 142 77.20 6th government agencies *, p < 0.001. Congested population in communities 113 61.41 8th Intercept, –0.77; Group centroid (0), –1.104; Group centroid (1), 1.193; Wilk’s Lambda, 0.43*; Test of equality of variance (Box’s M test), 176.90*; Canonical correlation Flooding from small streams whose catchment 72 39.13 9th (with dummies), 0.76; Canonical correlation (without dummies), 0.49; Effect of dummies areas lie within built-up areas taken as a set or whole, 0.27; Overall hit ratio, 87.5%. Lack of awareness of government environment 142 77.17 6th management practices within the community In addition, the number of coping or precautionary Experienced impacts of flooding measures undertaken by the residents to deal with Experience shortage of food during floods 61 38.10 7th Experience shortage of clean drinking water 40 25.00 9th occurring floods was used as a proxy for gauging their during flood adaptive capacity. Frequency analysis of dichotomous Toilet facilities are affected in time of flood 59 37.30 8th responses given on a list of subjective well-being measures Power (electricity) cut when flood occurs 137 88.40 1st presented to the residents to indicate the adaptive Loss of property during flood 101 63.10 2nd measures they had undertaken. As shown in Table 5, Most Suffered body injuries as result of flood 68 44.16 5th (88%) of the surveyed households fell into the low adaptive Loss of a loved one as a result of the flood 17 11.04 10th Flood disrupted household activities 89 57.79 3rd capacity category. These households generally were Household building was damaged as a result 86 55.84 4th associated with lower number of adaptive strategies to of flood mitigate the impacts of floods or respond to future Experience disruption of income generating 77 48.40 6th business because of flood flooding. Coping or adaptive measures Use of mechanical water pumps to remove flood 60 42.55 4th water from home Determinants of flood disaster preparedness Built temporary plank bridges between houses 75 52.08 3rd The summary of stepwise discriminant analysis showing and across roads to move about during flooding Constructed dykes or trenches to divert water 27 19.29 8th the variables influencing the disaster preparedness of a away from the house household is presented in Table 6. It could be seen from the Relocated to highest parts of community that are 55 39.29 6th more secure from flood F-statistics that the variables age, flood risk assessment, Purchased insurance policy to guard against 14 7.57 10th flood risk perception, household size and location were all disaster loss statistically significantly different between those who were Constructed drainages around property 106 58.56 1st disaster prepared and those who were not. In addition, Build walls around building to keep out water 98 54.44 2nd the pooled-within group matrices showed that the Planted vegetation around building to reduce or 21 11.41 9th prevent erosion intercorrelations were low, justifying the use of the Aware of government disaster management 77 41.85 5th independent variables. Four of the five significant variables agency to call on in the occurrence of disaster event have positive coefficients, which means they help to Involved in joint communal effort to combat flood 50 28.57 7th discriminate the households that are disaster prepared Adaptive capacity (they can drive preparedness behaviours), while the age Low adaptive capacity 157 87.71 - variable with a negative sign helps to predict the households High adaptive capacity 22 12.29 - that are not disaster prepared. In other words, households that have higher number of family members, are aware of characterising the households were majorly structural the risks associated with floods, have higher perception of (drainage issues and unplanned constructions), climatic flood risk and live in low-lying floodplain settlements will (excessive rainfall), environmental and public utility–related have the tendency to be disaster prepared against possible (inadequate and poorly maintained environment or drainages). future flooding disaster. However, younger heads of The result agreed with Uddin (2018) that poor implementation of urban plans and housing policies leads to the creation of households will have less tendency to be disaster prepared. slums and shanties lacking in basic facilities and impacting In addition, it can also be inferred that flood risk awareness negatively on human security (e.g. access to food and discriminates the most followed by flood risk perception electricity). Also, as noted by Echendu (2020), poor waste age, location and household size. The result agrees with the management contributes to flooding occurrence. There is no literature that location and risk perception can influence doubt that flooding will greatly impact areas where drainage disaster preparedness behaviours (Hashim et al. 2021; structures are deficient, as seen in many urban areas. However, Najafi et al. 2015). a flood-resilient community can emerge if the aforementioned challenges are addressed, thereby enhancing its ability to The discriminant function (equation) can be obtained from mitigate flood impacts and minimise vulnerability (Chong, the unstandardised coefficients shown in Table 6 and in this Kamarudin & Abd Wahid 2018). case, it is: http://www.jamba.org.za Open Access Page 8 of 10 Original Research TABLE 7: Classification results. a Disaster preparedness (DP) Predicted group membership Predicted group for analysis 1 0 1 Total 25 Non-disaster n % n % n % prepared Reference line Original 20 0 87 90.6 9 9.4 96 100 1 14 15.9 77 84.1 88 100 15 Cross-validated 0 86 89.6 10 10.4 96 100 10 1 17 19.3 71 80.7 88 100 NB, Numbers in parenthesis indicate row percentages. 5   –0.77  0  -4 -2 0 2 4 –0.06 D= X=[1 Age Awareness Floodperception    Canonical discriminant funcon 11.70 Householdsize Location  0 . 6 6   + e Predicted group for analysis 1 b   0.09   25 Disaster Reference line  0.79  prepared [Eqn3] 15 Furthermore, the group centroids (mean of canonical 10 variables) are different for each group. Disaster nonprepared group had a mean of −1.104 while the disaster prepared 5 group was 1.193. The canonical correlation measures the strength of relationship. With a canonical correlation of 0.76, it can be inferred that 58% (square of the correlation) of 0 -4 -2 0 2 4 the variance in the dependent variable is accounted for by this model, and about 42% is unexplained. As indicated Canonical discriminant funcon 1 by the Wilk’s lambda (Table 6), the discriminant function is FIGURE 2: Distribution of discriminant scores for disaster prepared and better than having chance separate the two groups. The nonprepared households. implication of the results is that government and relevant agencies should create more awareness on impending or (coded 0), while cases with scores higher than 0.09 are potential flood disaster and promote public sensitisation on classified under disaster prepared (coded 1). the dangers associated with flooding and the need to prevent or minimise it, especially in flood-prone areas. Conclusion Although the Box’s M significant value was an indication The study examined the vulnerability, flood risk perception of violation of the assumption of equality of covariance and disaster preparedness of urban households in flood- matrices, it was not considered a serious problem as the prone areas, as it becomes increasingly important to build sample was a large one. resilience to the impacts of climate change-related flood. It also determined the factors that propels a household to be The result in Table 7 showed that households that were not disaster prepared or otherwise. Descriptive statistics and disaster prepared were more accurately predicted (90.6%) stepwise discriminant analysis were employed on collected than those that were prepared (84.1%). Generally, on the data. Results showed that most of the households were average, 87.5% (hit ratio) of the original group cases were unaware of the risks associated with floods and those who correctly classified. This implies that overall, in three out of were aware became informed through mainstream media. A four times, the model is correct. The histogram in Figure 2 majority of the participants lived in low-lying floodplain shows the distribution of the discriminant function scores settlements and were employed; a majority had low-medium for each group. It shows how well the function discriminates perception of risk and showed low readiness to deal with a by the overlap and nonoverlap of the graphs. The cutting flooding disaster. It was also observed that more of the score was computed as follows: residents in floodplain settlements experienced flooding and Z 96 (1.193) + 88 (−1.104) were more willing to buy insurance than those living in cs = 96 + 88 elevated areas. Furthermore, given the extent to which they [Eqn 4] 114.528 − 97.142 had taken up coping and/or adaptive measures, a majority Zcs = = 0.09 184 of the households were classified as having low levels of adaptive capacity. Statistical significance of mean differences Thus, observations or cases with discriminate score lower was observed for all five predictors (age, flood risk awareness, than 0.09 are grouped together as non–disaster prepared risk perception, location and household size) out of http://www.jamba.org.za Open Access Frequency Frequency Page 9 of 10 Original Research 15 predictors regressed on the dependent variable. It was Akukwe, T., 2014, ‘Determinants of flooding in Port Harcourt metropolis, Nigeria’, IOSR Journal of Humanities and Social Science 19(2), 64–72. https://doi. shown that the five predictors accounted for 58% of the org/10.9790/0837-191186472 variation between the two groups variability, while the hit Akukwe, T., Krhoda, G. & Oluoko-Odingo, A., 2018, ‘Principal component analysis of ratio indicated that overall 87.5% of the cases were correctly the effects of flooding on food security in agrarian communities of south eastern Nigeria’, International Journal of Hydrology 2(2), 205–212. https://doi. classified. Important implications arising from the study are org/10.15406/ijh.2018.02.00070 that tackling the challenges (e.g. poorly constructed drainage Akukwe, T.I. & Ogbodo, C., 2015, ‘Spatial analysis of vulnerability to flooding in Port Harcourt metropolis, Nigeria’, SAGE Open 5(1), 21582440155. https://doi. and unplanned building of structures) increasing the org/10.1177/2158244015575558 vulnerability of the people or area to flooding will enhance Appleby-Arnold, S., Brockdorff, N., Jakovljev, I. & Zdravković, S., 2018, ‘Applying cultural values to encourage disaster preparedness: Lessons from a low-hazard the adaptive capacity of the households. Also, creating country’, International Journal of Disaster Risk Reduction 31, 37–44. https://doi. org/10.1016/j.ijdrr.2018.04.015 awareness on disaster risk helps to form the individuals’ Barquet, K. & Cumiskey, L., 2018, ‘Using participatory Multi-Criteria Assessments for perception of the risk and, in turn, their disaster preparedness assessing disaster risk reduction measures’, Coastal Engineering 134, 93–102. behaviours and subsequently building resilience. The study https://doi.org/10.1016/j.coastaleng.2017.08.006 Barrios, S., Bertinelli, L. & Strobl, E., 2006, ‘Climatic change and rural-urban migration: supports the need to mitigate flood disaster impacts on The case of sub-Saharan Africa’, Journal of Urban Economics 60(3), 357–371. households and the environment. It suggests an integrated https://doi.org/10.1016/j.jue.2006.04.005 approach that includes protective, preventive and control Baten, A., Arcos González, P. & Delgado, R.C., 2018, ‘Natural disasters and management systems of Bangladesh from 1972 to 2017: Special focus on flood’, OmniScience: A measures by all stakeholders, including government and Multi-disciplinary Journal 8(3), 35–47, viewed 21 September 2021, from www.stmjournals.com. relevant bodies, increasing public sensitisation of flood risk Brooks, N., 2003, Vulnerability, risk and adaptation: A conceptual framework, Working and its attending effects for greater awareness, supporting Paper 38, Tyndall Centre for Climate Change Research, Norwich, viewed 18 June 2022, from https://www.researchgate.net/publication/ 200032746. communities in regular clearing of drainage areas and Caparros-Midwood, D., Barr, S. & Dawson, R., 2015, ‘Optimised spatial planning checking unplanned construction of buildings, particularly to meet long term urban sustainability objectives’, Computers, Environment and Urban Systems 54, 154–164. https://doi.org/10.1016/j.compenvurbsys. in flood-prone areas. 2015.08.003 Chong, N.O., Kamarudin, K.H. & Abd Wahid, S.N., 2018, ‘Framework considerations for Acknowledgements community resilient towards disaster in Malaysia’, Procedia Engineering 212, 165–172. https://doi.org/10.1016/j.proeng.2018.01.022 Competing interests Cologna, V., Bark, R.H. & Paavola, J., 2017, ‘Flood risk perceptions and the UK media: Moving beyond “once in a lifetime” to “Be Prepared” reporting’, Climate Risk Management 17, 1–10. https://doi.org/10.1016/j.crm.2017.04.005 The authors declare that they have no financial or personal Conroy, R.M., 2018, The RCSI sample size handbook: A rough guide, Royal College of relationships that may have inappropriately influenced Surgeons, Ireland, viewed 26 June 2018, from https://doi.org/10.13140/RG.2.2.30497.51043. them in writing this article. De Brauw, A., Mueller, V. & Lee, H.L., 2014, ‘The role of rural-urban migration in the structural transformation of sub-Saharan Africa’, World Development 63, 33–42. https://doi.org/10.1016/j.worlddev.2013.10.013 Authors’ contributions Demeritt, D. & Nobert, S., 2014, ‘Models of best practice in flood risk communication and management’, Environmental Hazards 13(4), 313–328. https://doi.org/10.10 Z.A.E. conceptualised the research, drew up the 80/17477891.2014.924897 methodology, wrote the original draft of the paper and Douglas, I., Alam, K., Maghenda, M., Mcdonnell, Y., Mclean, L. & Campbell, J., 2008, ‘Unjust waters: Climate change, flooding and the urban poor in Africa’, participated in field survey. O.L. validated and helped with Environment and Urbanization 20(1), 187–205. https://doi.org/10.1177/ 0956247808089156 data curation, employed software to enhance analysis Duh, J.-D., Shandas, V., Chang, H. & George, L.A., 2008, ‘Rates of urbanisation and the and reviewed and edited the final draft of manuscript. resiliency of air and water quality’, Science of the Total Environment 400(1–2), 238–256. https://doi.org/10.1016/j.scitotenv.2008.05.002 Echendu, A.J., 2020, ‘The impact of flooding on Nigeria’s sustainable development Funding information goals (SDGs)’, Ecosystem Health and Sustainability 6(1), https://doi.org/10.1080/20964129.2020.1791735 This research received no specific grant from any funding Eiser, J.R., Bostrom, A., Burton, I., Johnston, D.M., McClure, J., Paton, D. et al., 2012, ‘Risk interpretation and action: A conceptual framework for responses to natural agency in the public, commercial or not-for-profit sectors. hazards’, International Journal of Disaster Risk Reduction 1(1), 5–16. https://doi. org/10.1016/j.ijdrr.2012.05.002 Etinay, N., Egbu, C. & Murray, V., 2018, ‘Building urban resilience for disaster risk Data availability management and disaster risk reduction’, Procedia Engineering 212, 575–582. https://doi.org/10.1016/j.proeng.2018.01.074 The data that support the findings of this study can by Federal Republic of Nigeria (FGN), 2011, Annual abstract of statistics, viewed 14 June 2019, from www.nigerianstat.gov.ng. made available by the corresponding author, Z.A.E., upon Gerald-Ugwu, G.C., Egolum, C.C. & Emoh, F.I., 2019, ‘An investigation of factors reasonable request. accelerating rise in building in the flood prone areas of Port Harcourt’, Iconic Research and Engineering Journals 3(1), 375–384. Hashim, H.M., Ng, Y.G., Talib, O. & Md Tamrin, S.B., 2021, ‘Factors influencing flood Disclaimer disaster preparedness initiatives among small and medium enterprises located at flood-prone area’, International Journal of Disaster Risk Reduction 60, 102302. https://doi.org/10.1016/j.ijdrr.2021.102302 The views expressed in this article are those of the authors Hoffmann, R. & Muttarak, R., 2017, ‘Learn from the past, prepare for the future: and not of the University of Port Harcourt or the publisher. Impacts of education and experience on disaster preparedness in the Philippines and Thailand’, World Development 96, 32–51. https://doi.org/10.1016/j. worlddev.2017.02.016 References Intergovernmental Panel on Climate Change (IPCC), 2007, Climate change 2007: Impacts, adaptation and vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, in Abass, K., Dumedah, G., Frempong, F., Muntaka, A.S., Appiah, D.O., Garsonu, E.K. M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. Van der Linden & C.E. Hanson (eds.), et al., 2022, ‘Rising incidence and risks of floods in urban Ghana: Is climate change p. 976, Cambridge University Press, Cambridge, UK. to blame?’, Cities 121, 1–41. https://doi.org/10.1016/j.cities.2021.103495 Kwak, Y., Muraoka, K. & Asai, K., 2018, ‘Community-based nationwide capacity building for Aderogba, K.A., 2012, ‘Qualitative studies of recent floods and sustainable growth and GIS-based integrated flood management’, in T. Tanzi, M. Chandra, O. Altan & F. Sunar development of cities and towns in Nigeria’, International Journal of Academic (eds.), International archives of the photogrammetry, remote sensing and spatial Research in Economics and Management Sciences 1(3), 1–25. https://doi. information sciences – ISPRS Archives, 06 March 2018, pp. 313–318, International org/10.17142/ijbas-2012.1.2.4 Society for Photogrammetry and Remote Sensing, Istanbul, Turkey. http://www.jamba.org.za Open Access Page 10 of 10 Original Research Lawal, O. & Arokoyu, S.B., 2015, ‘Modelling social vulnerability in sub-Saharan West Potts, D., 2015, ‘Urbanization in Africa’, International Encyclopedia of Social & Africa using a geographical information system’, Jàmbá: Journal of Disaster Risk Behavioral Sciences 24, 965–972. https://doi.org/10.1016/B978-0-08-097086- Studies 7(1), 155. https://doi.org/10.4102/jamba.v7i1.155 8.72096-4 Lawal, O. & Umeuduji, J.E., 2017, ‘Exploration of hydro-geomorphological indices for Rakib, M.A., Islam, S., Nikolaos, I., Bodrud-Doza, M. & Bhuiyan, M.A.H., 2017, ‘Flood coastal floodplain characterization in Rivers State, Nigeria’, Ghana Journal of vulnerability, local perception and gender role judgment using multivariate Geography 9(1), 67–87, viewed 23 September 2021, from https://www.ajol.info/ analysis: A problem-based “participatory action to Future Skill Management” to index.php/gjg/article/view/154654. cope with flood impacts’, Weather and Climate Extremes 18, 29–43. https://doi. Mmom, P.C. & Iluyemi, P.O., 2018, ‘Disaster risk of property development on org/10.1016/j.wace.2017.10.002 floodplains in Port Harcourt, Nigeria’, in IAIA18 Conference Proceedings/ Ramayah, T., Ahmad, N.H., Abdul-Halim, H., Zainal, S.R.M. & Lo, M.-C., 2010, Enironmental Justice in Societies in Transition, 38th Annual Conference of the International Association for Impact Assessment, Durban Internationl Convention ‘Discriminant analysis: An illustrated example’, African Journal of Business Center, Durban, May 16–19, 2018, viewed 22 June 2022, from www.iaia.org. Management 4(9), 1654–1667. Najafi, M., Ardalan, A., Akbarisari, A., Noorbala, A.A. & Jabbari, H., 2015, ‘Demographic Rollason, E., Bracken, L.J., Hardy, R.J. & Large, A.R.G., 2018, ‘Rethinking flood risk determinants of disaster preparedness behaviors amongst Tehran inhabitants, communication’, Natural Hazards 92(3), 1665–1686. https://doi.org/10.1007/ Iran’, PLOS Currents Disasters 7. https://doi.org/10.1371/currents.dis.976b0ab9c s11069-018-3273-4 9d9941cbbae3775a6c5fbe6 Santoro, S., Pluchinotta, I., Pagano, A., Pengal, P., Cokan, B. & Giordano, R., 2019, Netzel, L.M., Heldt, S., Engler, S. & Denecke, M., 2021, ‘The importance of public risk ‘Assessing stakeholders’ risk perception to promote Nature Based Solutions as perception for the effective management of pluvial floods in urban areas: A case flood protection strategies: The case of the Glinščica river (Slovenia)’, Science study from Germany’, Journal of Flood Risk Management 14(2), 1–22. https://doi. of the Total Environment 655, 188–201. https://doi.org/10.1016/j.scitotenv. org/10.1111/jfr3.12688 2018.11.116 Nur, I. & Shrestha, K.K., 2017, ‘An integrative perspective on community vulnerability Tingsanchali, T., 2012, ‘Urban flood disaster management’, Procedia Engineering 32, to flooding in cities of developing countries’, Procedia Engineering 198, 958–967. 25–37. https://doi.org/10.1016/j.proeng.2012.01.1233 https://doi.org/10.1016/j.proeng.2017.07.141 Uddin, N., 2018, ‘Assessing urban sustainability of slum settlements in Bangladesh: Nwankwoala, H.O. & Jibril, T., 2019, ‘Flood sensitivity analysis and impact assessment Evidence from Chittagong city’, Journal of Urban Management 7(1), 32–42. in parts of Obio/Akpor Local Government Area, Rivers State’, World Applied https://doi.org/10.1016/j.jum.2018.03.002 Sciences Journal 37(9), 716–741. https://doi.org/10.5829/idosi.wasj.2019.716.741 United Nations, 2016, Report of the open-ended intergovernmental expert working Obinna, V.C., Owei, O.B. & Mark, E., 2010, ‘Informal settlements of Port Harcourt and group on indicators and terminology relating to disaster risk reduction, viewed 26 potentials for planned city expansion’, Environmental Research Journal 4(3), June 2019, from http://www.preventionweb.net/drr-framework/open-ended- 222–228. https://doi.org/10.3923/erj.2010.222.228 working-group/. http://www.jamba.org.za Open Access