The main objective of the study was to assess the technical suitability of adaptation strategies toward flooding in Sangkrah and Keko Machungwa's informal settlements, with a view to improve adaptation strategies at the household level. Studying technical suitability was central for ensuring that adaptation strategies were reliable in protecting people and properties against flooding. However, before assessing the technical suitability of adaptation strategies, it was important to explore available adaptation strategies. Another aim was to enhance awareness amongst residents, public organisations and the governments of both countries in terms of viable options for mitigating flood-related risks and improving adaptation strategies in urban informal settlements.

The study primarily used qualitative rather than quantitative assessment methods; these consisted of household interviews, physical observations (including visual inspections of signs of damage and deterioration, e.g. cracks, dampness, etc.), mapping, photographing and in-depth interviews with selected respondents. Household interviews were held with flood victims to identify available adaptation strategies. The interviews were also meant to establish whether or not technical factors (in terms of use of experts, flood damage-resistant materials, standard measurements and proper maintenance) were incorporated in the design, implementation and management of flood protection structures.

The only quantitative assessment used was for taking simple measurements; results were then correlated with the BFE. Measurements of structural height of flood control structures were performed for retaining walls, plinths, raised pit-latrines, raised foundations and raised stairs. Because informal settlements have been frequently affected by flooding, one would have expected guidelines to be in place for flood deterrence that specified building materials, BFEs, building designs and so forth. Unfortunately, there were no national guidelines in Tanzania or Indonesia on the design and material requirements for buildings located in flood hazard areas and, as such, FEMA (2008) guidelines were used. However, it was acknowledged that FEMA guidelines might not be conducive for application in Indonesia and Tanzania considering the lower level of technical development compared to the USA.

Five criteria were developed in order to identify and select cases that were suitable for conducting the study; these were, (1) existence of informal housing development, (2) a settlement traversed by a river and thus susceptible to riverine flooding, (3) a settlement that experienced the problem of intensified flooding over time, (4) residents who demonstrated effective responses for containing flooding and (5) a settlement which needed critical intervention in terms of flood control measures. All informal settlements in Surakarta and Dar es Salaam were subjected to evaluation based on the aforementioned criteria and Sangkrah and Keko Machungwa informal settlements were chosen as potential in-depth case studies.

The word sangkrah in the Javanese language means any kind of materials deposited in the local depression, usually rubbish or garbage and might also include wood transported by water, usually during flood events. Therefore, geomorphologically, sangkrah literally means ‘floodplain’. Sangkrah is located in Surakarta City, within the Bengawan Solo River basin (SMERU Research Institute 2011), which is the longest and the largest river basin on Java Island, draining approximately 16 000 km² of the island (Takeuchi, Jayawardena & Takahasi 1995). The Sangkrah informal settlement is traversed by the Bengawan Solo River (Figure 1) and, thus, is located within the river's floodplain, making it a suitable study site.

On the other hand, Keko Machungwa is located in Dar es Salaam City in Tanzania. It is located within a valley of a seasonal stream (Figure 2) and is characterised as a wetland receiving storm water drainage from different elevated parts, eventually discharging into the Indian Ocean. The 1968 Dar es Salaam master plan (United Republic of Tanzania 1968) designated Keko Machungwa as a hazardous area in the sense that it was frequently waterlogged and, hence, not suitable for a human settlement development (Sakijege et al. 2012). It is important to note that informal settlement began at Keko Machungwa in 1960 – 8 years before the Dar es Salaam master plan became effective. Despite zoning of the area as hazardous, informal settlement development has continued. This situation has led to a concentration of buildings, which, in turn, increases the amount of surface water runoff and, consequently, increases the potential for flooding and associated impacts.

Owing to their location (in low-lying areas) and the nature of development taking place (informal), the two selected study areas are, therefore, affected by riverine and urban or local flooding, respectively.

This study employed qualitative procedures for establishing the appropriate number of households to be interviewed. The researcher determined that an appropriate sample size of 70 households in each location would be more than adequate for achieving the goals of the study. The researcher also applied a quantitative procedure involving a mathematical formula by Baradyana and Ame (2005) to establish whether or not the determined sample size was adequate and representative.

As stated above, Sangkrah lies within the Bengawan Solo River basin, which frequently experiences flooding caused by long periods of rains. The precipitation potential in the basin is therefore high and, when coupled with an inadequate drainage system, increases the risk of flooding. The Bengawan Solo River basin has a precipitation potential of 2100 mm per year, of which close to 1040 mm remains as overland flow. This has been one of the causes of floods in the basin, in general, and in the Sangkrah informal settlement in particular. The most severe floods recorded in the Bengawan Solo River basin occurred in 1966 and 2007, with the 1966 flood inundating an area of 142 000 ha (Network of Asian River Basin Organizations 2007) and the 2007 flood causing estimated damages of $200 million (Hidayat 2009). The Sangkrah informal settlement was one of the villages affected by both of those floods. During extreme rainfall, water depth can reach 4 m in Sangkrah; in these cases, the existing flood control infrastructures become overloaded and fail to protect people and properties. Survey results from this study showed that about 1415 houses in Sangkrah are located within the flood-prone area and, thus, are frequently affected.

Keko Machungwa has been affected by frequent flooding because of its location in a river valley and the absence of robust flood control infrastructures. The worst flooding experienced in Keko Machungwa and Dar es Salaam City, in general, was the flood event of December 2011, in which Government authorities (United Republic of Tanzania [URT] 2012) confirmed deaths of 40 people. It was also reported that over 1000 families were displaced and their livelihoods impacted during the 2011 floods (URT 2012). Keko Machungwa was amongst the Dar es Salaam settlements most affected by floods in 2011 and, in some cases, the settlement experienced the effects from flooding even if it did not rain in that area. Such flooding was caused by surface runoff from surrounding elevated areas. In extreme events, water depth normally reaches 1.5 m. As a result, surface runoff takes time to dry because of the nature of the soil which does not allow smooth infiltration. The findings of this study revealed that about 150 houses in the settlement are frequently affected by flooding.

Presently, as the problem of flooding continues to increase, flood victims choose to apply various adaptation strategies to minimise flood impact. Across the two cases, it was revealed that the adaptation strategies ranged from structural to non-structural strategies. The adaptation strategies applied by flood victims are summarised in Figure 3.

The main finding in both cases was observation of the limited use of experts for structural design, choice of building materials and construction techniques for specific adaptation strategies. Financial constraints were reported as a major factor that contributed to this situation. Instead of using experts, interviewed households reported that they used untrained local artisans, which was linked to the fact that their skills were relatively inexpensive and easily available within informal settlements. This was revealed by 82% and 89% of the interviewees in Sangkrah and Keko Machungwa settlements, respectively. One of the interviewees in Sangkrah reported that, ‘I can't afford the cost of hiring an engineer, I use our boys. They are so helpful’. Unfortunately, observations also revealed that flood protection structures and sometimes even entire buildings were destroyed by floods (Figure 4).

According to the flood histories of the two settlements, during extreme flood events, flood water normally reaches a depth of 4.0 m in Sangkrah and 1.5 m in Keko Machungwa. According to FEMA (2008), the water depth of the previous flood is considered the BFE and, thus, any structure for flood mitigation must be constructed one foot (approx. 0.3 m) higher than the BFE. Any flood protection structure constructed below the required BFE will not deter floods and, thus, is considered technically not suitable. Quantitative measurements from the two cases showed that all protective structures were constructed below the BFE (Figure 5). For example, the height of retaining walls and raised stairs in Sangkrah were lower than BFE by 2.8 m and 3.8 m, respectively. In Keko Machungwa, the average height of the retaining walls and raised stairs were lower than the BFE by 0.8 m and 1.3 m, respectively.

It was observed that there were variations in terms of building materials used to construct various parts (foundation, floor, wall and roof) of buildings which impacted adaptation strategies. For example, 88.6% of interviewees in Sangkrah reported that building materials used for construction of house foundations consisted of soil bricks with cement mortar. On the other hand, 78.8% of interviewees reported that materials widely used to build house foundations in Keko Machungwa were sand and cement blocks. In both cases, the same construction materials used to build home foundations were also chosen by residents for flood protection structures. According to FEMA (2008), all these materials are vulnerable to floods when there is a prolonged flood event.

Another observed area of variation from FEMA guidelines (2008) was technical suitability of materials used to construct home floors in the two cases. It was observed in Sangkrah that the majority of interviewees (72.9%) constructed their home flooring using ceramic materials, whereas in Keko Machungwa, about 81.4% stated that most were built with sand and cement flooring materials (see Table 1). This was a deviation from the recommended use of waterproof cement and ceramic materials in the construction of home flooring by FEMA (2008). Inadequately constructed sand and cement floors in Keko Machungwa increased the vulnerability of buildings as it allowed underground water to percolate into houses. This infiltration raised water depths and dampness levels in walls (Figure 6). About 52.9% of houses in Keko Machungwa were reported to be affected by water infiltration problems.

The use of ceramic flooring (72.9%, see Table 1) as observed in Sangkrah may be considered to be reasonable as far as technical requirements were concerned; however, it did not necessarily mean that buildings with such floors were resistant to flooding. Buildings should be constructed using appropriate materials, as well as taking into consideration all other necessary technical aspects. This was not the case in Sangkrah, where it was learned that, although most homes had ceramic floors, their walls were made of soil bricks and mud mortar which were not water-resistant building materials. Walls built with such materials are generally weak and not able to resist the hydrodynamic forces of flood waters. The wall-building materials used in the two cases did not conform to FEMA's guidelines, as Table 1 shows. However, overall, Sangkrah households did comply more with FEMA (2008) guidelines than Keko Machungwa, as Table 1 illustrates, in that 4.3% of Sangkrah interviewees used reinforced concrete to construct the foundations of their buildings, whilst 72.9% used ceramics to construct floors. According to FEMA (2008), reinforced concrete and ceramic are flood-resistant materials.

When maintenance of flood protection structures was examined, both cases revealed the lack of maintenance as reported by 72% and 79% of interviewees in Sangkrah and Keko Machungwa, respectively. This practice also indicated the disregard of factors meant to deliver technically suitable adaptation strategies. Thus, the absence of maintenance has led to weakening of flood protection structures, rendering them vulnerable to failure during floods.

As mentioned earlier, engineering aspects of structural adaptation strategies were not covered in this study. Therefore, it would be worthwhile to assess the robustness of the flood protection devices by conducting tests on characteristics of building materials and ratios used amongst other engineering aspects. Furthermore, efforts to deliver technically suitable adaptation strategies would require addressing various key issues, as elaborated hereunder.

Firstly, although the study has shown that the adaptation strategies at household level were technically unsuitable, the fact cannot be denied that households in flood-prone informal settlements did initiate efforts to establish adaptation strategies to reduce the impacts of flooding. As a way forward, the initiatives at household level need to be enhanced. In order to fully transform the reactive mindset into one that reduces risk and prevents impacts, continuous investment is required to enhance awareness. Government, in collaboration with other key stakeholders such as non-governmental organisations and development partners, can accomplish this through establishment of social learning process.

Secondly, the construction of structural adaptation strategies should be regulated and controlled to follow current procedures for infrastructure construction. That is, following a procedure that involves, (1) application of building permits to relevant authorities, (2) scrutiny of design drawings, (3) site inspections by qualified and authorised personnel, (4) issuing of construction consent and (5) monitoring the construction process to ensure that the prerequisite technical aspects are considered and implemented. In addition, there is a need to strengthen workforces through training of competent workforces so as to increase the level of awareness on disaster reduction issues.

Thirdly, existing urban land development control should be enforced to prevent human settlements development and related investments in flood-prone areas. In this case, urban planning laws, regulations and guidelines could be used to direct wise use of flood-prone areas. For reducing vulnerability of properties and people living in risk areas, there is a need to embed management of risky areas embedded into disaster reduction and environmental management laws. In this case, provisions of the Law Concerning Disaster Management (Law No. 24 of 2004) (Republic of Indonesia 2004) and Law Concerning Spatial Management (Law No. 26 of 2007) (Republic of Indonesia 2004) and the Environmental Management Act, 2004 (Act No. 20 of 2004) (United Republic of Tanzania 2004) should be effectively enforced by relevant authorities.