Flood risk susceptibility analysis in Larkana district Pakistan using multi criteria decision analysis and geospatial techniques

Study area

Study area map of Larkana district with river network

The Larkana district (study area) is located in the southeast side of the country and upper district in Sindh province and near to the west bank of the Indus River as shown in Fig. 2a. The map delineates the Larkana District as the specified study region, with its perimeters indicated in yellow. Rivers and streams are shown by blue lines, which shows the district’s natural drainage and hydrological network as shown in Fig. 2b.

(a) Location of study area in Pakistan map; (b) River streams in study area.

River stream orders network of study area

The GIS map of Larkana District shown in Fig. 3a and map Fig. 3b represents the river stream network in three layers: first order (red) for headwater tributaries, second order (yellow) from their converge, and third order (blue) for main channels. The topography, shown as a color gradient ranging from higher (green) in the northeast to lower (red) in the southwest, indicates a northeast-southwest flow direction. Stream numbers identify segments for hydrological investigation, while components such as the scale bar, compass rise, and stream labels promote and beautify the map and its utility for analyzing water flow as shown in Fig. 3b.

(a) Larkana district (study area) map; (b) River stream order network map.

Data collection and model setup

Table 1 below contains criteria for calculating flood risk, with various parameters assigned rank, rating, and weights for predicting flood probability. Each criteria’s range has been divided into classes, rated from 1 (Very Low) to 5 (Very High) based on flood probability, and weighted to indicate their respective importance in flood risk assessment.

Methodology flowchart-diagram

The flowchart (Fig. 4), describes the framework process for flood risk potential mapping and area computation, which works on dataset and GIS thematic mapping to assess flood hazards using the (AHP) approach.

Methodological framework of the current study.

This methodology uses spatial data analysis with GIS technique and AHP method to determine flood risk potential in study area. The flowchart explains a flood risk mapping technique that begins with collecting data from ASTER DEM Landsat 8 satellite image and open-source data from different web sources mentioned in Table 2. These datasets are then processed to generate thematic maps that illustrate various geographical and environmental aspects related to flood risk. The weights of class and sub-classes within each flood parameter are then determined in preparation for risk analysis. To quantify the relative relevance of the primary criteria, the weights are calculated using the AHP (Analytic Hierarchy Process) approach. Finally, the data are used to generate a flood risk potential map, allowing for flood risk evaluations.

Elevation and slope

In this study, a slope map was constructed using slope generating tools and the ArcGIS digital elevation model. The DEM for the study area was obtained from the NASA website’s Earth Data Search (https://search.earthdata.nasa.gov/search).

(a) Digital elevation model; (b) Reclass of elevation map; (c) Slope map; (d) Reclass of slope map.

The elevation and slope GIS maps of the Larkana District evaluate flood hazard levels by organizing topographical features. In Fig. 5a, the Digital Elevation Model shows elevations ranging from 12 to 84 m, whereas the Reclassified Elevation Map Fig. 5b categorizes these values into five flood risk levels (Very Low to Very High). Similarly, the Slope Map Fig. 5c shows slope percentages throughout the region, with flat to steep gradients affecting water runoff prospective. The Reclassified Slope Map Fig. 5d further categorizes slopes into five risk levels identifying flood-prone locations based on their steepness. These maps help to identify flood-prone areas in the district.

Soil texture and NDVI maps

The soil and NDVI GIS maps of Larkana District evaluate flood hazard levels using soil composition and vegetation index. In Fig. 6a, the Soil Map shows clay composition, with areas classified as 16–29%, which affects water retention and flood hazard, https://data.apps.fao.org/map/catalog/srv/eng/catalog.search#/home the soil data was obtained from FAO Map Catalog, The Reclassified Soil Map Fig. 6b simplifies this to flood risk levels by categorizing regions as Very Low or Very High. The NDVI Map Fig. 6c shows vegetation density, with NDVI values representing the health and existence of vegetation throughout the area, the vegetation index data was obtained from USGS Earth Explorer, Landsat Collection 2 Level 1 https://earthexplorer.usgs.gov/. The Reclassified NDVI Map Fig. 6d ranks flood risk levels from Very Low to Very High based on vegetation density, assisting in identifying flood-prone areas.

(a) Soil map; (b) Reclass of soil map; (c) NDVI map; (d) Reclass of NDVI map.

Curvature and topographic wetness index maps

The GIS maps of Larkana District evaluate terrain and flood hazard levels using curvature and topographic wetness indexes. In Fig. 7a, the Curvature Map shows slope differences, with convex (steep) sections in red and yellow, and concave (gentle) slopes in blue and green. The Reclassified Curvature Map Fig. 7b classifies these slopes as flat, moderate, or steep, which aids terrain interpretation. The Topographic Wetness Index (TWI) Map Fig. 7c shows regions where water can accumulate, with a green gradient representing wetness levels. The Reclassified TWI Map Fig. 7d divides wetness into low, moderate, and high flood hazard zones, providing a more detailed perspective of flood-prone areas.

(a) Curvature map; (b) Reclass of curvature map; (c) TWI map; (d) Reclass of TWI map.

Land use and land cover maps

The GIS maps of Larkana District contain information about land use/ land cover, and flood risks. The Land Use and Land Cover (LULC) Map in Fig. 8a divides areas into sections such as water bodies, forests, flooded vegetation, crops, built-up areas, barren land, and rangeland, with each class represented by a different color, the land use and land cover data was obtained from Esri-Sentinel-2 10 m LULC Download, https://search.earthdata.nasa.gov/search. The Reclassified LULC Map Fig. 8b divides these classes into flood risk intensity levels ranging from very low (green) to very high (red), highlighting areas with higher flood risk, providing a more detailed perspective of flood-prone areas.

(a) Land use and land cover map; (b) Reclass of land use and land cover map.

Distance to river and precipitation maps

The above GIS maps of the Larkana District give important terrestrial information for flood risk and river/ rainwater information. In Fig. 9a the Distance to River Map shows how close different regions are to rivers, with a color gradient indicating proximity, which is useful for evaluating flood risk (Fig. 9b). The Reclassified Distance to River Map organizes these distances into broader proximity categories, allowing for a more rapid assessment of flood-prone areas (Fig. 9c). The Precipitation Map shows rainfall distribution, showing areas of variable precipitation Climatic Research Unit, cru-uea data, https://crudata.uea.ac.uk/cru/data/hrg/ (Fig. 9d). The Reclassified Precipitation Map categorizes this data into flood risk levels ranging from very low to extremely high, providing useful information for flood prevention and planning in agriculture and infrastructure. These maps, used together, help to make informed decisions on how to manage Larkana’s environmental resources.

(a) Distance to river map; (b) Reclass of distance to river map; (c) Precipitation map; (d) Reclass of precipitation map.

Analytical hierarchy process (AHP)

AHP is a multi-criteria decision-making process that uses hierarchical structures to weigh and prioritize the features and alternatives respectively based on judgement. The AHP involves six steps: Parameters selection, evaluation criteria and developing the AHP hierarchy, pairwise comparison/ standardized matrix, estimate the relative weights, calculation and check the consistency, obtained overall rating/ rank as shown in Fig. 10.

Selection of flood risk parameters

The Analytical Hierarchy Process (AHP) is a systematic decision-making process for selecting criteria in flood risk assessment that prioritizes aspects based on their relative relevance. In AHP, experts first identify and prioritize important features that determine flood susceptibility, such as elevation, slope, land use, soil type, and distance from rivers as shown in Fig. 11.

List of selected parameters.

Criteria evaluation

After the selection of criteria’s for specific purposes, the pairwise comparisons are then made between criteria to calculate their weights, which indicate each factor’s influence on flood risk as mentioned in Tables 3 and 4. The weighted criteria are combined to generate a flood risk map, which guides flood management and mitigation actions. AHP ensures a systematic transparent approach, balancing expert judgement with quantitative facts to enable informed decision-making.

Characterization and standardization of the criteria

Normalizing comparison matrix and calculating criteria weights for each parameter by using Eq. (1),

whereas, N_ij represents normalized value of ith row and jth column,

X_ij represents element of ith row and jth column in pairwise matrix.

The pairwise compared matrix is normalized for eliminating error subjectivity by using Eq. (2). Then calculated relative importance and relative weights of each parameter by averaging the normalized matrix for each row as shown in Table 5.

whereas, X_ij represents element of ith row and jth column, n represents the number of elements.

Weighting criteria

The weights criterion step in AHP provides priority levels to each flood risk factor, the weighted sum score is calculated by adding the totals from each row from the product matrix as process shown in Table 5. Higher weights, such as rainfall (28%) and distance to rivers (22%), indicate stronger influence on flood susceptibility, guiding the attention on these critical factors in flood assessment and planning. To determine the estimated value’s precision, multiply the pair-wise assessment matrix with criteria weigh matrix. In Table 6, the required process shows for calculating the resulting matrix from the multiplication of pairwise matrix with weighted sum score obtained from above Table 5 to produce the required outcomes in new column of weighted sum values and then calculating the ratio of weighted sum value by dividing the weighted sum scores and the criterion weights produced the weighted sum score to criteria weight ratio. The last column of Table 6 shows the estimated ratios.

Calculation of Λ _Max, CI and CR and check of consistency

In this step the consistency of the matrix is checked to ensure that the judgment of making decisions is correct and consistent. The formula for consistency index is calculated as:

whereas, λ _max represents the maximum eigen value.

λ_max = (10.21 + 10.05 + 10.26 + 10.43 + 10.13 + 9.84 + 9,36 + 8.85 + 9.10)/9.

λ_max = 9.803.

λ_max = (10.21 + 10.05 + 10.26 + 10.43 + 10.13 + 9.84 + 9,36 + 8.85 + 9.10)/9.

λ_max = 9.803.

To calculate the value of consistency Index (CI), by using equation,

CI = (9.803–9)/(9 − 1).

CI = 0.100.

After the calculation to have determined the value of consistency ratio (CR) by using equation,

CR = 0.100/1.45.

CR = 0.069 or 6.9%.

Note, If the consistency ratio (CR) is less than 0.1 or 10% it means the matrix considered as consistent hence it is accepted so that the calculated value can be used. Otherwise, it may need to be revised for adjustment.

Obtained overall rating and ranking

The table shows the prioritization of flood risk assessment factors, according to their importance in contributing to flood risk. Rainfall has the highest priority (28%), followed by proximity to rivers (22%) and slope (16.7%), as these characteristics have a strong influence on flood susceptibility. Lower-ranking factors such as soil type (2.5%) and curvature (1.6%) have a minimal effect on flood risk as shown in Table 7. Each criterion’s rank and positive/negative variation percentages provide information about its relative importance, allowing effective flood management and mitigation efforts by focusing on the influential factors.

AHP model

AHP model, involves various flood risk indictors that contribute to the final weighted overlay output map.

AHP model for all parameters.

The AHP model, shown in Fig. 12, is a structured decision-making framework used in geospatial analysis to assess flood risk by incorporating multiple flood risk factors. The Topographic Wetness Index (TWI), slope, elevation, river distance, soil type, land use/land cover, terrain curvature, rainfall, and vegetation index (NDVI) are all important input features. Each parameter is subjected to “Pairwise Processing,” which determines its relative importance. This procedure weighs the inputs, which are then integrated into the “Weighted Overlay” output to provide a comprehensive map for analyzing flood vulnerability or contributing to other geospatial decision-making.

The Eq. (6) defines function W(Cm, pc). Where Cm denotes the criteria weights with base run whereas pc denotes the percentage changes in level, and the product of W (Cm, 0 ) and xp is a modified version of W(Cm​, 0), scaled by the factor xp in order change in weight of other criteria at change level in percentage. In equations 7 and 8 the weights of all flood risk features are used and add-ups.

Here W (Ci, pc) represents the weight of ith criteria at specific change level.

There are n criteria’s, with RPC min and RPC max representing the minimum and maximum values of the RCP.