Flood disasters are growing in frequency worldwide due to a variety of environmental and human factors. Natural disasters and floods in particular are becoming more frequent and destructive (Garcia-Castellanos et al. 2009; Li et al. 2014; Miller 1997; Smith 2001).

Physical causes of floods include the nature of precipitation, topography, vegetation, soil type and runoff patterns whilst human factors contributing to flooding are mostly associated with development and land use (Khan 2004; Smith 2001). Over the last 50, years anthropogenic factors have played a determining role in causing flood disasters (Clements 2009). The impact of flooding includes destruction of and damage to roads, bridges, buildings or sewage-disposal systems. Flood disasters affect agricultural production, bringing about food insecurity in communities (FAO–SAFR 2002).

Governments in southern Africa are devising mechanisms to cope with the impact of flooding, but these efforts are often limited and are affected by people’s perceptions which in turn influence their responses. Communities often resist relocation as flooding is an essential component of agriculture, and the ecological systems provide the basis for the regeneration of crops, plants and aquatic life (Wisner et al. 2004). A report by the Limpopo Department of Agriculture (2002) indicates that apartheid created conditions that were unfavourable to the welfare of rural Black people. This historical legacy led to the present situation where the Black majority stay in poor marginal land that is inadequate for subsistence farming and prone to flooding.

Thulamela municipality is vulnerable to hazards such as diseases, fires, drought, thunderstorms and flooding. This is associated with high levels of poverty, landlessness and poor infrastructure. The physiography of the area and the location of settlements render Thulamela communities prone to flooding. Before and after the 1999–2000 flood, disaster management for floods was found to be lacking and ineffective. This was the first major flood disaster, caused by tropical cyclone Eline, but the municipality was completely unprepared according to Mudinda (2010). Ten years later, flooding became a major disaster once again. This last flood (2010–2011) led to this study in the two villages of Maniini and Tshilungwi in the Limpopo Province. Figure 1 indicates the location of the Vhembe District Municipality and areas that are vulnerable to floods.

Research on the impacts of and responses to flooding have been undertaken at national, provincial and city level and focused on the causes and impact of floods (Mudau 2001; Sengani 2008). Few studies have been done at local level in the rural areas of South Africa. There are no studies on the impact of flood disasters in these two villages (Maniini and Tshilungwi), this despite the high level of vulnerability.

Thulamela Municipality is one of the four local municipalities in the Vhembe District of Limpopo (Figure 2). The major town in this municipality is Thohoyandou. Thulamela’s population is 618 462, 54.4% of whom are women (Statistics South Africa/Census 2011). The municipality is situated in the eastern subtropical region of the province, and it is generally hot and humid. It receives much of its rainfall during summer, from November to March, as the inter-tropical convergence zone (ITCZ) moves south (Kabanda 2004). The Maniini and Tshilungwi study sites are characterised by a dry season in the months of March to September. The average temperature ranges from 35 °C in summer to 18 °C in winter (Nethengwe 2007). The geology and soil of the sites consist of fertile clay soil, sandy soil and clay loam soil. The land use of the municipality includes residential, subsistence agriculture and commercial uses in Thohoyandou town (Figure 2). Informal businesses within and around the town plays a significant role in the people’s economic and social life.

We purposively sampled these villages because of experiences with flooding and community reports of the negative impact of the flood disaster of January–February 2011.

Most research on the impact of flooding is done at national and city levels, and few studies are done at the local level in rural areas in South Africa.

Various local-government units played a vital role in the recovery and rehabilitation measures as part of the response after the disaster had taken place. Food parcels, tents and other necessary commodities were supplied to the affected communities.

Local businesses also contributed to disaster-relief funds. However, all the effort made by the Municipality and other local-government units was considered to be temporary and short-term. This is because they mainly focused on the recovery and rehabilitation measures with little focus on disaster preparedness, mitigation and risk management. During the household survey, participants were asked about the role of government units before, during and after flood disasters. In response, 80% of Tshilungwi and 66% of Maniini respondents indicated that local-government units were actively involved only after the disaster. This is because the municipality lacked the capacity for the implementation of long-term strategic planning. Preparedness and preventive measures before the disaster were not adequately implemented as indicated by officials.

We utilised a household survey in examining the socio-economic status and coping mechanisms utilised to mitigate the impact of the flood. Particular attention was paid to socio-economic characteristics such as the age of respondents, their level of education, their type of house and the income and household size because these are important indicators for the assessment of a household’s vulnerability to a flood.

Survey results showed that households with a low annual income were the most vulnerable and affected more by floods (Figure 7). This is because coping with a flood disaster demands financial reserves that can buffer the household against the negative impact. Furthermore, the findings also indicate that households with higher levels of education had higher annual income and thus coped better than low-income households. This indicates the importance of education in vulnerability assessment.

Household incomes are divided into four different groups from the poorest, lower income to the highest as indicated in Figure 7. The specific categories for income per month are as follows: lower income (R500 and below), low income (R5001 – R10 000), middle income (R10 001 – R20 000) and high income (R20 001+). Results show that the majority of respondents from both villages (Tshilungwi 52% and Maniini 35%) fall in the lower-income category whilst only 4% of respondents in Tshilungwi and 16% in Maniini fall within the high-income category. One of the respondents in the low-income category noted:

The reason for increased vulnerability might have been the respondents’ level of education as the majority of respondents (Tshilungwi 60% and Maniini 63%) had only a secondary education whilst 8% and 17% respectively had tertiary education. Respondents with tertiary education have a better chance of obtaining a good jobs that pays well, consequently reducing their vulnerability.

We asked the heads of households to identify the adaptation strategies employed during the flood and how the local-government units addressed the impact of the disaster. We also asked households to choose one key adaptation strategy that they would use in a future flood. Finally, we assessed the effectiveness of all strategies and ranked them from the most preferred to the least preferred (Table 2).

Strategies identified for coping with future disasters were ranked using a multiple-response ranking of the strategies that households used to cope with the flood, from the most preferred to the least preferred (Table 2). Survey results from both villages indicate that most people are prepared to take precautionary measures against future threats from climate-related disasters. Their strategies are summarised in Table 2.

In Maniini Village, covering or protecting a housing unit to avoid destruction whilst not moving from their residence was the most preferred alternative. In order not to have to move, community members built levees and terraces, and only if this was not adequate, they would be willing to move to a place of safety. Only when these two efforts fail would they seek help from insurance or disaster aid. It is surprising that asking help from government, borrowing money or doing nothing ranked low in their adaptive strategies.

At Tshilungwi Village, evacuation to a place of safety was considered the first priority, indicating a higher vulnerability than in Maniini. Only if evacuation was not possible or delayed would they seek to protect their houses by building levees and terraces and seek help from neighbours, friends or family and borrow money. Seeking disaster insurance and restructuring or covering the houses did not seem feasible in this village.

In both cases, community members have a negative perception of local-government assistance as a way of coping with floods. A Tshilungwi resident believes that flooding is getting worse from year to year:

The vulnerability of communities to climate-related disasters is increasing, especially in the case of poor, rural communities living at the foot of mountains and in flood plains. Mechanisms to cope with flood disasters vary in scope and magnitude, depending on the impact of events and on the vulnerability of the community and households to such disasters (Canesio 2010). Our study, like that conducted by Mudinda (2010) on the flood disaster of 1999–2000, shows that Thulamela remains highly vulnerable to flooding, and hence, there is still need for improved and well-coordinated disaster risk management and improved risk-reduction strategies.

Households and communities in Thulamela were tremendously vulnerable during the flood disaster, and most damage was to public infrastructure. Apart from the physical impact, the flood disaster also had a significant impact on the households’ welfare, their socio-economic conditions and the livelihood systems of the communities. These findings confirm the findings of other studies in southern Africa where changing climatic conditions are a serious threat to communities (Motsholapheko, Kgathi & Vanderpost 2011).

The perceptions of households that had experienced flooding indicated a high vulnerability to the risk and threat of flooding, and their socio-economic status influenced their level of vulnerability. The unemployed, with low levels of education and low annual income, also tended to be more vulnerable. Our findings are similar to what Canesio (2010) found in Ormoc and Cabalian Bay in the Philippines concerning the experiences of households and communities. Rehabilitation and disaster-relief operations were the most common activities undertaken by municipalities immediately after the disaster. As such, disaster management by the municipality is limited to reacting after the disaster, and few early-warning systems were implemented. The most preferred option, namely temporary relocation or evacuation, might have been influenced by a lack of awareness programs that provide early-warning systems, hence people were neither aware nor prepared for the flood disaster. Cooperative mechanisms by communities are necessary to combat the impact of floods as most coping strategies were implemented by households. This is contrary to Wisner et al. (2004) who notes that effective strategies for preparedness and coping are required before, during and after the event.

Whilst predictions concerning climate change and its effects are made at a large spatial scale, there is a need to view their ramifications at local levels. Hence, this study has contributed to our understanding of the disastrous impact of floods and community perceptions at the local, rural-village level, taking into consideration the views of the communities and the responses by the local municipality. The study recommends the need for vulnerability and risk analyses, a focus on risk reduction, the identification of long-term adaptation strategies, early-warning systems and the introduction of disaster risk studies at all levels of education.