Introduction

Floods are considered the most calamitous natural hazards in the world, widely distributed leading to significant economic and social damages than any other natural phenomenon (Mateeul and Memon 2012). They are caused by a range of human vulnerabilities, inadequate development planning, climate variability, and watershed morphometry. Normally expected floods are generally welcomed in many parts of the world as they provide water-rich soil, but unexpected (damaging) floods with excessive frequency cause damage to life, infrastructure, and the environment. In recent decades, the trend of flooding has changed, becoming more frequent, intense, and unpredictable for local communities, especially as development issues and poverty have brought more people to live in areas vulnerable to flooding. The EM-DAT (Emergency Database 2007) registered a total of 210 hydrological events that have caused 4250 deaths and immense property damage.

Urban development is an important issue on a global scale. Urbanization, which means the growth of the population living in urban areas, is one of the major social changes affecting the globe (Chen 2007). Since 2007, half of the world’s population has lived in an urban environment. Once this level is reached, the total population living in cities is expected to double in the next 30 years (Sahani 2011).

In urban areas, naturally permeable soils are converted into impervious surfaces such as roads, parking lots, and building surfaces (Kang et al. 1998). Because of the reduction in infiltration, such surfaces cause increasing runoff. Often, the smooth surfaces allow an increasingly rapid surface runoff, and the areas of depressions and storage are reduced. Besides, natural drainage systems are modified into enlarged and straight artificial channels whose roughness is negligible. Generally, these modifications facilitate surface runoff and consequently the peak flood flow (Sauer and Thomas 1984).

To reduce losses in human lives and material, it is primordial to identify the flood risk and manage it in flood-prone areas (Zaharia and Costache 2017). This cannot be done without an appropriate approach to assess and map flood risk. In addition to the assessment, the flood hazard maps are useful tools for spatial planning and development of the city, especially in the area where it is necessary to identify infrastructure guidelines for the construction of a sewage drainage system and to determine the extent of the strict construction of new settlements (Gigovic and Pamucar 2017).

Contemporary, geographical information systems (GIS) are an indispensable tool; they have become a fundamental element in risk evaluation and mapping, notably flooding (Roopnarine and Opadeyi 2018). The flood visualization component of GIS enables engineers and emergency response planners to become familiar with the potential behavior of flooding, its rates of rising, evolving flood extents, and areas of high flood hazard with lead times prior to the area concerned being flooded (Djordjevic and Vojinovic 2014).

Decision-making is a major process for community managers, government agencies, and companies. It requires feedback from previous experiences, expert advice, and the use of traditional operational search techniques. These are mainly concerned with economic function (Benmena 2000). Decision-makers are more hesitant to make decisions based on feelings and intuitions; instead, they prefer to use analytical and quantitative tools to make and analyze their decisions on solid ground. Several methods based on applied mathematics are useful in decision-making. Among these methods, some require, as instruments, subjective judgments of a decision-maker or expert. In this context, the analytical hierarchy process (AHP) becomes a useful tool for analyzing and making decisions. In general, AHP is a theory and methodology for relative measurement. In relative measurement, we are not interested in the exact measurement of some quantities, but rather in the proportions between them (Brunelli 2015).

Flood risk management requires, first, quantification and knowledge of the propagation of this risk. This must explicitly go beyond the traditional approach, which is often limited to the simple mapping of flood zones and involve a more credible approach that relies on the analysis of the vulnerability of different issues. In this context, the AHP method was widely used (Willett and Sharda 1991; Pawattana and Tripathi 2013; Lawal and Matori 2012; Gigovic and Pamucar 2017). The integration of GIS and multicriteria decision analysis has attracted significant interest over the last 15 years or so. These two distinct areas of research, GIS and MCDA, can benefit from each other (Malczewski 2006). Multiple studies focused on using AHP as a decision-making tool to evaluate the weight and importance of natural aspects of the watershed such as morphometric parameters, slope, land use, soil texture, and rainfall intensity to flood forecasting (Sinha and Bapalu 2008; Pawattana and Tripathi 2013). The combined GIS-MCDA approach is an efficient tool to evaluate the degree of risk and to put suitable strategies to reduce its impact (Fernandez and Mourato 2018). However, the response variables that constitute the urban tissue such as buildings and roads must be taken into consideration to evaluate the vulnerability of a city to flood risk. Based on this concept, this paper aims to determine and map the spatial distribution of the vulnerability to flood risk according to the topography in Sigus city. This paper is organized as follows: the “Introduction” section gives a brief description of the study area. In the “Material and methods” section, we explain the methodology used to map the vulnerability of flood risk. The “Results and discussion” section is devoted to results and discussion. Finally, the “Conclusions” section summarizes the results of this work and draws conclusions.

Material and methods

Study area

Sigus city is located 56 km from Oum El Bouaghi, 36 km from the Constantine, east of Algeria. Its geographical location between two mountains downstream of a dense drainage network makes it susceptible to flood risk (Faregh and Benkhaled 2016). The city is crossed, in the center, by the main stream of this network. The city is divided into 5 districts (Fig. 1), the 1st one is the largest with 80.83 ha. Districts 02 and 04 are completely saturated. District 01 is not urbanized.

Fig. 1
figure 1

Location map of Sigus city

Full size image

Sigus city is situated downstream a watershed of 330 km2. This watershed is characterized by a dense dendritic drainage network with a main stream of order 5, and hypsometric features between 754 and 1350 m (Faregh and Benkhaled 2018). According to the Algerian department of geology, map 121, the geology of Sigus region is mainly composed of cone of red dejection of the great limestone massifs (including all the Quaternary up to the age of the medium terraces and, in part, recent alluvium). The subrecifal limestone, marl, and limestone formations belonging to the Cenomanian are found in the east of Sigus.

Information about rainfall is essential data for surface runoff estimation (Soulis and Valiantzas 2012). The data collected are those of the Fourchi rainfall station code 100511. They were collected from the National Agency for Water Resources (ANRH). We have considered the maximum daily precipitation (Pdmax) corresponding to the flood events of 22 September 1995 and 01 October 2001 shown in the chronicle (Fig. 2).

Fig. 2
figure 2

Chronicle of daily maximum precipitation of the Sigus area

Full size image

AHP method

AHP is a multicriteria analysis method used in decision-making introduced by the mathematician Thomas Saaty in 1980. The method allows integrating several criteria and arriving at a justified choice. The decision is rational, systematic, and correct. By giving the weight of each criterion, Saaty and Vargas (2000) has developed a pairwise comparison method (PCM) according to a scale of importance (Table 1). The first step in decision-making is to divide the general problem into sub-problems that form, later, the hierarchical structure. Figure 3 illustrates the flowchart of the proposed methodology.

Table 1 The scale of importance
Full size table
Fig. 3
figure 3

Diagram of the AHP method

Full size image

To make the pairwise comparison, we use a questionnaire to obtain the relative importance of the criteria between them. These judgments will be transformed into a matrix called “pairwise comparison matrix” indicated by the equation:

$$ {M}_p=\left[\begin{array}{ccc}1& {r}_{12}\kern0.5em \cdots & {r}_{1j}\\ {}\begin{array}{c}1/{r}_{12}\\ {}\vdots \end{array}& \begin{array}{c}1\\ {}\vdots \end{array}\kern0.5em \begin{array}{c}\cdots \\ {}\ddots \end{array}& \begin{array}{c}{r}_{2j}\\ {}\vdots \end{array}\\ {}1/{r}_{1j}& \begin{array}{cc}1/{r}_{2j}& \cdots \end{array}& 1\end{array}\right] $$

Eventually, the pairwise comparison matrix consists of values varying between 1/9 and 9. The elements of the diagonal always have the value 1. The values assigned to the criteria reflect opinions and preferences, which means that there is no rule for cross-checking them. But this can lead to a matrix that is not coherent especially in the case where the number of criteria is important. The coherence of the matrix is determined by the calculation of the “coherence ratio” CR. For good consistency of judgments, the CR should be <0.1 otherwise the judgments are considered inconsistent and the pairwise comparison must be checked.

Construction of the hierarchical model

Using the AHP method, to assess vulnerability to flood risk, requires, first, the design of the hierarchical structure (Pugnet 2015). Among other criteria, the topography is a decisive factor to aggravate flooding flows. This criterion involves, in its physical sense, three aspects that are distance from the main stream (C 1), the elevation of the site (C 2), and the slope (C 3) of the site. Also, urban land use (C 4) is a relevant criterion (Fig. 4).

Fig. 4
figure 4

The hierarchical model

Full size image

The first step is to establish the pairwise comparison matrix between the four criteria, and each criterion is assigned a weight compared to another criterion. Table 2 illustrates the pairwise comparison between the three criteria, Tables 3 and 4 show the pairwise comparison matrix.

Table 2 Pairwise comparison between the four criteria
Full size table
Table 3 Pairwise comparison matrix
Full size table
Table 4 Pairwise comparison matrix
Full size table

To establish the vulnerability maps, all the four criteria maps must be standardized. For that, we propose a scale of indexes starting from index 01 which means “very low vulnerability” to index 05 which corresponds to “very high vulnerability” as shown in Table 5.

Table 5 Vulnerability indexes
Full size table

Among the three criteria composing topography, distance from the main stream has the most impact. This distance is measured from each bank and not from the axis of the main stream. The indexes 5, 4, and 3 correspond to the regular intervals 0–50, 50–100, and 100–150. The distance 228 m corresponds to the point furthest of the flooded zone (Faregh and Benkhaled 2018). The range 150–228 m indicates a “low vulnerability” index. The distance 934.6 m marks the furthest point from the built-up area to the main stream. Finally, we assign the index “01” to the area between 228 and 934.6 m since it is very far from the flood zone.

The criterion “site elevation” means the altitude of the building. It relates to the different altitudes of the flood zone and the geographical position of the site. This criterion is less efficient, in terms of vulnerability assessment, compared to the previous criterion. If the elevation of a building is within the range of the flood zone but it is far from it, it will be less vulnerable. The limits of the flood zone are between 743 m and 804. This allows us to estimate that indices 2, 3, 4, and 5 reflect the variation in the importance of vulnerability within the flood zone. The index “01” corresponds to the range 804–897 since all the constructions included in this interval are considered less vulnerable.

The “slope” criterion is less influential because it is about local slope (a plot of a few meters) which does not generate a destructive flow. This criterion, in association with criterion 1, can play an aggravating or slowing role in the flow propagation.

One of the most essential elements to assess vulnerability to flood risk is urban land use. Urban elements have a great impact on runoff genesis. Buildings, parking, and urban forests are considered obstruction elements to flood flows. On the other side, roads act as open channels as they transfer massive flows to outlets. They are, also, vulnerable as the high velocity of the flow can damage their structure. In this study, we created the geodatabase “land use,” with ArcGIS software, by merging different urban element layer. Each layer was created separately. For this study, we classified land use by category into buildings, roads, forests, main stream, and empty lands.

Results and discussion

This paper presents a useful tool to assess the vulnerability to flood risk in Sigus city (Fig. 6). The main outcome of this study is the vulnerability to flood risk map showing the spatial distribution of all risk classes areas in Sigus city (Fig. 7). To that aim, we calculate the criteria weights using the AHP method and incorporating results with GIS-based technics that include overlaying and raster conversion. The AHP results (Fig. 6) indicate that criterion C 1 is assessed to be the most influential for vulnerability to flood risk with 62% followed by site criterion C 2 with 21% and C 3 and C 4 with the same score of 8%. The calculation of consistency gives a CR = 0.017 (CR < 0.10); this means that the judgment is rationally consistent.

We performed the spatial analysis by overlaying the different criteria maps to obtain the vulnerability map. To carry out a detailed analysis, we used the geoprocessing function “intersect” within the software ArcGIS Pro 2.5. This function allows extracting the value of the vulnerability index that corresponds to land use.

Generally speaking, the results reveal a significant difference between the five indexes of vulnerability as shown in Fig. 6. The most remarkable result to emerge from the vulnerability map is that almost half of Sigus city is situated in low-risk zones. This rate decreases to reach nearly 20% of high-risk zones. Among all rates, very high-risk zones present the lowest value of 0.13%. As shown in the table, the rate of very high-risk for buildings is negligible. This rate heightens in moderate-risk zones to attain 6% and continues to reach 20% in low-risk zones. Roads present the lowest values on all indexes with a maximum of 40% in low-risk zones. The highest rates are to be found in empty lands, and low-risk zones have the maximum with 52%. Moderate and high-risk zones have almost the same rate with nearly 16%. Very high-risk zones are insignificant (Table 6).

Table 6 Vulnerability indexes
Full size table

From the resulting map of “distance from the main stream” criterion (Fig. 5), it has been found that half of Sigus city is located in very low-risk zones, yet, 13% is situated in very high-risk zones. Moderate and high-risk zones present 11 and 10%, respectively. Districts 01, 02, 04, and 05 present, almost, the same rate of very low-risk of nearly 50%. The lowest rate of very low-risk is to be found in district 03. The rate of moderate-risk in the five districts is nearly the same by 10%. Among all districts, 04 shows the lowest rate of high-risk zones with 0.69%. This rate heightens to attain 17% on district 03. The rate of very high-risk is approximately the same on districts 01, 02, 04, and 05 with 10%. District 04 presents the maximum rate of 38.5%.

Fig. 5
figure 5

Maps of vulnerability: a distance from the main stream, b elevation of the site, c slope, and d land use

Full size image
Fig. 6
figure 6

Pairwise comparison between the four criteria

Full size image
Fig. 7
figure 7

Map of vulnerability to flood risk

Full size image

Broadly speaking, the most remarkable result to emerge from the “elevation of the site” vulnerability map (Fig. 5) is that half of Sigus city is situated in very high and high-risk zones. Moderate-risk zones have a remarkable rate of 26.5%. In detail, the map highlighted that districts 01, 02, and 05 show nearly the same rate of very low-risk zones with 17%. This rate diminishes to reach a negligible value of 0.7% on district 03. As mentioned earlier, high-risk zones are the most dominant, particularly in districts 03, 04, and 05. Also, very high-risk is interestingly significant with a minimum of almost 7% in district 01 and a maximum of 31% in district 05.

The vulnerability map according to the “slope” criterion (Fig. 5) revealed that low-risk zones are the most dominant with 46%. This rate decreases to attain 6% in high-risk zones. Unexpectedly, very high-risk is null. Among all districts, 01 presents the highest value of high-risk zones with 11%, flowed by nearly 9% on district 05. The lowest rate is to be found in district 04 with only 02%.

Conclusions

To determine the spatial distribution and vulnerability to flood risk, the use of a multicriteria approach is essential. In this paper, the aim is to map and assess flood risk vulnerability in Sigus city using the analytic hierarchy process (AHP).

In this study, we considered the following criteria: distance from the main stream (C 1), site elevation (C 2), slope (C 3), and land use (C 4). By using the PMC, it has been found that the criterion C 1 is the most important. By transforming the criteria to maps and overlay them, we obtain the final flood risk vulnerability map. The results show that the major part of the city is located in low-risk zones, which are far from the main stream. High-risk zones are close to the main stream. This is due to C 1 which is the most decisive with a weight of 64.34%.

The obtained map is a very useful resource for engineers, planning agencies, and local governments in managing risk and future constructions. Also, the technique applied in this study can be applied to other areas, where other factors may be considered, depending on the availability of data.