A qualitative analysis of cascading effects of critical infrastructure service failure post torrential floods in formal & informal settlement: the study-case of Medellin city, Colombia

ABSTRACT

Critical infrastructure (CI) services are crucial to the functioning of modern communities, and a single CI failure can cause cascading effects through a network of services, impacting an entire city. Challenges such as climate change and rapid urbanisation are increasing the risk of cascading effects of CI failure, particularly impacting the urban poor in peri-urban and informal settlements. This study uses a systems thinking approach to validate the hypothesis if CI services are susceptible to generating severe cascading effects in informal than formal settlements. Qualitative data was collected on cascading effects via interviews and focus group discussion in a formal and an informal settlement in Medellin, Colombia. The responses were analysed and structured using a service chain framework and a Causal Loop Diagram. Our findings show that cascading effects vary between the formal and informal contexts concerning CI service availability and accessibility and may be attributed to differences in demography, geography, socio-economic development, risk regulations, community coping capacities and other factors associated with urban development.

KEYWORDS:

• Critical infrastructure services
• cascading effects
• torrential rainfall/landslides
• informal settlements

1. Introduction

Urban critical infrastructure (CI) such as transportation, energy, water, wastewater, public health, and communication systems are, by definition, crucial to the socio-economic functioning of modern communities. CIs can be conceptualized as links in the built environment that connect the physical, functional and organizational attributes (Pescaroli & Alexander, 2016a). A single failure in a CI service system can have cascading effects (Zimmerman & Restrepo, 2009) and affect an entire network of services and, thereby, even an entire city (Serre & Heinzlef, 2018). CI systems’ complexity and integral interdependencies (Petit & Verner, 2001; Zimmerman & Restrepo, 2009) further exacerbate the losses caused by a natural hazard uncertain in time, location and magnitude. Consequently, varying socio-economic impacts of natural hazard events amplify the cascading effects of disrupted CI services on the affected population (Gill & Malamud, 2016; Porio, 2011). Multiple studies on cascading effects have stressed the importance of assessing types of interdependencies and dependencies of CI services, such as organisational, physical, social and technical (Nijs, 2015; O’Sullivan, Kuziemsky, Toal-Sullivan, et al., 2013; Pescaroli & Kelman, 2017).

In urban areas, the spatial propagation of cascading effects is likely to have severe effects (Petit & Verner, 2001). The urban poor, living in peri-urban areas and informal settlements in cities of the global South, are often exposed to the high impacts of potential hazards (Woolf, Twigg, Parikh, et al., 2016). They often endure complex challenges such as land insecurity, faulty shelters in high-risk areas, high population density, violence, inadequate sanitation, safe drinking water, poor hygiene practices and a general lack of economic resilience (Songsore, 2017; United Nations, 2015a). For example, research in Manila, the Philippines, showed how a typhoon generated cascading effects, disrupting water and sanitation services and leading to substantial and disproportionately large, long-term effects on informal settlers (Purwar, Sliuzas, & Flacke, 2020).

However, assessing the cascading effects of disrupted CI services in informal settlements and peri-urban areas is also dependent on its recognition by the government as a legal/illegal settlement (Chatterjee, 2010; Netto, Brigatti, Meirelles, et al., 2018). The recognition status determines CI service provision policies, laws and regulations (UN-HABITAT, 2016) that may or may not include legal access to services. A consequence may be low or negligible access to services, illegal connections, chaos and conflict over service access within the informal/illegal settlement (UN-HABITAT, 2003). Formal settlements, recognised by the government, have legal access to CI services with subsidies for different income groups (Heikkila & Lin, 2014). Thus, different CI service regulations and policies, availability and accessibility of services may influence cascading effects due to CI service failure in formal or informal settlements. The presence of CI services in a formal settlement creates scope to restore disrupted services after a disaster and could be seen as part of a community’s coping capacity (Murray, 2017). However, the complex nature of an informal settlement’s living environment and its CI deficiencies (Songsore, 2017) may worsen the impact of the disaster on an informal community. Thus, we hypothesize that CI services are more susceptible to failure and likely to generate severe cascading effects in informal settlements than in formal settlements.

However, ‘Capacity (or the lack of it) is central to reducing disaster risk and requires an all-of-society engagement and partnership’ (United Nations, 2015b). At the household level, capacities are often internal (or endogenous) to communities, meaning that people have more control over them to take action with available resources and skills (Marschiavelli, Hadi, McCall, et al., 2008). However, at the community level, resources and skills to cope may vary with the overall capacity to take actions for the benefit of the community in the wake of a disastrous event (Marschiavelli, Hadi, McCall, et al., 2008) influenced by the legal status of the settlement, income, access to CI services, housing, health, livelihood, and assets and the government support (Woolf, Twigg, Parikh, et al., 2016). Community participation is recognised as an effective method to enhance capacity for disaster risk reduction as an inherent part of city development plans or community-based disaster management (Twigg, 2004). Therefore, to explore potential or complement existing interventions to address cascading effects, it is important to understand the capacity of the community and the government plans and strategies to cope with cascading effects both in formal and informal settlements.

Various mathematical and statistical models have been used to assess the cascading effects of CI services and forecast the cause and effect of disrupted/failed interdependencies either within a web of networks or a single CI service (Arvidsson, 2015). Much research emphasises the economic and technical aspects of potential ways to reduce cascading effects on physical infrastructure (Hilly et al., 2018). However, there is less attention on how CI services interact with each other and with actors who co-exist or interact in urban surroundings, such as utility companies, government agencies, and communities. This re-enforces the need for a better understanding of disrupted interdependencies and interconnectedness between specific infrastructure services and after-effects on the community.

This paper compares the cascading effects of CI service failure caused by torrential flood/landslides on communities residing in formal or informal settlements in Medellin city, Colombia. The comparative analysis explores how formal and informal communities deal with cascading effects and whether and how each community is supported by government plans to respond to natural hazard events to reduce the impact and future risks of cascading effects. The paper is structured as follows: Section 1 introduces the concepts and significance of the study; section 2 presents the research methodology applied, combining a variety of qualitative methods for data collection and analysis and sets the context for the study area. Section 3 presents results and discussion on the identification of cascading effects, critical/key elements of the CI service chain, consequences of cascading effects on the community of both settlements, and local government support to reduce cascading effects. Other factors that may have a role in reducing or exacerbating cascading effects are also discussed. Finally, section 4 provides our conclusions and further possible research directions.

2. Research methods

The study adopts a systems thinking approach (Casella, Van Tongeren, & Nikolic, 2015) to understand the cascading effects of disrupted physical CI services on the communities utilising them (section 2.1). A systems thinking approach considers a system to be an interrelated and interdependent set of parts defined by its boundaries and is more than the sum of its parts (subsystems) (Schaveling, Bryan, Schaveling, et al., 2018). The research used the service chain management framework (Purwar, Sliuzas, & Flacke, 2020) with an overarching systems thinking approach to explain and explore these complex relationships in Medellin city, Colombia, as a study case (section 2.1). The study also applied focus group discussion and semi-structured interviews to collect and analyse qualitative data (sections 2.3 & 2.4).

2.1. Research design

A systems thinking approach presents urban actors and physical CI services as system components interacting with each other to deliver or utilise the services provided for a better quality of life. Changing one part of a system affects other parts and often ripples through the entire system (Grohs, Kirk, Soledad, et al., 2018). While the specific impact itself may not be predictable, there are many patterns and behaviours of cascading effects that are predictable (Davis, Challenger, Jayewardene, et al., 2014), spreading from the disrupted CI services themselves to the community utilising these services. To investigate such cascading effects, we apply a service chain management framework to identify the CI services and their users as components of interconnected systems. This framework facilitates the identification of the patterns and behaviours of cascading effects in the wake of a disruptive event in a relatively simple manner. The service chain management framework is based on a conceptual analogy of supply chain management (supply chain of goods to consumers) with infrastructure services as service chain management functions, presenting the CI service network as a system and parts of the service chain as subsystems. Any failure or disruption of service delivery directly or indirectly relates to the delivery process of services (Purwar, Sliuzas, & Flacke, 2020). The framework serves as a generic model characterising the service chain of infrastructure comprising distribution-based services (inflow of material/service, e.g., electricity) to its consumers and collection-based services (outflow of material/service – e.g., solid waste) from its consumers. For this study, the framework applies the qualitative data collected obtained using FGDs and interviews in Medellin city. The conceptual schematization of the service chain management framework applied in the study can be found in Annex 1. Based on the more linear representation obtained through service chain analysis, we then developed a causal loop diagram (CLD) as a systems thinking tool (section 3.5) to visualize how the different variables or factors in the CI system are causally interconnected (Vermaak, 2016). The CLD captures and represents the relationships between factors that drive the various outcomes and system behaviours (Rehman, Sohaib, Asif, et al., 2019) that are important to understand the progression of cascading effects through a community.

2.2. Study area

Medellín is the second-largest city and urban economy in Colombia. As the capital of Antioquia, one of the most populated states in the country (map1), it hosts major economic activities serving the region and beyond. Medellin is often considered a success story and an international benchmark for remarkable urban transformation and social innovation (Biczyńska, 2019). Though Medellín is globally recognised for its remarkable resurgence, it has high levels of inequality and social exclusion and still contends with its informal settlements, which continue to expand on hillsides and are prone to landslides and torrential floods (Medellin Comovamos, 2020). In the past three decades, development policies and regulations around informal settlements in Medellin have evolved from a focus on settlement clearance to upgrading programmes to improve their quality of life.

Map 1. Study area map (source: OpenStreetMap & Municipality of Medellin).

The city administration and public utility agencies have managed to establish public CI services in many informal settlements (Calderon, 2008); however, the coverage or quality of such services could not keep up with their rapid expansion, especially in hazard-risk zones. Therefore, to explore the impact of cascading effects of disrupted CI services on communities, the study has selected two locations, one formal and one informal settlement in the district Villa Hermosa (Comuna 8), known for its complex geography, frequent disaster events and low-income households.

Villa Hermosa (Comuna 8): is a densely populated district with a population of 154,579 in 18 barrios (neighbourhoods) (de Medellín, 2019). The socio-economic status of Villa Hermosa is very low, with low-income households, high vulnerability to natural hazards and a very high poverty index marked by SISBEN¹ category 1 and 2 (Garcia, Smith, Coupe, et al., 2018). SISBEN, Colombia’s unified vulnerability assessment and identification system for social assistance, allows the population to be classified according to their living conditions and income into five categories.

Map 2. Hazard and risk map for Villa Hermosa.

Eighty percent (80%) of Comuna 8’s population is from the two lowest categories, 1 and 2, representing Medellin’s most socio-economically vulnerable groups (Irazábal, Krassner, Calvin, et al., 2013). Villa Hermosa is mostly occupied by internally displaced populations from rural parts of Colombia (Irazábal, Krassner, Calvin, et al., 2013), residing in self-built housing along communal pathways. Although Comuna 8 is close to the city centre, it is not easily accessible due to the steep terrain. The area has been declared a high-risk area (map 2) due to the high number of landslide events that characterize large areas, aggravated by anthropogenic activity such as construction malpractices and unplanned human settlements (Irazábal, Krassner, Calvin, et al., 2013). The river, Quebrada Santa Barbara, is a source of flash floods and debris flows. The steep slopes, a dense network of streams and the soil’s poor water absorption capacity lead to rapid surface runoff during heavy rainfall. Villa Hermosa shares its border with the rural zone of Medellin municipality (map 2), and many informal settlements are located close to this border. Increasing construction and population density on the steep hill slopes also causes the loss of green spaces and land degradation due to the release of sewage, solid waste and other contaminants by households, factories (legal status unknown), construction sites and debris from damaged structures.

Picture 1: Villatina (source: D.Purwar).

Picture 2: El Faro (source: D.Purwar).

Study location 1: Villatina (formal settlement) is on a steep slope (picture 1), and part of the settlement is in a high-risk area adjacent to the river Quebrada Santa Barbara. Most of the households in Villatina have been living there for more than 40 years, having fled from conflict-affected rural parts of Colombia. The neighbourhood’s total population is 11,357, from approx. 2280 households. The majority are low-income households below SISBEN category 2, living in rented houses or houses constructed on private owners’ rented land. The high-risk area falls under an eviction and relocation programme for households living close to the river who face combined threats from fluvial flooding and landslides. Most households do not possess a land title, only house ownership, which puts them at risk of eviction due to risk zone regulations. Villatina, being a municipal urban zone, has full coverage of CI services and SISBEN subsidy schemes for water and sanitation services from Empresas Públicas de Medellín, a utility company (section 3.1).

Study location 2: El Faro (informal settlement), meaning lighthouse (map1), is situated on the steepest slope of Pan de Azúcar hill (picture 2) in the Medellin metropolitan area. As the geographical boundary of El Faro is not well defined, map 1 depicts the approximate location according to the community leader. El Faro has approximately 300 houses, inhabited by some 440 families (close to 1700 people), many of which are headed by young mothers with small children. El Faro is partially in the urban zone of Comuna 8 and partially in the Santa Elena rural zone, thus not part of the municipal map of the administrative boundary. The site is uneven, and with a high probability of landslides, and due to the high cost of installation of CI services given the difficult site, the community does not have access to CIs such as water and sanitation. The area was recently connected to electricity with a pre-paid subsidy scheme that is only accessible to those who can afford it.

Natural hazard event: Due to its tropical climate, Medellin experiences heavy rainfall during the long rainy season (April to December) and a short dry season (January to March). Locally, the phenomena referred to as torrential avenues are classified as (i) channelled debris flows, (ii) debris flooding and (iii) flash floods (Aristizábal, Velez, & Martinez, 2016).

2.3. Data collection

This study selected a reference period for hazard events, September to November 2019, as both study locations experienced torrential floods and debris flows with varying impacts during that period. The experiences of the two communities during torrential flood events were considered during the primary data collection. With the assistance of the community leaders, data were collected through seven (7) focus group discussions (FGDs) involving a total of thirty-five (35) people from the two communities (Table 1). Eighteen (18) expert interviews were conducted (annex 4 & 5 for interview questions), and the lead author performed field observations during several transect walks through the two communities.

Table 1. Overview of primary data collection methods and respondents.

Data collection method	Data source	Total FGDs/Interviews	Total participants	Data contributes to:
Focus group discussion (annex 3, section A)	Community (El Faro)	3	11 (10 female, 1 male)	Result section
	Community (Villatina)	4	24 (20 female, 4 male)
Key informant interview (annex 3, section B)	Community leader (El Faro)	1	1 (female)
	Community leader (Villatina)	1	1 (female)
Expert interview (annex 4 & 5)	AMVA	4	4	Discussion & conclusion section
	Medellin municipality	7	7
	EPM (CI service provider)	3	3
	Academia	4	4

The FGDs were used to collect data on CI services availability and accessibility to the community and the challenges related to CI services that emerged after a torrential flood. The initial interviews with the community leaders and the first FGDs in each community focused on identifying critical CI services during and after flood events and identifying the most critical services (refer to section 3.1) to assess cascading effects. Before the first FDG in each community, a service chain diagram (section 2.3) that illustrates the cascading effects of CI services on a community was explained to the community leader, who then further explained the cascading effects and the purpose of the FGD to the participants. The majority of FGD participants were females (30), as male members were less available during the selected days. When required, FGD questions were modified to dig deeper into an issue and bring out additional information from participants. Data on their preparedness and awareness of participants on natural hazards and risks in the study areas, government support for capacity building and disaster recovery measures to cope with hazard impacts, were also collected. Participants of the FGDs in El Faro were more reluctant to share their experiences and concerns about past disaster events as they were worried about being evicted if the study was connected to the municipality.

Transect walks in both study locations were used to observe the general condition of CI services, such as the functionality, accessibility and availability of water and sanitation services, solid waste management and waste disposal sites, drainages, hygiene management, access to public transport and maintenance of these services. The Villatina community leader also discussed drug mafias and drug addiction issues during a transit walk (3–4 km) close to the landslide memorial. A transect walk in El Faro of 2–3 km was also accompanied by a community leader.

Expert interviews (Table 1) were conducted with officials from different government departments and non-government organisations active in programmes and projects on CI services, disaster management, planning and climate change. The interviews focused on the department/organisation’s mandate, disaster response and preparedness plans, and the interdependency of the department (interviewee) on other departments for service delivery. Questions also covered the severity of disrupted CI services triggered by the failure of other supporting services due to hazardous events. Academicians from a local university were also interviewed on the urban transformation of Medellin city. Key informant interviews were carried out with community leaders from each study area about the hazard risks and community project implementation in their respective barrios. The study also used secondary data (Table 2) obtained from AMVA² (Area Metropolitana Valle de Aburra), Medellin municipality planning and disaster management departments.

Table 2. Overview of secondary data sources.

Secondary data
Data type/format	Data source
Risk assessment of Aburra valley, 2019 (PDF)	AMVA
Climate change plan 2019 (PDF)
Metropolitan strategy for emergency response support (PDF)
Plan de Ordenamiento Territorial 2014 (POT) (PDF)	Municipality of Medellin
Demographic profile 2016 to 2019 (PDF)
Municipal plan for disaster risk management 2015 to 2030 (PDF)
Maps & shapefiles for land use, hazards & public infrastructure

Figure 1. Flow diagram of data analysis.

2.4. Data analysis

Figure 1 presents a flow diagram of the analysis carried out by applying a service chain framework, thematic and content analysis leading to the creation of a causal loop diagram. Thematic coding and the service chain management framework were applied to the empirical data to present CI service status for both dry and rainy seasons (during and post torrential flood) for each settlement. Through this approach, critical/key elements responsible for initiating or propagating cascading effects were identified. The schematic presentations of the service chain frameworks for both settlements were compared to identify the types, patterns and severity of cascading effects between them. These results were used, together with the content analysis of literature and the key informant interviews, to investigate the communities’ coping mechanisms and government support to reduce cascading effects in each settlement. Content analysis was used to understand the viewpoints of the communities, the municipality and academia and to identify relationships and patterns between key words and themes communicated during interviews. The words and themes were categorised into sub-systems. The identified factors shaping these six sub-systems form the basis of a causal loop diagram (CLD) that shows how these triggers or propagate cascading effects and visualises the relations between them. The CLD combined factors that are common to both settlements and lead to cascading effects on residents. The factors are represented as text boxes, colour-coded by one of the six groups or sub-systems (geographical, institutional, infrastructural, socio-economic, natural hazard event and direct effect of hazardous event), and the causal relationship between them is represented as arrows. g. The CLD was developed by the authors, and the co-author from UNAL Medellin validated the identified links and relationships. Each link was labelled with a positive or negative causal sign: a positive causal link means that the two factors change in the same direction, i.e., if the origin node (factor) decreases, the destination node (factor) also decreases and vice versa. A negative causal link means that the two factors change in opposite directions, i.e., if the origin node increases, then the destination node decreases, and vice versa (Vermaak, 2016).

3. Results and discussion

Results obtained from focus group discussions and interviews in the study area are discussed in this section. First, the CI service chain in dry and rainy seasons is described. This is followed by a schematic representation of the CI service chain in each settlement, illustrating the cascading effects caused by torrential floods. Next section discuses current strategies and actions of the community and the city administration to address the impacts of cascading effects. The last section concludes with a CLD that shows a snapshot of the significant factors and systems contributing to exacerbating cascading effects.

3.1. CI services in the settlements

The data from FGDs showed that the CI services are water and sanitation. However, in a post-disaster situation, there is also a critical need for transportation services, electricity supply, health services and solid waste management. These were all considered in our analysis. Empresas Públicas de Medellín (EPM), a utility company owned by the Municipality of Medellin, provides water, electricity, and cooking gas services to its primary market, the metropolitan area (urban zones) of Medellin. Empresas Varias de Medellín (EVM), also owned by the municipality, is responsible for industrial and household solid waste management services (López, 2016). EVM collects, transports and disposes of ordinary waste from residential, commercial, industrial, construction, and educational entities. The main modes of motorised transportation in the area are Metro cable, public buses, and private taxis. Health care services for lower-income and homeless residents, including insurance, are available under the SIESBEN beneficiary programme (López, 2016). EPM has a public service subsidy programme for households in SISBEN categories 1–2 (De Mesa, 2014). However, to access these services, one must register via the SISBEN web portal. People who migrated from conflict-affected rural areas often lack identification documents and cannot access public services. In many cases, CI services such as electricity are disconnected due to delayed payment of service fees by families without a stable income (Fernando & Atehortúa, 2016; López, 2016).

3.2. CI service chain in the dry season

The CI service chain in the settlements (Figures 2a, b) shows the standard service chain for water, sanitation, sewage, solid waste, electricity, and transportation. The following section explains the dependency of CI services on other services and the current status of service quality and quantity in the study area during the dry season. A service chain management framework is used to assess the cascading effects of disrupted CI services on the community.

As Villatina is part of the municipal urban zone, it has access to CI services according to the city’s municipal regulations. Its service chain is presented in Figure 2a. The water supply chain has three components of distribution (W1, W2, W3), out of which W1 is dependent (Dep 1) on electricity (E1) for the treatment of water collected from various sources. Further, the water supply is managed via a gravity system fed from storage tanks located on the hilltops. The sewage facility (S1) is dependent (Dep 2) on individual water pipes (W3) at the household level. The waste transfer facility (SW3) is dependent (Dep 3) on the transportation services of EVM. Despite frequent waste collection, the level of solid waste management is poor. Waste segregation is not practised, and disposal is often done in the nearest site (creek, roadside and riverbank) instead of the dedicated disposal site, leading to clogged drainage.

Figure 2. (a) CI service chain and cascading effects in Villatina (b) CI service chain and cascading effects in El Faro.

El Faro has a community-managed gravity-based water supply network. The water source is a small stream locally known as La Castro de Santa Elena, falling along the slopes of hills above El Faro. Water is collected in a storage tank from where it is distributed through pipelines constructed by the community. However, because the pipes are of poor quality, they are constantly damaged by heavy vehicles. Also, as the stream’s water quality is low, it is used only for washing and bathing purposes and not for direct consumption. Although the residents boil or filter the water before consumption, health problems related to the skin or stomach often arise. The water quantity is also inadequate for the entire settlement. As sewage lines are not available, residents build sewage tanks or release sewage into small creeks throughout the area. At the household level, sewage is dependent (Dep 1) on water supply (CW3) (Figure 2b). EVM’s solid waste service is only available for part of the area and is dependent (Dep 2) on-road transportation service for EVM vehicles. Also, solid waste management at the community level is low, as waste is disposed of anywhere due to the lack of adequate dedicated disposal sites. Small creeks and streams in the neighbourhood are often blocked by solid waste. El Faro also has a few Empressa de Desarrollo Urbano (an urban development company) (EDU) drainage lines merged with community-managed sewage lines built under the Jardin Circumvular project (Anguelovski, Irazábal-Zurita, & Connolly, 2019) to collect rainwater. Nevertheless, these drainage channels are inadequate to channelize all wastewater towards the drainage lines of the urban zone.

The CI service network in Villatina (Table 3) shows a higher interdependency than that of El Faro (Table 4). For example, electricity and transportation are two important services in Villatina, on which other services are dependent for service delivery to the consumer. Whereas in El Faro, only sewerage and solid waste services are dependent on water and transportation, respectively; other services operate independently managed at the community level. Interdependent services are observed as a network system with linkages, which strongly influence the operational characteristics of service delivery.

Table 3. Service dependency matrix for Villatina (formal settlement).

Table 4. Service dependency matrix for El Faro (informal settlement).

CI service dependencies are also characterised as risk multipliers (O’Sullivan, Kuziemsky, Toal-Sullivan, et al., 2013); their failure can trigger cascading effects. The higher the interdependency between services, the higher the risk of cascading effects due to service failure (Petit & Verner, 2001). Factors that contribute to varying levels of cascading effects on CI services and their dependent communities, such as the causes of service failure, its duration and human behaviour, are elaborated in the next section.

3.3. CI service chain in the rainy season (during and post torrential floods)

The wet season brings an average rainfall of 1752 mm (SIATA, Medellin, 2020) and regularly creates damaging flash floods on the hill slopes. In 2018–19, torrential floods and landslides affected both settlements, with varying impacts in different parts of each one, including the disruption of CI services with cascading effects on services and the community.

The water and electricity supply in Villatina was generally unaffected during and after the heavy rainfall; although frequent power cuts did occur, these had no significant effect on the community (FGD participant). However, solid waste collection by EVM was disrupted for a considerable time due to flood-damaged roads. Waste from open disposal sites throughout the neighbourhood was therefore washed into the natural and built drainage channels (Figure 2a Cas 4). Clogged sewers often lead to the accumulation of surface runoff and sewage water wherever it finds space, leading to further surface runoff (Cas 1) with a tremendous amount of polluted water. Rapid surface runoff further blocks and sometimes damages roads, affecting the daily traffic (Cas 2) delaying the EVM waste collection vehicles (Dep 2), leading to unattended waste and clogging drainage (Cas 3). Cascading effects (Figure 2a Cas 1–4) reoccur in a cyclic pattern throughout the rainy season, leading to unhealthy surroundings, water and vector-borne diseases, flooded houses, disrupted livelihoods and education. Typical diseases that affect the community are dengue, viral fever, pneumonia or malaria. The many narrow sidewalks and staircases within the neighbourhood were flooded, blocking access to the local market, pharmacies, health centres and other services.

‘The surface runoffs in the rainy season are very dangerous, as most roads and sidewalks are slopes, making it difficult to either walk or drive. Every year, people get injured by getting dragged along the rainwater or skidding vehicles on the road. Though there are railings along the road in some areas for pedestrians. (Reyanaldo,³ Male, FGD participant, Villatina).

In El Faro, the quality of the community-managed water supply is further reduced during the rainy season due to the low quality of supply pipes (damaged/leaking) and muddy rainwater (Figure 2b) from the water source (Cas1 & Cas 2). Individual supply pipes (CW3) placed along the neighbourhood walls and stairways are often damaged (Dep 1) by mudflows or garbage-laden stormwater. This leads to clogged toilets, household sewage (Cas 3) and accumulated waste nearby houses. Open sewers (S2) also reduce the quality of the water supply. Poor solid waste management results in clogged drainage channels (Cas 7) (S2), leading to increased surface runoff (Cas 4). This situation increases the risk of mudflows and landslides due to the constant accumulation of surface water and public health problems due to the proliferation of vermin and mosquitoes. FGD respondents reported a high incidence of skin problems, malaria, dengue, and many more health problems during the wet season.

‘Because of poor water quality, our children fall sick very frequently during the rainy season. It does impact our expenditure, but frequent sickness also affects their education with their inability to attend school’. (Anna, Female, FGD participant, El Faro).

Open sewers also exacerbate the water retention capacity of the loose soils found on the hill slopes (Figure 2b) in El Faro, particularly during torrential rain. As a result, debris flow or landslide events are common during monsoons. Because of the steep slopes, downhill surface runoff may carry large amounts of soil and debris at high speed, causing considerable public and private property damage. Torrential flows obstruct roads (Cas5), block traffic (Dep2), leading to delayed EVM waste collection vehicles (Cas 6), which nevertheless is anyway insufficient in normal conditions due to inadequate collection routes. These cascading effects are further exacerbated through the loss of income, additional expenditures for purchasing drinking water, reduced access to schools and health care compounded by the health risks associated with garbage-laden mud deposits. Severe downpours often damage some houses, which are typically made of mixed materials, rendering them uninhabitable.

‘Most of the households in informal settlements are conflict-affected families, without any government support and lack an identification document, which makes it difficult for the municipality to support after a disaster event’. (Consultant, Environment planning, Municipality of Medellin)

The patterns and types of cascading effects that occurred from CI service disruption are similar in both neighbourhoods. From the analysis, it is evident that the sewerage chain (S2) is a critical element in both the planned and the unplanned neighbourhoods. Moreover, human behaviour is important as the residents are responsible for keeping the sewer lines clear of solid waste to prevent clogging. In addition, the inappropriate solid waste disposal facility (SW2) also contributes to drainage clogging. Both communities exhibit poor management of solid waste and sewage, and neither practices waste segregation, resulting in plastic waste clogging household and public drains. Also, EVM’s response to clean solid waste after heavy rainfall is often delayed in both study areas, further perpetuating cascading effects. demonstrating cascading effects due to organisational dependency (Boaru & Bădiţa, 2008), a type of dependency where the community is dependent on EVM for solid waste collection (Pescaroli & Alexander, 2016b).

In both neighbourhoods, the patterns of cascading effects on water and sanitation services have short-term and long-term impacts. The short-term impact was mostly related to the immediate access to potable water and sanitation services resulting in health issues. Expensive health care services increase the burden of post-flood morbidity, leaving households unable to cope economically. Residents find themselves in a vicious cycle of a poverty trap and persistently reduced income that undermines their resilience over time. The severity of cascading effects is similar in both study locations, as their exposure to torrential rains and the geography of the settlements are quite similar. Moreover, in both settlements, most are low-income households with limited education. Similar patterns and types of cascading effects in both neighbourhoods were also found. Thus, for these two settlements, there is no evidence to suggest that the cascading effects of disrupted CI services are more severe in informal settlements.

A key difference between the two settlements is that the recovery period after a flood/landslide event is shorter in Villatina than in El Faro, as the CI service providers are better able to repair any damages. By contrast, recovery in El Faro depends on its residents’ skills to restore disrupted services. As many residents may themselves have suffered losses and given their limited financial and technical resources, recovery is slower. Nevertheless, the lack of legal access to CI services in El Faro has pushed that community to adapt to the constraints and better use of available resources. Though the El Faro community is unaware of the municipal disaster preparedness and recovery measures, they do have their own coping mechanisms. For example, residents treat their drinking water (boiling/filtration) throughout the year, sewer lines/tanks are constructed for those who can afford them, and many people are skilled in the siting and construction of gravity-based supply lines on steep slopes and in protecting housing structures during rainstorms. These observations suggest that the El Faro residents have a higher ability to adapt and cope with post-disaster situations than those of Villatina. Therefore, the cascading effects of disrupted CI services may vary in severity when comparing two different communities or two different hazard events in the same community. These variations in cascading effects may be attributed to demographic context, geography, socio-economic development, regulations and policies, community coping capacities, risk awareness, and other factors associated with urban development (section 3.5).

3.4. Capacities and government support in the study area to recover from cascading effects

Based on our empirical data, the Villatina community is better informed about natural hazard events than El Faro. In Villatina, the residents owning smartphones may access early warning information from SIATA.⁴ Although, according to FGD participants, the distribution of this information is not uniform, those receiving early warning messages do inform others. In comparison (Table 5), there is no direct access to SIATA’s early warning system in El Faro, due to the lack of access to the internet and fewer smartphones. Community mobilisation in Villatina is active and strong due to past landslide events, and their community emergency committee regularly organises awareness and preparedness events. This committee also developed an evacuation plan, identified safe locations as refuges and stimulates residents to stock basic food for the rainy season to reduce the need to travel over inundated roads. As shared by a community leader in Villatina, ‘because of the past tragic event of landslides and torrential floods, people are aware of the risks and damages, thus participate pro-actively in awareness and preparedness meetings organised by the emergency committee of the community’.

Table 5. Community capacities and govt. support in study area.

Sr.No.	Community capacities & preparedness	Villatina	El Faro
1	Hazard & risk knowledge	Yes	No
2	Disaster preparedness	Yes	No
3	Early warning	Yes	No
4	Evacuation plan	Yes	No
5	Collective efforts to reduce risk	Yes	No
6	Access to CI services	Yes	Partial
7	Skills to repair & restore CI services	No	Yes
8	Land tenure security	Yes	No
9	Government support
	a. Capacity building	No	No
	b. Participatory activities	Yes	No
	c. Access to risk information	Partial	No
	d. Access to resources (response & recovery)	Yes	No

The legal status of Villatina gives its residents access to citizen participation in city development and disaster management planning process, capacity building and awareness programmes on risk reduction and risk knowledge. However, access to these is limited to community leaders and citizens actively involved in community development. Moreover, participants expressed difficulty to fully comprehend risk information due to the complex language used (scientific words). In El Faro, FGD participants collectively shared that they are unaware of the necessity of disaster awareness and preparedness and are not aware of any meetings organised by the municipality on these issues due to the illegal status.

A key difference between the capacities and government support in both settlements is the legal status of the settlement. Villatina recovers better and faster with accessible and available resources from cascading effects post-disaster than El Faro. However, land tenure security and legal access to CI services are often a prioritised concern in informal settlements globally (Deely, Dodman, Hardoy, et al., 2010), thus keeping disaster preparedness to reduce risk and cascading effects lowest on the priority ladder, as is the case in El Faro. With the increasing frequency of torrential floods and landslides in the study area, challenges emerging post-disaster are considered as part of the routine life in El Faro and are not a priority to prepare for the hazards. However, as discussed in section 3.3, people in El Faro are skilled in restoring community-managed water and sanitation services, paving the way for recovery, albeit at a slower pace than in Villatina due to the lack of government support and resources. Households with relatively low incomes often suffer more severe consequences with a longer impact on family income and coping capacity for future events.

At the administrative level, as part of the municipal land use and disaster management plan, a risk zone approach is adopted and implemented at the neighbourhood level, with mandatory eviction of households living in the risk zone. However, this zoning approach is relatively static and does not consider the cascading effects of CI services that go beyond the direct impacts of torrential floods and landslides. Rather, the Medellin Risk Management Plan 2016–2030 adopts the concept of a ‘systemic risk scenario’ (Schweizer, 2019) based on concentrated threats after continuous rain for over two weeks. This approach focuses on limiting urbanisation by agronomic interventions such as creating ecological gardens on the periphery of Medellin. However, these interventions do not go beyond the geological aspects of the risk zone, and new immigrant residents continue to occupy risk zones such as slopes or banks of the stream (de Medellin, 2015). Therefore, it is important to take a broader perspective to address all issues arising from natural hazard events in an integrated manner. This would also entail that broader community interests are taken more seriously than is currently the case.

3.5. Other factors contributing to cascading effects

To synthesise the above analysis, the factors identified from the analysis of interviews (Table 6) were used to create a CLD that presents a snapshot of the significant factors and sub-systems contributing to exacerbating cascading effects (Figure 3).

Table 6. Factors contributing to cascading effects (for colour code, refer to Figure 3).

The positive and negative relationships between these factors are shown in Figure 3. Each group of factors, also observed as a sub-system, influences two or more sub-systems, and thus creates a chain of relationships or a non-linear representation of cascading effects. For example: the institutional factor, poor institutional engagement, and coordination, has a positive relation with the level of community participation. However, poorly performing community participation reduces disaster awareness and leads to a negative relationship resulting in an extended recovery period after a hazardous event. Similarly, geographical causes of flood events may arise due to torrential rainfall that leads to saturation of water absorbing capacity of the soil, increased surface water run-off, and the drainage system overflow and, subsequently, torrential floods and landslide events. Institutional and infrastructure factors show how the lack of CI services overwhelms the existing capacity of the CI service system and results in the disruption or failure of CI services, which is further influenced by the lack of disaster preparedness and awareness of the community on how to cope with torrential floods.

Figure 3. Causal loop diagram to present additional factors.

Moreover, each factor contributing to cascading effects behaves differently, making it complex and unpredictable to determine the cause or origin of cascading effects that progress through the interconnected sub-systems. Furthermore, as the CLD only depicts the general behaviour and the relationships between different factors and not their strength, the severity or extent of cascading effects, with respect to other systems such as government institutions, geography or socio-economic conditions, cannot be determined. However, the CLD does help to better understand the relationships that exist between the general public, physical CI services and government authorities and the way they interact, complementing the more linear visualisation of the service chain framework (see 3.2 and 3.3). Therefore, the CLD facilitates the identification and visualisation of causal links related to torrential floods and landslides. These features may be used to support discussions on intervention points in the system to minimise cascading effects, such as where and how various actors can contribute to reducing cascading effects and their short- and long-term impacts.

3.6. Research limitations

Boundary constraints have played a role in this research. There is a possibility of cascading effects occurring from effects on service chains outside of the geographical context. For example, water supply from a dam on the outskirts of Medellin affected by flooding at the source may cause cascading effects in the study area. For this research, it was not possible to include a more extensive network of the service chain, as this would have shifted the emphasis to the broader physical infrastructure rather than the affected community. Similar boundary constraints influence the CLD analysis, which is limited to geographical factors within the study area. Translation between Spanish and English languages has shaped the primary data collection process, which was dependent on a local translator. Inevitably, while much effort was made to secure the narratives of FGD participants, some of the details of their experiences have been lost in translation.

4. Conclusion

This study shows that, despite their differences in legal status, a formal settlement (Villatina) and an informal settlement (El Faro) of Medellin city experience similar cascading effects of disrupted CI services caused by torrential floods and landslides. However, cascading effects may vary across different contexts. The location, duration and severity of cascading effects depend upon several factors, including natural conditions, human factors and government regulations and policies. The nature and relations between such factors and cascading effects found in the two settlements which we studied has been captured in a causal loop diagram that highlights those factors that are exacerbate or ameliorate cascading effects and visualises the complexity of their cause-effect relations. While the community in the formal settlement, Villatina, was found to be more mobilised and aware of disaster risk reduction practices, the informal community in El Faro was more self-reliant and able to adapt to a crisis, though constrained by its limited resources. Therefore, in each community, different approaches are needed to address cascading effects. In general, the municipality led participatory processes, which are now used as a token to get development plans approved, could be strengthened by seeking local knowledge and local solutions in collaboration between the community and municipality. El Faro residents who are skilled in gravity-fed water network construction can provide their knowledge on risk reduction strategies for infrastructure services and houses. On the other hand, the relatively strong community network in Villatina can be mobilised to share local knowledge and influence community behaviour through improved community awareness and preparedness for disasters and the functioning of CI services.

Future research directions can emphasize enhanced community participation, and how collaboration with relevant stakeholders could be exploited to address cascading effects at the local level. As emphasized by two scholars from Universidad Nacional de Colombia, Medellin (code 15 & 16, annex 2), a broad range of potential stakeholders (planners, disaster managers, academicians, CI service providers, community and other relevant participants) should contribute ideas, local knowledge and perceptions so that the government can make development plans and implement measures with the benefit of information that is widely dispersed in society (Prothro, 2009). Therefore, reducing the cascading effects of CI service failure at the community level requires a more collaborative approach and participation. This can be supported by a systems thinking approach that builds a shared understanding of key factors that lead to CI service failure and their interrelations. This approach should involve all stakeholders directly engaged in the operation, maintenance, provision and usage of CI service as separate systems, i.e., municipality, community and CI service providers that interact with each other for the growth & development of the city. Such stakeholder collaboration improves government effectiveness by encouraging partnership and cooperation (Alves, Patiño Gómez, Vojinovic, et al., 2018). Community-based disaster risk management is one such complementary approach to stakeholder collaboration, wherein the community identifies the problem, ideates and develops solutions appropriate for the community to adapt and implement at individual and community levels (Proudlock & Mitchell, 2012). Realising such an approach can help to reduce the cascading effects of torrential floods and landslides in Medellin.

Though COVID-19 was not directly studied in this research, as the pandemic emerged in Medellin after the data collection, it is impossible not to mention it. Subsequently, the pandemic has significantly affected Medellin, particularly low-income neighbourhoods like those in Villa Hermosa. The study area was severely affected by an extended lockdown that affected people’s livelihoods, access to food, health centres, education, and CI services (Ramirez, 2020). For instance, residents could not practice frequent handwashing due to pre-existing water supply problems or could not follow social distancing when living in small, crowded houses. Therefore, the inclusion of risks and the pandemic’s impact on the community could be significant for assessing and managing additional cascading effects. Improved awareness of disaster risk reduction in communities, efficient coordination and communication between all relevant stakeholders, and strong self-development concerning disasters/pandemics and community development could significantly reduce cascading effects.

Ethical approval

The research project received positive ethics advice from the ethics review committee, Faculty of Geo-information and Earth Observation, University of Twente (reference no. 2020012101, 21/01/ 2020).

Acknowledgments

The authors wish to thank all the individuals who participated in this study for their time and collaboration. We would especially like to thank the community members of Villatina and El Faro for their time and patience with FGD sessions, and Dr Veronica Botero, Dr Francoise Coupé and Dr Edier Aristizabal from the National University of Colombia, Medellin for their support and guidance in selecting study areas and data collection. We are also grateful for all the support we received from the staff of Medellin City administration and AMVA, who agreed to be interviewed, and Mr Christhian Vitola, for helping us with language translation during data collection in Medellin.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data used in this research is published in DANS Easy repository (https://doi.org/10.17026/SS/9A57WR) and access can be granted, taking the ethical considerations and privacy restrictions into account.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/23789689.2024.2340924

Additional information

Funding

The research paper is part of PhD research undertaken at the Faculty of Geo-Information Science and Earth Observation, University of Twente, The Netherlands.

Notes on contributors

Deepshikha Purwar

Deepshikha Purwar holds a master’s in Disaster Management (2011) from Tata Institute of Social Sciences, India and received a second master’s in Urban Planning and Management (2016) from the Faculty of Geo-information Science and Earth Observation, University of Twente, Netherlands. She has brief experience leading and managing humanitarian responses in India. She is currently a PhD candidate (2018-present) at the University of Twente studying the cascading effects of critical infrastructure failure due to natural hazards and how cascading effects impact the community level in Medellin City, Colombia. This work allows her to develop action research regularly to stay close to the experiences and processes in her fields of study. Her main research interests are humanitarian aid, urban resilience, disaster risk reduction, critical infrastructure services, cascading effects, and spatial planning in the Global South.

Johannes Flacke

Johannes Flacke holds a Diploma degree in Geography (1994) from Ruhr-University Bochum in Germany. He received his PhD in Geosciences at the Ruhr-University Bochum in 2002 with a dissertation on information systems promoting sustainable land use based on spatial indicators. Before joining ITC at the University of Twente as Assistant Professor Spatial Planning and Decision Support Systems in 2007 he worked as a postdoc researcher at the Department of Urban and Regional Planning, TU Dortmund University. Dr. Johannes Flacke has over 20 years of theoretical and practical experience in urban and regional planning in the global north and south and the development and use of geo-information and planning support systems. Fields of application have been urban land use planning, informal settlements, health and environmental inequalities, sustainable land management, climate change adaptation, hazard and risk management, and poverty alleviation and targeting. His research focuses on the transformation of urban areas into sustainable and resilient places for all. Main research interests are currently informed spatial decision making for urban planning in the global south, climate change adaptation planning at the local level, the application of collaborative spatial decision support systems in urban planning and micro-level modelling of urban dynamics.

Elizabeth Arboleda Guzman

Elizabeth Arboleda Guzmán is an anthropologist at the University of Antioquia, Colombia, and a doctoral candidate in Geography at the University of Toulouse, France. She obtained her Master’s in Habitat from Universidad Nacional de Colombia (UNAL), Sede Medellin. She is also a professor at Escuela del Hábitat, UNAL. Her work with disaster risk management over the last fifteen years is framed within this perspective. Currently, as a result of her doctoral work, she is developing research related to the energy transition and its impacts on territorial planning in Colombia.

Richard Sliuzas

Richard Sliuzas is Professor of Urban Planning for Disaster Risk Reduction. He is an urban planner, specialised in the use of geo-spatial technologies for urban planning and management. His research interests and activities are focused on the use of geo-spatial technologies in spatial planning for sustainable urban development with an emphasis on issues related to urban informality, urban poverty alleviation and the relationship between spatial planning and disasters. He has supervised PhD’s working on social-ecological resilience in Rafsanjan, Iran; integrated urban and flood modelling in Kampala and Kigali; the use of UAVs for slum mapping in Kigali, Rwanda; Planning support for urban resettlement projects in Kigali, Rwanda; Risk perception and flood mitigation in Kampala; modelling impacts of urban growth in the Nile River valley; disasters and urban spatial planning issues in Lalitpur, Nepal and Medellin, Colombia, amongst others. Richard has worked in numerous international development and capacity building projects including the following countries: China, Egypt, Ethiopia, India, Indonesia, Malawi, Mozambique, Rwanda, Tanzania, Uganda and Vietnam. From 2012-2013 he was team leader of the Integrated Flood Management in Kampala project undertaken under the UN-HABITAT Cities and Climate Change Initiative. He is currently coordinator of the Resilience and Risks Management Strategies thematic group of the Association of European Schools of Planning (AESOP) and a member of the scientific committee for the Global Human Settlement Initiative.

Notes

1. SISBEN: The System of Identification of Social Program Beneficiaries.

2. AMVA is the Autonomous Urban Authority, that represents ten (those municipalities located in the Aburrá Valley) of the 125 municipalities in Antioquia.

3. Names are changed to protect participants’ identity.

4. SIATA: is an early warning system, as flagship project of the Metropolitan Area of the Aburra Valley (AMVA) and the Mayor’s Office of Medellin to manage risk of the región: https://siata.gov.co/siata_nuevo/