## Abstract

The present study aims to assess flood depth, building risk analysis, and the effectiveness of various flood adaptation strategies to attenuate building risk caused by urban floods in climate change scenarios. A framework is proposed where a hydraulic model, Hydrologic Engineering Center's-River Analysis System 2D (HEC-RAS 2D), is applied for 2-dimensional flood modeling to estimate (a) submerged areas, (b) flood depth, and (c) building risk for extreme events corresponding to two representative concentration pathways (RCPs), 6.0 and 8.5. Greater Hyderabad Municipal Corporation (GHMC), India, is chosen for demonstration. Percentages of buildings in GHMC under high, medium, and low risks for RCP 6.0 are 38.19, 9.91, and 51.9% in the respective order, and these are 40.82, 10.55, and 48.63% for RCP 8.5. Six flood proofing (FP) strategies (S1–S6) are proposed for attenuating building risk along with the required capital cost. The capital investment required for FP to achieve the ideal situation of no risk for all buildings (strategy S6) works out to Rs. 3,740 × 107 and Rs. 3,800 × 107 for RCPs 6.0 and 8.5. It is observed that the effect of adaptation strategies is significant.

## HIGHLIGHTS

• Generation of flood inundation maps using Hydrologic Engineering Center's-River Analysis System 2D (HEC-RAS 2D) for the urban catchment of GHMC under two representative concentration pathways (RCPs).

• Building risk analysis and formulation of adaptation strategies under two RCPs.

## INTRODUCTION

Over the past 30 years, urban floods occurring in catchments have been the more prominent natural disasters, which account for almost 43% (EM-DAT, CRED 2015). This can be due to low-lying areas, sedimentation in drainages, clogged drains, and imperviousness of surfaces due to increased urbanization. In addition, natural streams deviated from the original flow paths, resulting in excessive urban runoff (Ahmed et al. 2013; Liu et al. 2019a). Another dimension of urbanization is the enormous increase in the number of buildings that experience different levels of alarming risk during floods. Building risk maps should be developed for different rainfall scenarios, including climate change-related, to assess this complex flood situation. These will help prioritize areas for mitigation measures and serve as a guide map to the policymakers.

The most important parameters for developing building risk maps are elevated terrain levels, contour maps, and corresponding submerged areas. The area under submergence is the extent of inundation in terms of the percentage of area occupied. However, it does not provide information about the water depth above the ground level. Flood depth varies from location to location based on the terrain–rainfall interaction. Risk estimation can be quantified on flood depth (Rangari et al. 2019). The higher the flood depth, the more risk of damage to ecology, roads, buildings, and infrastructure. This makes flood depth analysis a distinct necessity. However, flood depth alone does not serve the purpose while analyzing the infrastructural damage. For example, high flood depth in uninhabited areas does not cause much damage to infrastructure compared to buildings with low to medium flood depth. Hence, flood depth levels around buildings and flood inundation maps can be used to identify locations where buildings require more maintenance than usual.

Accordingly, the present study aims to assess flood depth, building risk analysis, and effectiveness of various flood adaptation strategies to attenuate building risk caused by urban floods in a climate change scenario. For demonstration, we have chosen Greater Hyderabad Municipal Corporation (GHMC), India. GHMC witnessed many flood events in the past, resulting in damage to property. This necessitated studying the response of GHMC toward extreme rainfalls for flood preparedness. Many studies were performed in the past for hydraulic modeling of urban floods for producing flood inundation depth, flood risk maps, and depth damage relationships. Some of the relevant literature reviews are as follows.

Mosquera-Machado & Ahmad (2007) evaluated flood hazard (FH) for Atrato River, Quibdó, northwest Colombia. They employed geographic information systems (GIS), statistical approaches, and Hydrologic Engineering Center's-River Analysis System (HEC-RAS) for evolving three FH maps having different return periods (RPs). Water depths on the left and right banks were 3.7 and 3.1 m for a 50-year RP, which could be used as the basis for designing flood protection structures. Fan et al. (2009) applied Qual2 K with HEC-RAS 1D for assessing the water quality of a tidal river, northern Taiwan, and the models exhibited good agreement. Timbadiya et al. (2011) employed HEC-RAS for the lower Tapi River to simulate the flow for 1998, 2003, and 2006. Simulated flows were closely matched with the observed data. Qasim (2013) simulated free flow over the broad-crested single-step weir using HEC-RAS 1D. Simulated and observed flow profiles were matched with reasonable accuracy. Klimeš et al. (2014) simulated hydrological characteristics of floods due to glacial lake outburst using HEC-RAS 1D along the Chucchun River, Peru, resulting in a simulated peak discharge of 580 m3/s. Authors have only considered 1D flow (depth) for flood analysis but have not applied 2D flow (depth and flood extent for submergence areas) using HEC-RAS 2D.

Quirogaa et al. (2016) applied HEC-RAS 2D to simulate the flood events of February 2014 for a case study of Llanos de Moxos, Bolivian Amazonia. They verified the same with available satellite images. Bhandari et al. (2017) used HEC-RAS 2D in a lower region of the Brazo River watershed, Richmond, TX. Maximum water depth and flood velocities were 13.1 m and 16.1 m/s, respectively. An increase in flooding was expected in climate change scenarios. Song et al. (2017) investigated the effect of Manning's roughness coefficient in HEC-RAS 1D for a case study of the German lowland area. They found that the roughness coefficient played a significant role in enhancing the model output's quality. Vojtek et al. (2019) evaluated the sensitivity of the event-based approach for small and ungauged basins using HEC-RAS 1D. They simulated flood area and flood volume and stressed the importance of roughness and cross-section parameters, besides excess rainfall determination for effective mitigation. Shustikova et al. (2019) compared LISFLOOD-FP (Local Inertia Solver for FLOOD across Flood Plain) and HEC-RAS (River Analysis System) for Secchia River, Italy. HEC-RAS performed better than LISFLOOD-FP for all resolutions except for a 25-m resolution. Kumar et al. (2020) employed the Global Flood Monitoring System and HEC-RAS 2D for the confluence of Ganga and Yamuna, Prayagraj, India. They identified critical and most critical areas based on a 100-year RP. The authors used Manning's roughness coefficient to enhance the model output quality in river basin analysis. However, no studies are reported in urban areas.

Rangari et al. (2019) applied HEC-RAS 2D to analyze flood inundation depths for different rainfall scenarios for zone 13 of GHMC. Seventeen percent of the total catchment area was likely to flood, of which 9% was under high risk. Surwase & Manjusree (2019) assessed the Storm Water Management Model (SWMM), HEC-RAS 2D, and Height Above the Nearest Drainage (HAND) for simulation of urban flooding for zone 12 of GHMC. All the models complemented each other. They determined the depth of the stream. None of them used the GHMC area nor the climate change impacts in their study.

Swathi et al. (2020) employed SWMM and HEC-RAS 2D to assess the impacts of land use land cover (LULC) and climate change on runoff in GHMC. Madhuri et al. (2021) applied five machine learning algorithms to a case study of GHMC under four representative concentration pathways (RCPs) for predicting flood risk probabilities, where XGBoost was found superior. The mentioned works did not focus on flood depth, building risk analysis, and prioritized areas under climate-based scenarios.

Dau & Kuntiyawichai (2015) simulated flood risks due to climate change in Central Vietnam. They used the global climate model (GCM), namely, HadCM3, for A2 and B2. A potential increase in runoff and water levels was observed. Shrestha & Lohpaisankrit (2016) assessed the FH potential based on three GCMs in Yang River Basin of Thailand using TOPMODEL and HEC-RAS 1D. The total flooded area was found to increase for RCP 4.5 as compared to the baseline period. Mondal et al. (2018) applied SWAT, HEC-RAS, and Delft3D to three major rivers of Bangladesh under RCP 8.5. An increase in peak flood levels by 25–72 cm was expected by this century's end. Dau & Kuntiyawichai (2020) assessed the climate change impacts on the operations of a reservoir in Huong River Basin, Vietnam, for RCP 8.5. They coupled HEC-HMS and HEC-RAS 2D. Mean rainfall, mean water level, mean annual runoff, and temperature would increase in the future. Hanif et al. (2020) examined flood risks for 2030 and 2050, respectively, under RCPs 4.5 and 8.5 using HEC-HMS and HEC-RAS. The study also provided potential adaptation strategies. Malik & Pal (2021) analyzed the flood susceptibility of the lower Dwarkeswar River for historical and future floods for one GCM using the HEC-RAS rain-on-grid model. They concluded that Weibull's method was the best suited for flood susceptibility mapping. These studies could not address the need for better adaptation strategies required to mitigate the risk of future floods.

Aerts et al. (2018) computed the associated costs of various adaptation measures for coping with sea-level rise in coastal LA County. Adaptation costs range between $4.3 and$6.4 billion. de Ruig et al. (2020) combined cost–benefit analysis (CBA) and developed a flood risk model for Naples and Venice Beach in Long Beach and Los Angeles. An optimal mix of 35–45 and 55–65% of dry-flood proofing (FP) and building elevation measures was proposed. Ventimiglia et al. (2020) developed a methodology for determining an optimal combination of FP measures using CBA for the Mela river, northeastern Sicily. Results supported the possibility of decreasing the damage caused due to floods by implementing FP measures. Han (2021) evaluated building-level adaptation strategies by incorporating the latest information on the sea-level rise through a dynamic programming-based CBA analysis for Bay County, Florida. He concluded that investing in adaptive measures decreased the average annual damage. The studies mentioned did not take into consideration the exclusive FP measures based on building risk analysis.

In summary, the research gaps identified in the climate change scenario are as follows:

• Building risk analysis based on flood depth

• Adaptation measures for buildings at risk

Hence, the objectives chosen for GHMC are (a) to develop flood inundation maps and identify the submerged areas for RCPs 6.0 and 8.5, using HEC-RAS 2D, (b) to compute building risk analysis based on flood depth, and (c) to formulate suitable adaptive measures and assess their effectiveness for attenuation of building risk in the urban catchment.

The following sections discuss the methodology, study area description, data collection and processing, and description of HEC-RAS 2D along with calibration and validation, results & discussion, and summary of the work.

## METHODOLOGY

HEC-RAS 2D was calibrated and validated using the National Remote Sensing Centre (NRSC) flood vulnerability maps and flood depth data. Two extreme event rainfalls corresponding to RCPs 6.0 and 8.5 were simulated using calibrated HEC-RAS 2D to generate flood inundation maps. These contain information related to submergence areas and flood depths at all locations of GHMC. Flood depths at the building locations were obtained by overlaying the GHMC building shapefile with the flood inundation map in Arc GIS. The resulting building risk map emphasizes classifying buildings into three risk categories based on corresponding flood depths. In addition, six FP strategies were also defined, and the positive impacts of their implementation were assessed. Investment required and the equivalent annual costs (EAC) for the implementation were also computed for each building category.

### Study area, data collection, and processing

The catchment area of GHMC, India, is 625 km2. It consists of 16 zones based on a stormwater drainage network, as shown in Figure 1. Elevations of the area fall between 462 and 635 m from mean sea level. Higher elevations are found in zones 9 and 11, the northern parts of zones 12, 14, and 7. Similarly, lower elevations are found near the Musi river, especially in the southern part of zones 1, 4, 5, 6, 12, 13, and 15. The river Musi flows between the catchment area from west to east. All natural streams drain the water to the Musi river, the main tributary of the Krishna River Basin.

Figure 1

DEM/pictorial representation of zones in GHMC.

Figure 1

DEM/pictorial representation of zones in GHMC.

Several factors affect the flooding process, which include terrain, slope, LULC, and imperviousness. Variation in these factors leads zones 1, 2, 3, 4, 6, and 7 to more flooding. With more imperviousness percentage area, the water holding capacity of soil decreases, resulting in excess water on the surface, causing inundation. LULC map is shown in Figure 2 (refer to Table 1 for more details). A significant amount of data is necessary to facilitate satisfactory flood risk analysis of buildings. Table 1 presents information about significant data collected/processed, including rainfall (historic and climate-based), digital elevation model (DEM), percentage imperviousness land use, LULC, and cost/m2 for water proofing (WP).

Table 1

Information on data collected and their sources

S. No.DataInformationDescription/sources
Historic rainfall (2016) for validation of HEC-RAS 2D September 20–28, 2016 (in mm) (3.8, 44.6, 9.2, 103.2, 21.4, 18. 4, 2.8, 7.8, and 4.7), totaling to 215.9 mm GHMC, Directorate of Economics and Statistics
Future rainfall (2006–2100) for generation of flood inundation mapping using HEC-RAS 2D RCP 6.0: mid-century extreme event, 2050; 440.35 mm occurred in 17 days (0.30, 7.21, 79.42, 117.61, 181.69, 35.39, 3.40, 2.33, 3.30, 2.71, 1.26, 0.88, 1.33, 1.99, 0.78, 0.57, and 0.18 mm)
RCP 8.5: mid-century extreme event, 2064; 624.2 mm occurred in 19 days (0.21, 0.10, 1.29, 1.39, 0.83, 0.89, 2.82, 3.93, 5.27, 10.55, 17.78, 17.42, 15.70, 33.04, 48.59, 324.14, 135.86, 3.45, and 0.94 mm)
GCM, Geophysical Fluid Dynamics Laboratory – Coupled Physical Model (GFDL-CM3) and nonlinear regression-based downscaling approaches (Swathi 2020).
Bias correction was done using the standardization approach (Wilby et al. 2004; Salvi et al. 2011).
DEM for developing DTM in HEC-RAS 2D 30 m × 30 m resolution Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) (USGS 2016).
Percentage impervious land use, LULC for generating land-use map, and resulting curve number (CN) (historic) 30 m resolution imperviousness percentage (years): 55 (1995), 59.62 (2002), 60 (2005), 69.02 (2015), 73 (2016), and 80 (2031)
CN: 78 (2000), 83 (2006), and 85 (2016)
(refer to Figure 2
Figure 2 presents the LULC of the study area with a drainage network represented as black arrows. Land usage is classified as open spaces, buildings, roads, railways, water bodies, and streams. Black arrows represent the drainage network overlaid on natural streams (blue lines). This LULC analysis helps in providing essential information on land-use changes.
Sannigrahi et al. (2018), Nayan et al. (2020), Open Street Maps (2016), and Master plan of Hyderabad Metropolitan Development Authority (HMDA) 2031 (HMDA 2019).
Percentage impervious land use, LULC for forecasting CN, and impervious percentage (future) Land imperviousness up to 84.85% by 2050 and 87.04% by 2064
Corresponding CNs were found to be 88 and 89
For CN and imperviousness percentage prediction, four types of curve fitting were performed to capture the variation of land-use impacts on flood routing. These included linear, power, exponential, and sigmoidal curves where corresponding R2 values were 0.9145, 0.8798, 0.9202, and 0.9567. Hence, the sigmoidal curve was determined to be the best fit based on which CN was forecasted.
Cost/m2 for WP for estimating capital cost and EAC Rs. 412/m2 (specialized high-performance acrylic polymer modified elastomeric cementitious waterproof coating to the terrace/walls) Telangana Public Health & Municipal Engineering Department Telangana (TGPHMED 2021).
S. No.DataInformationDescription/sources
Historic rainfall (2016) for validation of HEC-RAS 2D September 20–28, 2016 (in mm) (3.8, 44.6, 9.2, 103.2, 21.4, 18. 4, 2.8, 7.8, and 4.7), totaling to 215.9 mm GHMC, Directorate of Economics and Statistics
Future rainfall (2006–2100) for generation of flood inundation mapping using HEC-RAS 2D RCP 6.0: mid-century extreme event, 2050; 440.35 mm occurred in 17 days (0.30, 7.21, 79.42, 117.61, 181.69, 35.39, 3.40, 2.33, 3.30, 2.71, 1.26, 0.88, 1.33, 1.99, 0.78, 0.57, and 0.18 mm)
RCP 8.5: mid-century extreme event, 2064; 624.2 mm occurred in 19 days (0.21, 0.10, 1.29, 1.39, 0.83, 0.89, 2.82, 3.93, 5.27, 10.55, 17.78, 17.42, 15.70, 33.04, 48.59, 324.14, 135.86, 3.45, and 0.94 mm)
GCM, Geophysical Fluid Dynamics Laboratory – Coupled Physical Model (GFDL-CM3) and nonlinear regression-based downscaling approaches (Swathi 2020).
Bias correction was done using the standardization approach (Wilby et al. 2004; Salvi et al. 2011).
DEM for developing DTM in HEC-RAS 2D 30 m × 30 m resolution Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) (USGS 2016).
Percentage impervious land use, LULC for generating land-use map, and resulting curve number (CN) (historic) 30 m resolution imperviousness percentage (years): 55 (1995), 59.62 (2002), 60 (2005), 69.02 (2015), 73 (2016), and 80 (2031)
CN: 78 (2000), 83 (2006), and 85 (2016)
(refer to Figure 2
Figure 2 presents the LULC of the study area with a drainage network represented as black arrows. Land usage is classified as open spaces, buildings, roads, railways, water bodies, and streams. Black arrows represent the drainage network overlaid on natural streams (blue lines). This LULC analysis helps in providing essential information on land-use changes.
Sannigrahi et al. (2018), Nayan et al. (2020), Open Street Maps (2016), and Master plan of Hyderabad Metropolitan Development Authority (HMDA) 2031 (HMDA 2019).
Percentage impervious land use, LULC for forecasting CN, and impervious percentage (future) Land imperviousness up to 84.85% by 2050 and 87.04% by 2064
Corresponding CNs were found to be 88 and 89
For CN and imperviousness percentage prediction, four types of curve fitting were performed to capture the variation of land-use impacts on flood routing. These included linear, power, exponential, and sigmoidal curves where corresponding R2 values were 0.9145, 0.8798, 0.9202, and 0.9567. Hence, the sigmoidal curve was determined to be the best fit based on which CN was forecasted.
Cost/m2 for WP for estimating capital cost and EAC Rs. 412/m2 (specialized high-performance acrylic polymer modified elastomeric cementitious waterproof coating to the terrace/walls) Telangana Public Health & Municipal Engineering Department Telangana (TGPHMED 2021).
Figure 2

LULC map of GHMC.

Figure 2

LULC map of GHMC.

### Description of HEC-RAS 2D, calibration and validation

In the present study, HEC-RAS 2D, a 2-dimensional hydraulic model, is employed. The model considers hydrodynamics, horizontal and vertical flows, and 2D flow visualization, which are not possible in a 1D model (Quirogaa et al. 2016; Şensoy et al. 2018; Dasallas et al. 2019). HEC-RAS 2D also efficiently solves shallow water and diffusion wave equations (HEC-RAS 2016; Liu et al. 2019b).

The model requires rainfall, DEM as a boundary condition, and Manning's roughness coefficient derived from LULC for calibration. ASTER DEM of 30 m × 30 m resolution is processed in Arc GIS. Model is discretized into an equal number of grid cells of 30 m × 30 m, i.e., equal geometry to maintain spatial uniformity. In addition, DEM of the same resolution is adopted, as it is appropriate for hydrological modeling studies (Buakhao & Kangrang 2016; Rocha et al. 2020; Abili 2021). Simultaneously, other topographical features and infrastructural layers are also made similar to maintain uniformity. This helped in making 2D flow calculations simpler.

DEM is used as an input in the RAS mapper of HEC-RAS 2D to develop Digital Terrain Model (DTM). DTM developed from DEM is also taken into account natural drainage flow. DEM data are added to the New Terrain Layer dialogue of the RAS mapper in HEC-RAS 2D. A new terrain layer is created with terrain files that are used for evaluation. This information is saved in the terrain folder in GeoTiff format. In addition to GeoTiff, two more files are created in .hdf and .vrt formats. The file .hdf is created in the RAS mapper, which contains information on the raster data. The .vrt file helps visualize and display multiple data. For visualizing flood plain, the model geometry can be coupled to DTM. This DTM acts as a basis to create a 2D mesh using the polygon shapefile consisting of 19,717 cells. The DEM and DTM are presented in Figures 1 and 3 that contain submerged area and water depths at each mesh grid. These can be used for computing the water surface elevations to visualize flood plain geometry and flood risk analysis of buildings.

A cumulative rainfall of 165 mm occurred at Qutubullapur on September 21, 2016 for 12 h is used for calibration (Surwase & Manjusree 2019). Manning's value between 0.01 and 0.4 is used as input to the model during calibration (HEC-RAS 2016). It is based on a combination of land-use classes and soil properties (Dorn et al. 2014; HEC-RAS 2016).

The calibration of the 2D model is performed by comparing with the simulated inundation extents at each location using NRSC maps (Carisi et al. 2018). Maps related to only zone 12 are used for calibration due to the non-availability of maps for other zones. Total 11 locations are used for calibration and 8 for validation. F-index, which measures the model's simulating ability, is used (Quirogaa et al. 2016; Liu et al. 2019b). Mathematically, it is expressed as follows:
(1)
where is the simulated flood inundated area (km2), is the observed flood inundated area (km2), and is the inundation area that is common to both observed and simulated (km2). The value of F varies between 0 (no match) and 100 (perfect match). Table 2 presents the computation of the F-index for zone 12.
Table 2

F-Index computations for zone 12

Locations (1)HEC-RAS area (Am) (km2) (2)NRSC maps area (Ao) (km2) (3)Aom (km2) (4)Ao + AmAom (km2) (5)Aom/(Ao + AmAom) (km2) (6)F-index (7)
1.00 0.82 0.82 1.00 0.8243 82.43
0.15 0.09 0.09 0.15 0.6113 61.13
0.16 0.11 0.11 0.16 0.6567 65.67
0.26 0.21 0.21 0.26 0.8114 81.14
0.13 0.11 0.11 0.13 0.8501 85.01
0.16 0.14 0.14 0.16 0.9250 92.50
0.14 0.13 0.13 0.14 0.9174 91.74
0.13 0.11 0.11 0.13 0.8325 83.25
0.07 0.05 0.05 0.07 0.7704 77.04
10 0.07 0.06 0.06 0.07 0.9500 95.00
11 0.10 0.08 0.08 0.10 0.7626
Average
76.26
81.02
Locations (1)HEC-RAS area (Am) (km2) (2)NRSC maps area (Ao) (km2) (3)Aom (km2) (4)Ao + AmAom (km2) (5)Aom/(Ao + AmAom) (km2) (6)F-index (7)
1.00 0.82 0.82 1.00 0.8243 82.43
0.15 0.09 0.09 0.15 0.6113 61.13
0.16 0.11 0.11 0.16 0.6567 65.67
0.26 0.21 0.21 0.26 0.8114 81.14
0.13 0.11 0.11 0.13 0.8501 85.01
0.16 0.14 0.14 0.16 0.9250 92.50
0.14 0.13 0.13 0.14 0.9174 91.74
0.13 0.11 0.11 0.13 0.8325 83.25
0.07 0.05 0.05 0.07 0.7704 77.04
10 0.07 0.06 0.06 0.07 0.9500 95.00
11 0.10 0.08 0.08 0.10 0.7626
Average
76.26
81.02

Calibration is reasonable as understood from the average F-index value of 81% (Quirogaa et al. 2016; Liu et al. 2019b; Rahimzadeh et al. 2019) with Manning's value of 0.015. It is considered representative of the whole area due to uniformity in catchment characteristics in terms of sandy-loam soil.

Validation is carried out at eight locations of zone 12. Historic rainfall of 215.9 mm occurred during September 20–28, 2016 is used for validation. Simulated flood depths vary from 0 to 4 m at Nizampet, Qutubullapur, and Jeedimetla and agree with NRSC maps (Bhatt & Rao 2018).

## RESULTS AND DISCUSSION

Results related to flood inundation mapping, building risk analysis based on flood depth, and corresponding adaptation measures are presented for two extreme events of RCPs 6.0 and 8.5 in this section.

### Flood inundation mapping and submergence areas

After validating HEC-RAS 2D, the model facilitates flood inundation maps for extreme future rainfall events, i.e., 440.35 and 624.2 mm corresponding to RCPs 6.0 and 8.5 (refer to Table 1). Inundation maps are presented in Figure 4(a) and 4(b). It is observed that areas most vulnerable to flooding are located near the Musi river, its tributaries, and Hussain Sagar (highlighted in circles on the map). GHMC-wise submergence areas for RCPs 6.0 and 8.5 are 334.23 and 357.97 km2, respectively, as observed from inundation maps (Figure 4(a) and 4(b)). This can be attributed to their higher rainfall magnitudes. Areas most vulnerable to flooding are in the northern sections of zones 1 and 5 and southern sections of zones 13 and 15.

Figure 3

DTM of GHMC.

Figure 3

DTM of GHMC.

Figure 4

Flood inundation map of RCPs (a) 6.0, (b) 8.5, (c) submerged area of all zones for RCP 6.0, and (d) submerged area of all zones for RCP 8.5.

Figure 4

Flood inundation map of RCPs (a) 6.0, (b) 8.5, (c) submerged area of all zones for RCP 6.0, and (d) submerged area of all zones for RCP 8.5.

In addition, zone-wise submergence areas and catchment areas are shown in Figure 4(c) and 4(d) for RCPs 6.0 and 8.5. Zones 3, 9, 11, and 14 were less submerged due to their high elevation levels in the ranges of 500–564, 561–602, 559–612, and 525–613 m, respectively. This ensures that water does not accumulate. Among these, zones 9 and 11 are relatively less submerged (4.77 and 7.27 km2 for RCP 6.0; 5.43 and 8.53 km2 for RCP 8.5).

It is observed that a higher slope leads to more inundation, as evident from Figures 1 and 4(c) and 4(d). Submerged areas of zone 12 for RCPs 6.0 (69.2 km2) and 8.5 (71.45 km2) are more than those of the other zones. This is due to the higher catchment area (138.12 km2) and terrain characteristics. Zones 1, 2, and 4 were found to have terrains that aid the flooding process, whereas it is vice-versa for zones 8, 9, and 11. Zones 1 (elevation level: 462–539 m) and 2 (elevation level: 492–527 m) are highly submerged (15.69 and 3.68 km2 for RCP 6.0; 17.72 km2 and 3.99 km2 for RCP 8.5).

This situation is challenging for the residents in the vicinity of inundated areas of GHMC. Complimentarily, buildings were 526,545 in number with an area of 84.91 km2. Hence, it was felt to analyze building risk associated with flood depths. These were then categorized as low-risk buildings (LRB) with flood depth of <0.5 m, medium-risk buildings (MRB) with flood depths of ≥0.5 and <1 m, and high-risk buildings (HRB) with flood depth of ≥1 m as discussed in the following section.

### Building risk analysis

Building risk analysis is carried out for all zones of GHMC. Flood depths were found to vary from zone to zone. High flood depth values near buildings were observed in zones 1, 5, 6, 7, 13, and 15; low flood depth values near buildings exist in zone 9. The percentages of HRB, MRB, and LRB for RCP 6.0 in GHMC-wise are 38.19% (201,110), 9.91% (52,144), and 51.9% (273,291), and these are 40.82% (214,943), 10.55% (55,531), and 48.63% (256,071) for RCP 8.5. Values in parentheses represent the number of buildings in that category. The zone-wise building risk analysis for RCPs 6.0 and 8.5 is depicted in Figure 5(a) and 5(b).

Figure 5

(a) Zone-wise percentage of buildings under high, medium, and low risks for RCP 6.0. (b) Zone-wise percentage of buildings under high, medium, and low risks for 8.5.

Figure 5

(a) Zone-wise percentage of buildings under high, medium, and low risks for RCP 6.0. (b) Zone-wise percentage of buildings under high, medium, and low risks for 8.5.

It is observed from Figure 5(a) and 5(b) that for RCP 6.0, the range of zone-wise percentages of HRB, MRB, LRB, respectively, are, 19–48, 9–13, and 41–70%. The highest number of HRB was noticed in zone 1 due to its low elevation levels of 462–539 m. The lowest percentages of HRB and LRB were found in zone 9 due to its high elevation levels of 561–602 m. For RCP 8.5, the ranges of zone-wise percentages of HRB, MRB, and LRB, respectively, are 31–58, 8–12, and 34–64%. This increase in the percentage of HRB is similar to that of RCP 6.0 for zone 1. The most negligible percentage of HRB was found in zones 3 and 6. This can be due to the medium elevation (500–564 m) for zone 3 and a moderate slope of 11.3° for zone 6 where minor flooding is possible. The maximum number of LRB is found in zone 9, which is similar to RCP 6.0. The minimum number of MRB was found in zone 1 for RCPs 6.0 and 8.5.

Accordingly, suitable adaptation measures are necessary for the existing buildings, which are at risk. The adaptation strategy required to reduce damages to the buildings is discussed in the following section.

Adoption of appropriate measures can help in dealing with the risk in the existing buildings. As an effect of flooding, buildings often get degraded in strength and durability and incur repair costs. This damage can be found to increase with the frequency and flood depths. FP is one of the adaptation measures that can deal with this situation (Hochrainer-Stigler et al. 2019). Adaptation strategies are defined based on flood depth near the buildings. An attempt is made to bring HRB to MRB, HRB to LRB, and HRB to no risk buildings (NRB) by applying the following adaptation strategies:

1. S1 – No FP

2. S2 – 2 m FP for all HRBs + rooftop FP

3. S3 – 2 m FP for all HRBs + 1 m FP for all MRBs + rooftop FP

4. S4 – 1.5 m FP for all HRBs + 1 m FP for all MRBs + 0.5 m FP for all LRBs + rooftop FP

5. S5 – 2 m FP for all HRBs + 1 m FP for all MRBs + 0.5 m FP for all LRBs + rooftop FP

6. S6 – FP equal to flood depth for all buildings + rooftop FP

Details of WP charges are available from TGPHMED (2021). The capital investment was based on the summation of WP costs for each building, i.e., Rooftop WP (RWP) and Wall WP (WWP).
(2)
(3)
(4)

FP strategies and their corresponding results are shown in Table 3.

Table 3

FP strategies and their corresponding results

Strategy (1)Climate-based scenario RCP (2)Number of HRBs (≥1 m direct flood contact) (3)Number of MRBs (≥0.5 m and <1 m flood contact) (4)Number of LRBs (>0 m and <0.5 m flood contact) (5)Number of NRBs (No direct flood contact) (6)HRB cost Rs. (USD) × 107 (7)MRB cost Rs. (USD) × 107 (8)LRB cost Rs. (USD) × 107 (9)Total cost of FP Rs. (USD) × 107 (10)EAC/year for a building Rs. (USD) (11)
S1 6.0 201,110 52,144 62,270 211,021 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
8.5 214,943 55,531 64,694 191,377 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
S2 6.0 87,898 71,014 86,561 281,072 2,123 (29) 0 (0) 0 (0) 2,123 (29) 15,030 (203)
8.5 85,316 76,775 92,606 271,848 2,264 (31) 0 (0) 0 (0) 2,264 (31) 14,999 (203)
S3 6.0 87,898 18,870 86,561 333,216 2,123 (29) 507 (7) 0 (0) 2,630 (36) 14,785 (200)
8.5 85,316 21,244 92,606 327,379 2,264 (31) 526 (7) 0 (0) 2,791 (38) 14,689 (199)
S4 6.0 106,768 24,294 29,935 365,548 1,922 (26) 507 (7) 552 (7) 2,981 (40) 13,453 (182)
8.5 106,560 27,913 35,351 356,721 2,050 (28) 526 (7) 575 (8) 3,152 (43) 13,388 (181)
S5 6.0 87,898 18,870 24,291 395,486 2,123 (29) 507 (7) 552 (7) 3,182 (43) 14,359 (194)
8.5 85,316 21,244 27,912 392,073 2,264 (31) 526 (7) 575 (8) 3,366 (45) 14,298 (193)
S6 6.0 526,545 2,743 (37) 478 (6) 518 (7) 3,740 (51) 16,875 (228)
8.5 526,545 2,765 (37) 496 (7) 540 (7) 3,800 (51) 16,144 (218)
Strategy (1)Climate-based scenario RCP (2)Number of HRBs (≥1 m direct flood contact) (3)Number of MRBs (≥0.5 m and <1 m flood contact) (4)Number of LRBs (>0 m and <0.5 m flood contact) (5)Number of NRBs (No direct flood contact) (6)HRB cost Rs. (USD) × 107 (7)MRB cost Rs. (USD) × 107 (8)LRB cost Rs. (USD) × 107 (9)Total cost of FP Rs. (USD) × 107 (10)EAC/year for a building Rs. (USD) (11)
S1 6.0 201,110 52,144 62,270 211,021 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
8.5 214,943 55,531 64,694 191,377 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
S2 6.0 87,898 71,014 86,561 281,072 2,123 (29) 0 (0) 0 (0) 2,123 (29) 15,030 (203)
8.5 85,316 76,775 92,606 271,848 2,264 (31) 0 (0) 0 (0) 2,264 (31) 14,999 (203)
S3 6.0 87,898 18,870 86,561 333,216 2,123 (29) 507 (7) 0 (0) 2,630 (36) 14,785 (200)
8.5 85,316 21,244 92,606 327,379 2,264 (31) 526 (7) 0 (0) 2,791 (38) 14,689 (199)
S4 6.0 106,768 24,294 29,935 365,548 1,922 (26) 507 (7) 552 (7) 2,981 (40) 13,453 (182)
8.5 106,560 27,913 35,351 356,721 2,050 (28) 526 (7) 575 (8) 3,152 (43) 13,388 (181)
S5 6.0 87,898 18,870 24,291 395,486 2,123 (29) 507 (7) 552 (7) 3,182 (43) 14,359 (194)
8.5 85,316 21,244 27,912 392,073 2,264 (31) 526 (7) 575 (8) 3,366 (45) 14,298 (193)
S6 6.0 526,545 2,743 (37) 478 (6) 518 (7) 3,740 (51) 16,875 (228)
8.5 526,545 2,765 (37) 496 (7) 540 (7) 3,800 (51) 16,144 (218)

Notes: Column No. 11 has been rounded off to the nearest integer. 1 US\$ = 74 Rs. as of August 8, 2021.

The EAC has also been estimated along with the capital investment required for each strategy. EAC refers to the cost incurred per year as a result of owning an asset over its lifespan. The estimation of EAC was done using the following equation:
(5)
where C is the capital investment/cost, i is the discount rate [considered 3% above repo rate of 4% for government projects, i.e., 0.07 (7%)], and n is the life of FP (considered 10 years) (Swathi 2020). In addition, as a symbolic value, EAC/building has been calculated as per the following equation:
(6)

An apparent comprehension of the calculation of EAC per building can be obtained from the following example.

The total cost incurred for S2, S3, and S4 under RCP 8.5 is estimated to be Rs. 2,264 × 107, Rs. 2,791 × 107, and Rs. 3,152 × 107, respectively. The corresponding EAC required can be estimated using Equation (5) as Rs. 322.39 × 107, Rs. 397.31 × 107, and Rs. 448.72 × 107, respectively. However, EAC required per building is of greater significance at an individual level than the total EAC. This can be facilitated by Equation (6).

EAC/building for RCP 8.5 under S2, S3, and S4 is obtained as follows:
(7)
(8)
(9)

It should be noted that the denominator in Equation (6) varies across strategies depending on the number of buildings where FP is implemented. The utilization of capital cost for implementing FP results in the reclassification of buildings from higher to lower risk. Analysis was done across various strategies for RCPs 6.0 and 8.5, as shown in Table 3, which results in the following.

Capital investment of Rs. 3,740 × 107 and Rs. 3,800 × 107 are required, respectively, for RCPs 6.0 and 8.5 for achieving the ideal situation of all buildings under no risk (strategy S6). Corresponding EAC/building is Rs. 16,874 and Rs. 16,144. Strategy S5 results in an increase in the number of NRBs by 87.42 and 104.87% for RCPs 6.0 and 8.5, respectively, employing EAC/building of Rs. 14,360 and Rs. 14,298. HRBs decrease from 38.19 to 16.69% and 40.82 to 16.2%, respectively, for RCPs 6.0 and 8.5. RCP 8.5 results in shifting 21,244 from HRB to MRB; 27,912 from HRB to LRB; and 80,471 from HRB to NRB with Rs. 2,264 × 107 for strategy S2. The implementation of strategy S3 requires an amount of Rs. 2,791 × 107, which is Rs. 526 × 107 more than S2 under RCP 8.5. The implementation of strategy S4 for RCP 8.5 leads to a decrease in the number of HRBs, MRBs, and LRBs by 50.42, 49.73, and 45.36%, respectively, and an increase in NRBs by 86.40% by expecting investment, EAC/building of Rs. 3,152 × 107 and Rs. 13,388, respectively.

Overall, the application of FP is an effective measure for flood adaptation, ensuring preparedness toward urban floods. Effective implementation of FP can help decrease the vulnerability of building risk and increase the life span. Additional benefits are the incentives provided in the form of flood insurance. Such schemes will encourage the city dwellers to participate in FP actively. This can also be added to the existing government schemes as a policy, thereby reducing the fear of floods and resultant infrastructural damages.

## SUMMARY AND CONCLUSIONS

The problem of urban flooding is a complex one; hence, it is very challenging to model such situations. HEC-RAS 2D was presented for analyzing extreme future rainfall events of RCPs 6.0 and 8.5. Flood inundation maps based on buildings were studied and analyzed. The percentages of HRB, MRB, and LRB in GHMC-wise for RCP 6.0 were 38.19% (201,110), 9.91% (52,144), and 51.9% (273,291). These were 40.82% (214,943), 10.55% (55,531), and 48.63% (256,071) for RCP 8.5. Adaptation measures for the buildings at risk were also analyzed using six FP strategies. The capital investment required for FP has been calculated as Rs. 3,740 × 107 and Rs. 3,800 × 107 for RCPs 6.0 and 8.5 to achieve the ideal situation of no risk for all buildings, i.e., strategy S6. Information provided in this study will help residents of the catchment in Hyderabad to remain alert and be warned of future calamities, and combat the same in time. Future scenarios of risk maps for buildings and flood inundation maps will help the policymakers before planning any projects. The investments in strategy S6, as suggested, will be comparatively small but are certainly for better efficiencies. The contribution of this study is (a) in emphasizing flood risk of buildings by identifying the flood depths associated with each building in climate change scenarios and (b) in assessing the effectiveness of adaptation strategies for attenuating this risk.

The present work is based on chosen GCM, hydrologic model, RCPs, and costs from various sources. The outcome may vary based on different data/approaches. However, the methodology remains the same, which is the main focus of the present study. All the building conditions were considered the same here, and we wish to study in-depth by considering buildings’ life cycle/age in our future work.

## ACKNOWLEDGEMENTS

The present work is supported by Information Technology Research Academy (ITRA), Government of India under ITRA-water grant ITRA/15(68)/water/IUFM/01. Acknowledgments go to GHMC officials and other agencies for their help. Special acknowledgments to Prof D. Nagesh Kumar, Dept. of Civil Engineering, IISc., Bangalore, for providing critical suggestions for improving the manuscript and Dr Swathi Vemula for providing necessary data support.

## CONFLICT OF INTEREST

The authors declare that there are no conflicts of interest.

## DATA AVAILABILITY STATEMENT

Data cannot be made publicly available; readers should contact the corresponding author for details.

## REFERENCES

Abili
H. Z.
2021
Comparison of vertical accuracy of open-source global digital elevation models: a case study of Adama City, Ethiopia
.
Turkish Journal of Computer and Mathematics Education (TURCOMAT)
12
(
4
),
1731
1744
.
Aerts
J. C.
,
Barnard
P. L.
,
Botzen
W.
,
Grifman
P.
,
Hart
J. F.
,
De Moel
H.
,
Mann
A. N.
,
de Ruig
L. T.
&
N.
2018
Pathways to resilience: adapting to sea level rise in Los Angeles
.
Annals of the New York Academy of Sciences
1427
(
1
),
1
90
.
Ahmed
Z.
,
Rao
D. R. M.
,
Reddy
K. R. M.
&
Raj
Y. E.
2013
Urban flooding-case study of Hyderabad
.
Global Journal of Engineering, Design and Technology
2
(
4
),
63
66
.
Bhandari
M.
,
Nyaupane
N.
,
Mote
S. R.
,
Kalra
A.
&
S.
2017
2D unsteady flow routing and flood inundation mapping for lower region of Brazos River Watershed
. In:
World Environmental and Water Resources Congress 2017
. pp.
292
303
.
Buakhao
W.
&
Kangrang
A.
2016
DEM resolution impact on the estimation of the physical characteristics of watersheds by using SWAT
.
Advances in Civil Engineering
2016
,
1
9
.
Carisi
F.
,
Schröter
K.
,
Domeneghetti
A.
,
Kreibich
H.
&
Castellarin
A.
2018
Development and assessment of uni-and multivariable flood loss models for Emilia-Romagna (Italy)
.
Natural Hazards and Earth System Sciences
18
(
7
),
2057
2079
.
Dau
Q. V.
&
Kuntiyawichai
K.
2015
An assessment of potential climate change impacts on flood risk in Central Vietnam
.
European Scientific Journal
1
,
667
680
.
de Ruig
L. T.
,
Haer
T.
,
de Moel
H.
,
Botzen
W. W.
&
Aerts
J. C.
2020
A micro-scale cost-benefit analysis of building-level flood risk adaptation measures in Los Angeles
.
Water Resources and Economics
32
,
100147
.
EM-DAT, CRED
2015
Centre for Research on the Epidemiology of Disasters. The International Disaster Database. Available from: https://www.emdat.be/ (accessed 6 February 2019)
.
Han
Y.
&
Mozumder
P.
2021
Building-level adaptation analysis under uncertain sea-level rise
.
Climate Risk Management
32
,
100305
.
Hanif
A.
,
Dhanasekar
A.
,
Keene
A.
,
Li
H.
&
Carlson
K.
2020
Flood risk assessment methodology for planning under climate change scenarios and the corresponding change in land cover
.
Journal of Water and Climate Change
11
(
4
),
1370
1382
.
HEC-RAS
2016
River Analysis System: Hydraulic Reference Manual.
USACE Version: 5.0. US Army Corps of Engineers, CPD-68
.
HMDA
2019
Master Plan of the City
.
Available from: hmdaprojects.hmda.gov.in/masterplan. (accessed September 2019)
.
Hochrainer-Stigler
S.
,
Linnerooth-Bayer
J.
&
Mochizuki
J.
2019
Flood proofing low-income houses in India: an application of climate-sensitive probabilistic benefit-cost analysis
.
Economics of Disasters and Climate Change
3
(
1
),
23
38
.
Kumar
N.
,
Kumar
M.
,
Sherring
A.
,
Suryavanshi
S.
,
A.
&
Lal
D.
2020
Applicability of HEC-RAS 2D and GFMS for flood extent mapping: a case study of Sangam area, Prayagraj, India
.
Modeling Earth Systems and Environment
6
(
1
),
397
405
.
Liu
Y.
,
Wang
C.
,
Yu
Y.
,
Chen
Y.
,
Du
L.
,
Qu
X.
,
Peng
W.
,
Zhang
M.
&
Gui
C.
2019a
Effect of urban stormwater road runoff of different land use types on an urban river in Shenzhen, China
.
Water
11
(
12
),
2545
.
R.
,
Sistla
S.
&
Raju
K. S.
2021
Application of machine learning algorithms for flood susceptibility assessment and risk management
.
Journal of Water and Climate Change
.
doi:10.2166/wcc.2021.051
.
Mondal
M. S.
,
Islam
A. S.
,
Haque
A.
,
Islam
A.
,
Biswas
S.
&
Mohammed
K.
2018
Assessing high-end climate change impacts on floods in major rivers of Bangladesh using multi-model simulations
.
Global Science and Technology Journal
6
(
2
),
1
14
.
S.
&
S.
2007
Flood hazard assessment of Atrato River in Colombia
.
Water Resources Management
21
(
3
),
591
609
.
Nayan
N. K.
,
Das
A.
,
Mukerji
A.
,
Mazumder
T.
&
Bera
S.
2020
Spatio-temporal dynamics of water resources of Hyderabad Metropolitan area and its relationship with urbanization
.
Land Use Policy
99
,
105010
.
Open Street Maps (OSM)
2016
.
.
Qasim
J. M.
2013
Laboratory and HEC-RAS simulations of a single-step weir
.
ARPN Journal of Engineering and Applied Sciences
8
(
12
),
1005
1011
.
Rangari
V. A.
,
Umamahesh
N. V.
&
Bhatt
C. M.
2019
Assessment of inundation risk in urban floods using HEC RAS 2D
.
Modeling Earth Systems and Environment
5
(
4
),
1839
1851
.
Rocha
J.
,
Duarte
A.
,
Silva
M.
,
Fabres
S.
,
Vasques
J.
,
Revilla-Romero
B.
&
Quintela
A.
2020
The importance of high resolution digital elevation models for improved hydrological simulations of a Mediterranean forested catchment
.
Remote Sensing
12
(
20
),
3287
.
Salvi
K.
,
Kannan
S.
&
Ghosh
S.
2011
Statistical downscaling and bias-correction for projections of Indian rainfall and temperature in climate change studies
. In:
4th International Conference on Environmental and Computer Science
. Vol.
19
, pp.
16
18
.
Sannigrahi
S.
,
Bhatt
S.
,
Rahmat
S.
,
Uniyal
B.
,
Banerjee
S.
,
Chakraborti
S.
&
Bhatt
A.
2018
Analyzing the role of biophysical compositions in minimizing urban land surface temperature and urban heating
.
Urban Climate
24
,
803
819
.
Şensoy
A.
,
Uysal
G.
&
Şorman
A. A.
2018
Developing a decision support framework for real-time flood management using integrated models
.
Journal of Flood Risk Management
11
,
S866
S883
.
Shrestha
S.
&
Lohpaisankrit
W.
2016
Flood hazard assessment under climate change scenarios in the Yang River Basin, Thailand
.
International Journal of Sustainable Built Environment
6
(
2
),
285
298
.
Shustikova
I.
,
Domeneghetti
A.
,
Neal
J. C.
,
Bates
P.
&
Castellarin
A.
2019
Comparing 2D capabilities of HEC-RAS and LISFLOOD-FP on complex topography
.
Hydrological Sciences Journal
64
(
14
),
1769
1782
.
Song
S.
,
Schmalz
B.
,
Xu
Y. P.
&
Fohrer
N.
2017
Seasonality of roughness-the indicator of annual river flow resistance condition in a Lowland catchment
.
Water Resources Management
31
(
11
),
3299
3312
.
Surwase
T.
&
Manjusree
P.
2019
Urban flood simulation – a case study of Hyderabad city
. In:
National Conference on Flood Early Warning for Disaster Risk Reduction
,
May 30–31, 2019
,
, pp.
133
143
.
Swathi
V.
2020
Analysing the Linkages Between Urban Floods, Climate Change and Land Use
.
PhD Thesis
,
BITS Pilani Hyderabad campus
.
Swathi
V.
,
Srinivasa Raju
K.
&
Sai Veena
S.
2020
Modelling impact of future climate and land use land cover on flood vulnerability for policy support–Hyderabad, India
.
Water Policy
22
(
5
),
733
747
.
TGPHMED
2021
Telangana Public Health & Municipal Engineering Department Website
. .
P. V.
,
Patel
P. L.
&
Porey
P. D.
2011
HEC-RAS based hydrodynamic model in prediction of stages of lower Tapi River
.
ISH Journal of Hydraulic Engineering
17
(
2
),
110
117
.
USGS
2016
ASTER Data
.
Available from: http://earthexplorer.usgs.gov/ (accessed January 2016)
.
Vojtek
M.
,
Petroselli
A.
,
Vojteková
J.
&
Asgharinia
S.
2019
Flood inundation mapping in small and ungauged basins: sensitivity analysis using the EBA4SUB and HEC-RAS modeling approach
.
Hydrology Research
50
(
4
),
1002
1019
.
Wilby
R. L.
,
Charles
S. P.
,
Zorita
E.
,
Timbal
B.
,
Whetton
P.
&
Mearns
L. O.
2004
Guidelines for use of climate scenarios developed from statistical downscaling methods
. In:
Supporting Material of the Intergovernmental Panel on Climate Change.
Available from the DDC of IPCC TGCIA, p. 27
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).