Urbanization and climate change increase the frequency of urban floods worldwide which in turn poses a great challenge in the management of the urban stormwater drain system. This study investigated the impact of climate change and surface imperviousness in the urban stormwater runoff and the impact of washed off pollutants load in the conduit efficiency for two census wards Triplicane (ward 116) and Valasaravakkam (ward 152) in Chennai City. The climate change data for the study area wards were taken from the regional climate model (RCM) of Representative Concentration Pathway (RCP) 4.5. Storm Water Management Model (SWMM) with quality and quantity scenarios were developed and the results of the present year scenario were validated using a questionnaire survey. Intensity Duration Frequency (IDF) curves were generated to forecast the observed rainfall and climate model data for various return periods, i.e., 2, 5, 10, 50, and 100 years for future climatic scenarios (2030s and 2050s). The model results show that infiltration is only 3–7% of the total rainfall in the study areas and the maximum blockage and reduction in conduit capacity were estimated to be 65 and 40%, respectively. This paper presents the observations and suggestions for improving the conduit efficiency from an Integrated Water Management Approach (IWMA).

  • Imperviousness of the study area was calculated based on LULC classification using Google Earth.

  • Bias correction was carried out for the climate data using observed rain gauge data.

  • Evaluation of the impact of pollutants load in the conduit carrying capacity using the SWMM.

  • Questionnaire survey outcomes were used to understand the past extreme events and also used in conjunction with the model outputs to derive suggestions.

Graphical Abstract

Graphical Abstract
Graphical Abstract

Unplanned infrastructure development and paved surface in urban area increase imperviousness and it is one of the fundamental reasons for urban floods (Ahamed & Agarwal 2019). Land-use changes such as loss in natural water bodies, encroachment of rivers/streams, and other drainage channels, along with uncontrolled development of built-up areas have been identified as contributory factors for the increasing flood events in Chennai City (Gupta & Nair 2010). The built-up area in Chennai City has increased up to 173.83% in 2017 which has significantly increased the imperviousness (Mathan & Krishnaveni 2020). For a 44% increase in the imperviousness area, the total runoff volume increases by 72% in an urban catchment (Lee & Heaney 2003). It is identified that Land Use Land Cover (LULC) changes and an increase in imperviousness of an urban region increase the urban stormwater runoff by manyfold.

The urban regions in developing countries are more prone to the adverse effects of climate change due to their vulnerability, less adaptivity, and increase in population density (Pickett et al. 2001). Repercussions of climate change on the urban areas led to a reduction in the water quantity, quality, and increased flood events (Miller & Hutchins 2017). The number of extreme precipitation events is likely to be increased and there will be a decrease in the annual mean precipitation in the Asian region landmass, as per the Intergovernmental Panel on Climate Change (IPCC) Assessment Report 5 (IPCC 2013). This shows the importance of using climate models in urban studies. The daily observed rainfall data also can be used to study the impact of climate change in a major urban area (Shahid et al. 2016) and the trend of rainfall in Chennai City can be studied using Coordinated Regional Climate Downscaling Experiment (CORDEX) data for Representative Concentration Pathways (RCPs) 4.5 and 8.5 (Athira & Agilan 2020).

Increase in the imperviousness and climate change made the stormwater network-related study a complex process. The models used for these studies are the Hydrologic Modelling System (HEC-HMS), Soil and Water Assessment Tool (SWAT), Storm Water Management Model (SWMM), MIKE System Hydrologique European (MIKE SHE), and MIKE URBAN. Among these SWMM is one of the most widely used models for stormwater network studies (Janga et al. 2007; Ahamed & Agarwal 2019; Brendel et al. 2021). The SWMM can be used to study the impact of climate change on the quality of stormwater runoff and the flood-carrying capacity of the stormwater drain network in urban areas for both the present and future scenarios (Leboutillier et al. 2000; Andimuthu et al. 2019; Sonavane et al. 2020). The SWMM can also be used to estimate the pollutants load using EMC values for different land use and land cover data (Ryu et al. 2016; Tu & Smith 2018; Tuomela et al. 2019; Choi et al. 2021). The four water quality indicators, namely biological oxygen demand (BOD), total suspended solids (TSS), total phosphate (TP), and total nitrate (TN) were predominantly used in the evaluation of stormwater quality in urban regions in various studies (LeBoutillier et al. 2000; Li et al. 2015; Tu & Smith 2018; Tuomela et al. 2019; Choi et al. 2021). So, we have taken only these four quality indicators in this study.

Several studies (Agarwal & Kumar 2019; Hussain et al. 2022) were carried out to evaluate the impact of imperviousness and climate change on the urban stormwater runoff and storm drain network using the extreme precipitation event. It is noted that some of these studies were carried out without adjusting the climate data for the local climatic conditions. These approaches will not be adequate to understand the future impact on the stormwater drain network completely. In order to estimate the accurate impact of climate change on the stormwater runoff and pollutant load, bias correction needs to be carried out for the regional climate model (RCM) data to adjust for the local climatic conditions. To the best of our understanding, the combined impact of climate change, surface area imperviousness, and pollutants load need to be studied for the effective evaluation of storm drain network performance. This paper presents the findings based on this combined impact on the storm drain network performance with an integrated water management approach.

Study area description

Chennai is an ancient city with rich history and cultural heritage. The town was named Madras by the British East India Company on 22 August 1639. Today, it is one of the most populous urban agglomerations and is located on the East Coast of India. Chennai mainly depends on the North East (NE) monsoon during which around 65% of the annual rainfall is received every year and it is prone to flood during this season. Chennai City is now divided into 15 zones, and each zone has more than 10 wards accounting for 200 wards in the city. Among these wards, ward 116 (Triplicane) and ward 152 (Valasaravakkam) from zones 9 and 11, respectively, were chosen as the study area and these wards are within the urban limits of Chennai City.

Triplicane (ward 116), also known as Thiruvallikeni, is one of the oldest neighbourhoods of Chennai. It is situated on the Bay of Bengal coast, about 3 km (1.86 miles) from Fort St. George, and the average elevation of the neighbourhood is 14 m above sea level. Along with Mylapore and the surrounding regions, Triplicane is historically much older than Chennai City itself, and it is mentioned in records as early as the Pallava period (Ramanujam 2011). Valasaravakkam (ward 152) is a residential suburb and a municipality in the Chennai district and it is located approximately 15 km from the geographical center of Chennai City. The study area is presented in Figure 1.
Figure 1

Study area zones in Chennai City.

Figure 1

Study area zones in Chennai City.

Close modal

Data used for the study

The secondary data collected for this study from various sources are presented in Table 1. Precipitation data for the study area were collected from the Institute for Water Studies (IWS) and the gridded data were downloaded from the online database of the Indian Meteorological Department (IMD) Pune. The SWMM uses the topographical, LULC, soil type, and storm network data for structuring the model. The LULC data required for this study were taken from Google Earth for the study area and it was georeferenced and digitized using ARCGIS10.5. The elevation and slope values for the study area were taken from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) Digital Elevation Model (DEM) data. The soil property data for the study area are extracted from the Harmonized World Soil Database (HWSD) of the Food and Agriculture Organization (FAO). The stormwater network map was downloaded from the Greater Chennai Corporation online database, and it was digitized using QGIS 3.16.

Table 1

Data used for the study and their source

DataTime period/versionSpatial resolutionSources
Precipitation 1990–2020 0.25° × 0.25° IWS & IMD 
Climate data 2020–2100 0.44° × 0.44° ESGF-CORDEX 
DEM data Version 3 30 m × 30 m ASTER DEM 
LULC Real time 15 cm Google Earth 
Soil property Version 1.2 1 km FAO 
Storm network – – Chennai Corporation 
DataTime period/versionSpatial resolutionSources
Precipitation 1990–2020 0.25° × 0.25° IWS & IMD 
Climate data 2020–2100 0.44° × 0.44° ESGF-CORDEX 
DEM data Version 3 30 m × 30 m ASTER DEM 
LULC Real time 15 cm Google Earth 
Soil property Version 1.2 1 km FAO 
Storm network – – Chennai Corporation 

The global climate model (GCM) data are available for different RCP emission scenarios, namely RCP 2.6, RCP 4.5, RCP 6.0, and RCP 8.5. In which, the RCP 2.6 scenario is very stringent with carbon dioxide emissions because it assumes CO2 emission to decline from 2020 and becomes zero by 2100. RCP 4.5 is a medium stabilization scenario with a carbon dioxide emission equivalent concentration range of 580–720 ppm for the year 2100. RCP 6.0 is a stabilization scenario where total radiative force is stabilized before 2100 without overshoot employment of a range of technologies and strategies for reducing greenhouse gas emissions, whereas RCP 8.5 is a very high emission scenario with a carbon dioxide equivalent concentration of 1,000 ppm for the year 2100 (IPCC 2014). One can observe, various measures being taken around the globe to cut down carbon emissions by 2050, and the RCP 4.5 model seems to reflect these measures for the future scenario. So, in this study, we have only considered RCP 4.5 data for our modelling purposes.

In this study, climate data from the World Climate Research Programme (WCRP), CORDEX dataset for South Asia (WAS 44) from the Atmosphere Ocean coupled General Circulation Model (AOGCM) with a resolution of about 50 km were shortlisted for this study. The models were further filtered based on the criteria such as the availability of historical paths, different RCP values, and models with the horizon year 2100. Based on the review of earlier studies (Komaragiri & Kumar 2014; Shashikanth & Sukumar 2017; Bhadran et al. 2020; Vishwakarma et al. 2022) carried out in India for evaluating the impact of climate change on precipitation, four models, namely CNRM-CM5, ICHEC-EC-EARTH, MIROC-MIROC5, and MPI-M-MPI-ESM-LR, were chosen for this study and the details are presented in Table 2. All the necessary data were georeferenced, georectified, geocorrected, and the data resolution was homogenised using the data resampling method in ARCGIS 10.5 before preparing the model inputs and analysis.

Table 2

Description of selected RCMs

RCMGCM nameInstitution
CORDEX CNRM-CM5 National Centre for Meteorological Research and Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique 
MIROC-MIROC5 National Institute for Environmental Studies, and Japan Agency for Marine-Earth Science and Technology, MIROC), Japan 
ICHEC-EC-EARTH Consortium of European Research Institution and Researchers 
MPI-M-MPI-ESM-LR Max Planck Institute for Meteorology, low resolution (Germany) 
RCMGCM nameInstitution
CORDEX CNRM-CM5 National Centre for Meteorological Research and Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique 
MIROC-MIROC5 National Institute for Environmental Studies, and Japan Agency for Marine-Earth Science and Technology, MIROC), Japan 
ICHEC-EC-EARTH Consortium of European Research Institution and Researchers 
MPI-M-MPI-ESM-LR Max Planck Institute for Meteorology, low resolution (Germany) 

Bias correction

GCM is the fundamental source of information for developing climate scenarios and the basis for the impact assessment due to climate change at all levels. For impact analysis, GCM output cannot be used directly because the model exhibits a systematic error (bias). It is necessary to do a bias correction for the precipitation data to use it for hydrological studies. The RCM is a sub-area model of the GCM, and its accuracy depends on the accuracy of the GCM for the corresponding sub-area.

In this study, RCM data were used for the climate scenarios. There are various methods available for bias correction of RCM simulated data. In general, the available methods are classified into parametric and non-parametric methods. The parametric methods estimate the extreme precipitation more accurate than the non-parametric methods (Teutschbein & Seibert 2012; Kumar et al. 2022). The various parametric methods used for bias correction are (1) Linear Scaling (LS); (2) Local Intensity (LOCI) scaling; (3) Power Transformation (PT); (4) Delta–Change (DC); and (5) Distributing Mapping (DM) method.

The LS method is one of the simplest bias correction methods, which is used in several studies (Schmidli et al. 2006; Teutschbein & Seibert 2012), but this method has a disadvantage in that it bias corrects the RCM simulated data only based on the monthly mean precipitation values. The LOCI scaling method improves the raw data by correcting the wet frequency and intensity based on the monthly wet day threshold. However, this method is only suitable if the study area has several drizzle days (Schmidli et al. 2006). The PT method corrects the variance using the standard deviation of the data (Teutschbein & Seibert 2012), but this method is ineffective in correcting the wet day data. The DC method is a widely used method to correct the precipitation of the RCM data (Bosshard et al. 2011; Teutschbein & Seibert 2012). However, this method does not allow for a change in the extreme precipitation value due to its limitation of using a simple transformation function and multiplicative function to correct the precipitation. The DM method corrects the RCM simulated data by adjusting the mean, standard deviation, and quantiles (Teutschbein & Seibert 2012). The DM method uses a quantile mapping technique to improve the fitting of RCM simulated data to observed data, regardless of extreme values. So based on our best understanding, this method provides a more accurate and micro-level estimate of the data required for the study area.

The CMhyd tool helps to correct the precipitation data with respect to the metrological gauges in the study area (Takele et al. 2021). The CMhyd tool is used for bias correction of the precipitation data using the DM method.

Significance test for the rainfall data

Various statistical techniques are available for testing the significance of the forecasted rainfall from climate models. These techniques can be broadly classified into parametric and non-parametric methods. The parametric methods are sensitive to the skewness and also assume a normal distribution of the variables, whereas the non-parametric method does not follow the normal distribution and is not affected by the skewness of the data. So, in this study, the widely used non-parametric method Mann–Kendall and Sen's slope method (Panda & Sahu 2019; Rathnayake 2019) was used to show the statistical significance of the projected precipitation from different climate models for the confidence intervals of 95 and 90%.

IDF curve generation

An urban area with a low level of flood risk should be studied for a return period of 5–10 years, whereas an urban area with a high level of flood risk should be studied for a return period of 50–100 years (Ponce 2016). This approach has been adopted in various studies for assessing the impact of climate change in the urban region (Athira & Agilan 2020; Andimuthu et al. 2019). Since 2000, Chennai is more prone to flood with an extreme flood event in 2015. So, in this study, IDF curves were generated for both the observed and future climate data from the GCMs for five different return periods 2, 5, 10, 50, and 100 years. From the daily rainfall data, a shorter duration of rainfall was calculated using the empirical formula given by IMD which is shown in Equation (1). Hourly rainfalls for this study, i.e., 1, 2, 4, 6, 12, and 24 h single event were calculated from the annual mean precipitation value.
(1)
where Pt is the required rainfall depth in mm at t-h duration, P24 is the daily rainfall in mm, and t is the duration of rainfall in h.
Among the distribution models such as the Generalized Extreme Value, Gumbel or Extreme Value Type 1 (EV1), Log-Normal and the Log Pearson type III distributions, the Gumbel model provides the best fit for extreme values. Because, Gumbel distribution was found to be effective for smaller sample sizes (Onen & Bagatur 2017) as well as for infinite sample sizes (Samantaray & Sahoo 2020). As this study evaluates the impact of climate change in an urban region from 2 to 100 years return period, the Gumbel's distribution method was chosen for generating the IDF curves (Andimuthu et al. 2019) using Equation (2).
(2)
The intensity or the depth of rainfall for the given return period was calculated to estimate the IDF. The intensity of the rainfall was calculated using the frequency factor derived using Equation (3).
(3)
where KT is the frequency factor, XT is the rainfall intensity at a given return period, is mean at a particular time, and S is the standard deviation.

SWMM simulation

The SWMM is a dynamic rainfall runoff simulator used to estimate a single-event or a continuous simulation, estimate storm runoff quantity, and quality for urban areas (Rossman 2005). Pollutant loading and washoff from sub-catchment areas are determined using the land use assigned to the sub-catchment (Rossman 2010). In this study, the time-series rainfall data, the infiltration flow derived using the curve number (CN) method, and the flow routing developed using dynamic wave method were used to estimate the surface runoff. For pollutants load estimation, land use was assigned to the SWMM using the mixed-land use method. The power function method in SWMM was used to calculate the pollutants load accumulated in the sub-catchment and the washoff pollutants load from the sub-catchment was estimated using the Event Mean Concentration (EMC) method in SWMM.

Blockage estimation

Blockage in the conduit reduce its discharge carrying capacity. Fathy et al. (2020) calculated the reduction in the carrying capacity of the conduits based on the percentage of reduction in the discharge volumes from the conduits as an experimental study. In this study, the percentage reduction in the carrying capacity of the conduits were calculated from the percentage reduction in the volume of discharge from the conduits using the following equations.
(4a)
(4b)
(4c)
where QR is the percentage of the discharge reduction and QSWMM is the discharge estimated by the SWMM model.

Performance evaluation of the models

In this study, the Nash–Sutcliffe Model Efficiency (NSE) coefficient, coefficient of determination (R2), and Percentage of Bias (PBIAS) were used for the performance evaluation of the GCMs before and after bias correction. The NSE coefficient value is considered to be satisfactory if the value is greater than 0.5. The value of R2 is considered to be satisfactory if the value is closer to 1. The optimal value of PBIAS is ‘0’ and the PBIAS value between ±25 is considered to be satisfactory. The values are adopted based on the performance rating given by Moriasi et al. (2007). In this study, the performance evaluation of climatic models before and after the bias correction was carried out using the following equations (Fang et al. 2015; Mendez et al. 2020).
(5)
(6)
(7)

Parameters used in the SWMM

ASTER DEM data were used to find the elevation and slope for the study area. The sub-catchments for the study area were delineated using the Arc Hydro tool available in Arc GIS 10.5 by adjusting the distribution of conduit features. CN was calculated using the LULC, ASTER DEM, and Soil Property data with the help of the Arc Hydro tool. Manning's Roughness coefficient value was taken from the study carried out by Andimuthu et al. (2019). The imperviousness area were calculated using the vector analysis method in QGIS 3.16. For quality analysis, the pollutant EMC values for BOD, TSS, TN, and TP were taken from various studies carried out for urban areas (Tu & Smith 2018; Tuomela et al. 2019; Choi et al. 2021). The parameters required for the SWMM were calculated using various methods and software as explained above and it is presented in Table 3.

Table 3

Parameter values used in the SWMM

ParameterValues for ward 116Values for ward 152
Slope for conduit 0.01–0.028 0.005–0.01 
Imperviousness percentage 70–89 67–84 
Impervious manning 0.024 0.024 
Pervious manning 0.018 0.018 
Impervious depression storage 0.05–0.10 0.05–0.1 
Pervious depression storage 0.10–0.20 0.10–0.20 
Curve Number 81–100 75–86 
Manning's roughness for conduit 0.015 0.015 
TSS (EMC) 28.8–273 28.8–273 
TP (EMC) 0.092–0.8 0.092–0.8 
TN (EMC) 1.048–2.42 1.048–2.42 
BOD (EMC) 200–1,000 200–1,000 
ParameterValues for ward 116Values for ward 152
Slope for conduit 0.01–0.028 0.005–0.01 
Imperviousness percentage 70–89 67–84 
Impervious manning 0.024 0.024 
Pervious manning 0.018 0.018 
Impervious depression storage 0.05–0.10 0.05–0.1 
Pervious depression storage 0.10–0.20 0.10–0.20 
Curve Number 81–100 75–86 
Manning's roughness for conduit 0.015 0.015 
TSS (EMC) 28.8–273 28.8–273 
TP (EMC) 0.092–0.8 0.092–0.8 
TN (EMC) 1.048–2.42 1.048–2.42 
BOD (EMC) 200–1,000 200–1,000 

Questionnaire survey

A questionnaire is a research tool used to conduct surveys and it includes close-ended, open-ended, short-form, and long-form questions. In this study, the questionnaire survey includes specific questions to understand the past extreme events, present situation of the conduit system, and grates in the study area from the people's perspective. Due to the pandemic situation, some of the survey samples were collected via a telephonic interview and the rest of the samples using household surveys. Efforts were taken to ensure random sampling in both the methods. Around 50 samples were collected for each of the study area wards. A set of 22 questions were prepared as a questionnaire for this study, and it is presented in Figure 2.
Figure 2

Questionnaire used for the survey.

Figure 2

Questionnaire used for the survey.

Close modal
The analysis of responses for ward 116 is presented in Figure 3. The survey outcomes show that most of the ward 116 was flooded heavily only during 2015 extreme rainfall. It was also observed that, there was no water stagnation in some parts of the ward, whereas in the rest of the area there was water stagnation from 1 to 5 ft. The duration of water stagnation varied spatially from 1 day up to 1 week and there was a heavy loss of property in the areas where there was prolonged stagnation of water. Improper maintenance of stormwater drain system and improper solid waste management are identified as the main reasons for floods in this ward. It was also observed that the majority of households in this ward lack a proper rain water harvesting system. This survey suggests the enforcement of installation of a proper rain water harvesting system, which not only reduces the runoff conditions but also improves the ground water level and water quality in this ward.
Figure 3

Reponses from questionnaire survey for the ward 116.

Figure 3

Reponses from questionnaire survey for the ward 116.

Close modal
The analysis of responses for ward 152 is presented in Figure 4. Similar to ward 116, ward 152 was flooded heavily only during 2015. It was also observed that, most of ward 152 had water stagnation up to 10 ft and only a negligible proportion of the total ward area had no water stagnation. Majority of the ward had water stagnation for at least a day and in some cases, there was water stagnation for more than a week which led to severe property loss. Ward 152 is a newly extended urban area and the unavailability of a proper stormwater drain network was identified as the main reason for water stagnation. After the flood, the stormwater drain network for this ward was partially completed in 2018 and most of the households in ward 152 have proper rain water harvesting systems.
Figure 4

Reponses from questionnaire survey for the ward 152.

Figure 4

Reponses from questionnaire survey for the ward 152.

Close modal

Present scenario results

Estimation of imperviousness area

In this study, LULC is categorized into five land-use types, namely Recreation/Park, Developed (Residential and Road), Green Space, Open Space and Waterbody, based on the Indian Space Research Organization (ISRO) land use classification system. LULC classification for the study area wards 116 and 152 was carried out with the aid of Google Earth, and it is presented in Figure 5. From the classified LULC, the pervious and the impervious areas for study area wards were calculated using the area of each LULC type. Impervious area for the study area wards 116 and 152 was calculated as 89.22 and 83.27%, respectively, and the values are more or less similar for both the wards. LULC characteristics for the study area wards are presented in Table 4.
Table 4

Impervious area for the study area wards

LULU characteristicsWard 116
Ward 152
Area (m2)Percentage (%)Area (m2)Percentage (%)
Developed land use types 
Residential 6,05,541 71.38 11,40,286 66.61 
Road 1,21,108 17.85 2,28,057 16.85 
Undeveloped land use types 
Park/playground 5,703 0.84 4,007 0.296 
Open space 41,417 6.10 1,40,512 10.38 
Green space 19,941 2.94 35,719 3.647 
Water bodies 6,074 0.001 32,663 2.41 
Total area (A) 6,78,676 100 13,53,187 100 
Total pervious area (B) 73,135 10.78 2,12,901 16.73 
Total impervious area (A−B) 6,05,541 89.22 11,40,286 83.27 
LULU characteristicsWard 116
Ward 152
Area (m2)Percentage (%)Area (m2)Percentage (%)
Developed land use types 
Residential 6,05,541 71.38 11,40,286 66.61 
Road 1,21,108 17.85 2,28,057 16.85 
Undeveloped land use types 
Park/playground 5,703 0.84 4,007 0.296 
Open space 41,417 6.10 1,40,512 10.38 
Green space 19,941 2.94 35,719 3.647 
Water bodies 6,074 0.001 32,663 2.41 
Total area (A) 6,78,676 100 13,53,187 100 
Total pervious area (B) 73,135 10.78 2,12,901 16.73 
Total impervious area (A−B) 6,05,541 89.22 11,40,286 83.27 
Figure 5

LULC map classifications of wards 116 and 152.

Figure 5

LULC map classifications of wards 116 and 152.

Close modal

Runoff estimation using SWMM for observed rainfall data

Ward 116

Study area ward 116 was delineated into several sub catchments using Arc Hydro tool in ArcGIS. The total area of 67.99 ha of ward 116 was delineated into 15 sub-catchments as shown in Figure 6 and parameters such as area, width, and average slope were derived for each sub-catchment. The outlet of the sub-catchment is connected with the nodes. Imperviousness for the ward 116 is 89.22% and the elevation of this ward ranges from 5 to 24 m. The SWMM was simulated as a quantity model for both single event and continuous simulation using the observed rainfall data from 1998 to 2020. The model run for ward 116 was carried out with the following characteristics, viz., 15 sub-catchments, 13 junctions, 13 conduits, and 4 outfalls. The continuous simulation model for the study area ward 116 was simulated for 22 years. The model results show that most of the rainfall resulted in runoff and only a small amount of rainfall (about 3%) infiltrated.
Figure 6

Details of subcatchment, conduits, and junctions in the study area ward 116.

Figure 6

Details of subcatchment, conduits, and junctions in the study area ward 116.

Close modal

The single event model runs were carried out for both extreme rainfall conditions in 2015 and minimum rainfall conditions in 2012. The event-based model results show that conduits C3, C9, and C13 (storm drain network) flooded for more than a day, and the remaining conduits flooded for 1–5 h. Out of 13 nodes, 12 nodes flooded in the 2015 extreme rainfall model. In specific, nodes E2 and E3 flooded for more than 10 h, and other nodes flooded for 1–5 h. The conduits (storm drain network) C3 and C13 are in Dr Beasant Road and C9 in Bharathi Salai, respectively, shown in Figure 6. All the outfalls are connected to Buckingham Canal and the maximum and minimum outflow from ward 116 was observed in E1OUT3 and F1OUT4, respectively.

The percentage of infiltration in the minimum scenario model (2012) is similar to other modelled scenarios. In the minimum rainfall model, conduit C3 flooded for more than an hour, and the nodes E2 and E3 flooded for 1 h. All three model results show that infiltration is very minimum due to the high imperviousness of the study area and the results are presented in Table 5.

Table 5

Runoff quantity results for observed rainfall data from the SWMM for the ward 116

Model results for the Ward 116Continuous model runoff quantity (1998–2020)Event-based model runoff quantity – extreme scenario (2015)Event-based model runoff quantity – minimum scenario (2012)
Average annual depth (mm)Depth (mm)Depth (mm)
Precipitation 1,358.39 1,168.49 222.29 
Infiltration loss 71.32 9.252 22.36 
Surface runoff 1,287.02 1,109.252 200.43 
Model results for the Ward 116Continuous model runoff quantity (1998–2020)Event-based model runoff quantity – extreme scenario (2015)Event-based model runoff quantity – minimum scenario (2012)
Average annual depth (mm)Depth (mm)Depth (mm)
Precipitation 1,358.39 1,168.49 222.29 
Infiltration loss 71.32 9.252 22.36 
Surface runoff 1,287.02 1,109.252 200.43 

A typical example of the Water Elevation Profile for Dr Beasant Road during extreme rainfall conditions (2015) showing the failure of the existing drainage network is presented in Figure 7. The figure shows that the nodes E2 and E1Out3 and conduits C3 & C13 reached full capacity on the maximum rainfall day and the excess surface runoff stagnated in the near vicinity of the nodes.
Figure 7

Water elevation profile for the conduits in Dr Beasant Road.

Figure 7

Water elevation profile for the conduits in Dr Beasant Road.

Close modal

Ward 152

The total area of ward 152 is 135.71 ha which was further delineated into 54 sub-catchments as shown in Figure 8 and parameters such as area, width, and average slope were derived for each sub-catchment. Imperviousness of the ward 152 is 83.27% and the elevation of the ward ranges from 7 to 26 m. The model runs were performed with the following characteristics, viz., 54 sub-catchments, 49 junctions (nodes), and 51 conduits (storm drain networks) for the observed rainfall data. The stormwater drain system is under construction for the nearby wards and the exact outfall is still unknown, so based on the available information from Greater Chennai Corporation only the flow direction is assumed for the study area ward 152. The SWMM continuous model was carried out only for 2 years (2018–2020) for the observed rainfall data, this is because the stormwater drain system for the ward 152 was completed only by 2018 and the conduits for the surrounding wards are yet to be completed. The model results show that the infiltration is only 7% the and rest of the rainfall resulted in surface runoff. Due to the absence of stormwater drain system in ward 152 in 2015, the event-based model scenarios were not carried out for the ward 152.
Figure 8

Details of subcatchment, conduits, and junctions in the study area ward 152.

Figure 8

Details of subcatchment, conduits, and junctions in the study area ward 152.

Close modal

The SWMM results for the observed rainfall data show that the infiltration is only 3 and 7% for the wards 116 and 152, respectively, the highly impervious nature of the study area wards reduced the infiltration of stormwater runoff significantly. The runoff for both the study area wards was more than 90% which is mainly because of the low infiltration. From the questionnaire survey, we found that the study area wards also lack a proper rainwater harvesting system which leads to an increase in the stormwater runoff.

Validation of study

Due to the unavailability of a discharge gauge for the study area wards, the event-based extreme rainfall model 2015 was validated using real-time flood mapping data and a questionnaire survey. The survey results show that ward 116 was heavily flooded during the 2015 extreme rainfall event and the streets worst affected were near Dr Beasant Road and Bharathi Salai. The SWMM results show that the conduits C3, C9, and C13 flooded for more than a day and it is important to note that the conduits C3 and C13 are located in Dr Beasant Road and C9 in Bharathi Salai. The real-time flood map (Karmegam et al. 2021) for ward 116 during the 2015 extreme rainfall event shows a water depth between 0.98 and 1.89 m. The depth of water estimated by the SWMM was 1.109 m which is within the specified range from the real-time flood map. The performance evaluation based on the surface runoff depth was found to be statistically satisfactory with NSE, R2, and PBIAS values of 0.807, 0.87, and 11.52, respectively. Hence, the SWMM developed is validated and observed to have produced accurate results.

Estimation of reduction in conduit efficiency for present scenario

Ward 116

The pollutants load was estimated using the SWMM quality model and it is used to find the reduction in conduit efficiency due to blockage. For the present scenario, the observed rainfall data for 20 years (1998–2020) was used to estimate the pollutants load. The pollutants load was estimated from the SWMM in terms of TSS, TN, TP, and BOD. The average annual pollutants load and daily pollutants load estimated for the entire catchment area from the SWMM for ward 116 is presented in Table 6. The BOD and TSS values are higher than the TN and TP pollutants load because ward 116 is completely residential. The pollutants load such as BOD, TN, and TP only affects the quality of the drain storm, but TSS not only affects the quality of the drain storm and it also blocks the conduit due to large suspended solids in it.

Table 6

SWMM results

Pollutant load parametersAnnual pollutant load (kg)Daily pollutant load (kg)
TSS 88,573 242.66 
TP 391.1 1.07 
TN 725.8 1.99 
BOD 76,162 208.66 
Pollutant load parametersAnnual pollutant load (kg)Daily pollutant load (kg)
TSS 88,573 242.66 
TP 391.1 1.07 
TN 725.8 1.99 
BOD 76,162 208.66 

In the current study, the estimation of reduction in conduit carrying capacity is carried out only using TSS pollutant load. The change in the volume of discharge for all the 13 conduits in ward 116 was estimated using the mathematical formulas (Equation (4a)–(4c)). The change in maximum discharge and reduction in the carrying capacity of the conduits are presented in Figure 9 for the observed rainfall data. Conduits C3, C7, C8, C9, and C10 have a noticeable reduction in the carrying capacity, but there is a negligible change in the carrying capacity of other conduits. The reduction in the carrying capacity of the conduits due to pollutant load in ward 116 ranges from 1 to 28.09%. Conduit C3 has a blockage of about 55% of its area and the carrying capacity was reduced by 28.09%. Experimental study carried out by Fathy et al. (2020) shows a reduction in the efficiency of the conduits from 2.27 to 24.06% which is similar to the reduction in conduit efficiency estimated in this study.
Figure 9

Change in maximum discharge of conduits for observed rainfall data.

Figure 9

Change in maximum discharge of conduits for observed rainfall data.

Close modal

Forecast scenario

Bias correction for RCMs

The GCM/RCM data should be bias-corrected with respect to the local climatic condition of the selected study area which improves the accuracy of the data. This bias corrected data will help to analyze the impact of climate change on water resources precisely. The average rainfall data for a series of 20 years (1998–2018) from the RCMs were bias corrected with the aid of the average observed rainfall data for 20 years (1998–2018) using the DM method. The average rainfall data for a series of 20 years (1998–2018) before and after bias correction are presented in Figure 10.
Figure 10

Bias correction adjustment of RCM data for the study area wards 116 and 152.

Figure 10

Bias correction adjustment of RCM data for the study area wards 116 and 152.

Close modal

The performance evaluation of the GCM before and after bias correction is presented in Table 7. The results show that the RCM data fit well with the observed data after bias correction. The value of R2 and NSE has improved significantly with the bias corrected data. For example, the NSE value of MPI-M-MPI-ESM-LR has improved from 0.59 to 0.94 after bias correction. Table 7 also shows that the PBIAS values of the corrected data lies within the recommended range. Hence, the model performance is found to be statistically satisfactory.

Table 7

Performance evaluation of RCM data before and after bias correction

Climatic modelRaw data
Bias-corrected data
NSER2PBIASNSER2PBIAS
CNRM-CM5 0.845 0.762 −25.75 0.969 0.960 10.10 
ICHEC-EC-EARTH 0.754 0.665 −32.74 0.959 0.97 14.54 
MIROC-MIROC5 0.791 0.794 −44.74 0.952 0.956 5.40 
MPI-M-MPI-ESM-LR 0.599 0.682 −46.86 0.937 0.951 11.26 
Climatic modelRaw data
Bias-corrected data
NSER2PBIASNSER2PBIAS
CNRM-CM5 0.845 0.762 −25.75 0.969 0.960 10.10 
ICHEC-EC-EARTH 0.754 0.665 −32.74 0.959 0.97 14.54 
MIROC-MIROC5 0.791 0.794 −44.74 0.952 0.956 5.40 
MPI-M-MPI-ESM-LR 0.599 0.682 −46.86 0.937 0.951 11.26 

Projected change in the precipitation

The 10-year average rainfall pattern forecasted by selected GCMs for RCP 4.5 from 2021 to 2090 (70 years) is presented in Figure 11. The projected precipitation is different for each GCM, and it is mainly because of the underlying assumption of each GCM. To justify the trend statistically, a performance evaluation was carried out for the projected precipitations. The statical significance of the projected change in precipitation was carried out for the confidence intervals of 95 and 90%. For a confidence interval of 95%, except ICHEC-EC-EARTH all other models are statistically insignificant. So, we checked the significance of the models for a confidence interval of 90%, and all models except MPI-M-MPI-ESM-LR were found to be statistically significant. For both the confidence intervals the base line data were found to be statistically insignificant. The significant test result details are presented in Table 8.
Table 8

Significant test result details for baseline and RCMs

Modelsα = 0.05
α = 0.1
Z-statSlope (β)Statistical significanceZ-statSlope (β)Statistical significance
Base Line (1998–2018) 1.177 16.34 No 1.177 16.34 No 
CNRM-CM5 1.896 7.74 No 1.896 7.92 Yes 
ICHEC-EC-EARTH 2.382 4.65 Yes 2.382 5.11 Yes 
MIROC-MIROC5 1.754 7.29 No 1.754 7.28 Yes 
MPI-M-MPI-ESM-LR −0.304 −0.74 No −0.304 −0.21 No 
Modelsα = 0.05
α = 0.1
Z-statSlope (β)Statistical significanceZ-statSlope (β)Statistical significance
Base Line (1998–2018) 1.177 16.34 No 1.177 16.34 No 
CNRM-CM5 1.896 7.74 No 1.896 7.92 Yes 
ICHEC-EC-EARTH 2.382 4.65 Yes 2.382 5.11 Yes 
MIROC-MIROC5 1.754 7.29 No 1.754 7.28 Yes 
MPI-M-MPI-ESM-LR −0.304 −0.74 No −0.304 −0.21 No 
Figure 11

Projected rainfall data by RCMs from 2021 to 2090.

Figure 11

Projected rainfall data by RCMs from 2021 to 2090.

Close modal

Design storm

Design storm was estimated for various return periods using the observed (Nungambakkam rain gauge) and future climate data. IDF (Intensity Duration Frequency) curves were developed for different duration (1, 2, 4, 6, 12, and 24 h) and for five return periods of 2, 5, 10, 50, and 100. The mean and standard deviation values for the observed rainfall data are presented in Table 9. The KT values used in the model are −0.164, 0.719, 1.304, 2.592, and 3.136 for the return periods 2, 5, 10, 50 and 100 years, respectively. The IDF curves developed for different return periods are presented in Figure 12 and the intensity of rainfall for 2-year and 100-year return periods estimated from the IDF curves are presented in Table 10. It was found that the intensity of rainfall forecasted using ICHEC-EC-EARTH model for 2-year and 100-year return periods were close to the forecasted rain gauge data.
Table 9

Mean and standard deviation for observed rainfall data for various duration

Rainfall durationMeanStandard deviation
1 h 46.17 25.32 
2 h 17.58 31.90 
4 h 73.29 39.53 
6 h 83.90 45.25 
12 h 105.70 57.96 
24 h 133.18 73.02 
Rainfall durationMeanStandard deviation
1 h 46.17 25.32 
2 h 17.58 31.90 
4 h 73.29 39.53 
6 h 83.90 45.25 
12 h 105.70 57.96 
24 h 133.18 73.02 
Table 10

Intensity of rainfall for 2-year and 100-year return period

DataIntensity of rainfall – 2-year return period (mm/h)Intensity of rainfall – 100-year return period (mm/h)
Observed rainfall 42.01 157.08 
CNRM-CM5 84.57 331.38 
ICHEC-EC-EARTH 54.36 263.52 
MIROC-MIROC5 84.45 294.23 
MPI-M-MPI-ESM-LR 73.19 316.31 
DataIntensity of rainfall – 2-year return period (mm/h)Intensity of rainfall – 100-year return period (mm/h)
Observed rainfall 42.01 157.08 
CNRM-CM5 84.57 331.38 
ICHEC-EC-EARTH 54.36 263.52 
MIROC-MIROC5 84.45 294.23 
MPI-M-MPI-ESM-LR 73.19 316.31 
Figure 12

IDF curve for observed and selected RCM.

Figure 12

IDF curve for observed and selected RCM.

Close modal

Estimation of stormwater runoff and conduit efficiency for single event

In order to estimate the carrying capacity of the conduits during various flood scenarios, single-event SWMM runs were conducted for five different return periods (2, 5, 10, 50, and 100 years) for both observed and forecasted rainfall data for the study area wards 116 and 152. In addition, the effect of pollutants load on the conduit efficiency was also studied. The model results for various rainfall return periods are presented in Table 11. The simulation results show that most of the rainfall became runoff and the main reason for the flood are inadequate stormwater drain and impervious nature of the study areas.

Table 11

SWMM quantity model results for wards 116 and 152

Model ResultsWard 116
Ward 152
2-year return period100-year return period2-year return period100-year return period
Rainfall (mm) 435.68 1,627.37 435.67 1,627.37 
Runoff volume (mm) 426.67 1,618.19 420.96 1,613.11 
Infiltration loss (mm) 8.99 9.20 14.71 14.42 
Model ResultsWard 116
Ward 152
2-year return period100-year return period2-year return period100-year return period
Rainfall (mm) 435.68 1,627.37 435.67 1,627.37 
Runoff volume (mm) 426.67 1,618.19 420.96 1,613.11 
Infiltration loss (mm) 8.99 9.20 14.71 14.42 

The number of conduits flooded in wards 116 and 152 without considering the washed off pollutants is presented in Figures 13 and 14. The results show that all the conduits have flooded for a considerable amount of time in which few conduits flooded for more than 20 h in both the wards. It was observed that the 100-year return period storm caused the maximum flood in the conduits. This shows that, the effect of climate change on the urban stormwater runoff and its impact on the storm drain system increases with increase in the intensity of rainfall. So, evaluation of conduits for 100-year return period floods from climate models could provide more insight in managing flood events.
Figure 13

Number of conduits flooded under different return periods in ward 116.

Figure 13

Number of conduits flooded under different return periods in ward 116.

Close modal
Figure 14

Number of conduits flooded under different return periods in ward 152.

Figure 14

Number of conduits flooded under different return periods in ward 152.

Close modal
The results of the event-based SWMM considering washed off pollutants are presented in Figures 15 and 16. It was observed that the maximum discharge of the conduits with and without pollutant loads are almost same. In other words, the impact of pollutants load using the event-based method on conduits is negligible. This is because, the event-based method only considers washed off pollutants and rainfall volume in a particular day, and it does not take into account the pollutants already accumulated in the conduits over a period of time. Hence, the impact of pollutants load on the maximum discharge volume of the conduits was found to be negligible.
Figure 15

Change in maximum discharge of conduits for ward 116.

Figure 15

Change in maximum discharge of conduits for ward 116.

Close modal
Figure 16

Change in maximum discharge of conduits for ward 152.

Figure 16

Change in maximum discharge of conduits for ward 152.

Close modal

Estimation of stormwater runoff and conduit efficiency for continuous simulation

Estimation of reduction in the conduit efficiency using continuous simulation method was carried out for two different scenarios, i.e., 2030s (2021–2040) and 2050s (2041–2060) to replicate the real-time situation where the conduits already have accumulated pollutants load.

Ward 116

For both the 2030s and 2050s scenario, the rainfall predicted by the MIROC-MIROC5 data are higher than other GCMs and ICHEC-EC-EARTH has the lowest rainfall volume. The pollutants load estimated for ward 116 in the 2030s (2021–2040) and 2050s (2041–2060) scenarios are presented in Table 12. The trend shows that with an increase in rainfall, there is an increase in the washed off pollutants load from each sub-catchment. So, the total washed off pollutants load from the entire study area has also increased.

Table 12

Pollutant load estimated from the model for 2030s and 2050s for the ward 116

ScenarioTSS (kg)TP (kg)TN (kg)BOD (kg)
Model results for 2030s 
CNRM-CM5 1,183,909 5,231 9,702 937,675 
ICHEC-EC-EARTH 880,780 3,890 7,219 697,801 
MIROC-MIROC5 2,176,706 9,615 17,835 1,723,731 
MPI-M-MPI-ESM-LR 1,263,996 5,584 10,357 1,001,065 
Model results for 2050s 
CNRM-CM5 1,552,323 6,858 12,719 1,229,247 
ICHEC-EC-EARTH 1,051,695 4,645 8,618 833,115 
MIROC-MIROC5 2,204,306 9,737 18,061 1,745,636 
MPI-M-MPI-ESM-LR 1,189,415 5,254 9,747 942,038 
ScenarioTSS (kg)TP (kg)TN (kg)BOD (kg)
Model results for 2030s 
CNRM-CM5 1,183,909 5,231 9,702 937,675 
ICHEC-EC-EARTH 880,780 3,890 7,219 697,801 
MIROC-MIROC5 2,176,706 9,615 17,835 1,723,731 
MPI-M-MPI-ESM-LR 1,263,996 5,584 10,357 1,001,065 
Model results for 2050s 
CNRM-CM5 1,552,323 6,858 12,719 1,229,247 
ICHEC-EC-EARTH 1,051,695 4,645 8,618 833,115 
MIROC-MIROC5 2,204,306 9,737 18,061 1,745,636 
MPI-M-MPI-ESM-LR 1,189,415 5,254 9,747 942,038 

The pollutants transported by each conduit were used to calculate the blockage in each conduit using the mathematical formulas (Equation (4a)–(4c)). The maximum discharge and reduction in the carrying capacity of the conduits for the 2030s scenario is presented in Figure 17. Conduits with noticeable change in maximum discharge conditions are C3 (17 and 16%), C9 (24.5 and 29%), and C13 (25 and 35%) for CNRM-CM5 and MIROC-MIROC5 models. Similarly, for ICHEC-EC-EARTH and MPI-M-MPI-ESM-LR the conduit with maximum change in the discharge are C9 (15%), and C13 (15%), respectively.
Figure 17

Change in maximum discharge of conduits at 2030s in ward 116.

Figure 17

Change in maximum discharge of conduits at 2030s in ward 116.

Close modal
In the 2050s scenario, the maximum change in discharge and the reduction in carrying capacity of the conduits are presented in Figure 18. The conduits with a significant change in the carrying capacity for the CNRM-CM5 and MIROC-MIROC5 model scenarios are C3 (10.61 and 18.9%), C9 (20.90 and 29.76), and C13 (21.66 and 40.03%). Similarly, for ICHEC-EC-EARTH and MPI-M-MPI-ESM-LR are C9 (13.84 and 14.91), and C13 (18.96 and 15.28), respectively. MIROC-MIROC5 simulated the maximum amount of rainfall for which the conduit efficiency reduced up to 40%. The study area's imperviousness is very high, so the majority of the rainfall volume became runoff which led to increase in the washed off pollutants load. As a result of increase in both the runoff volume and washed off pollutants load, maximum blockage occurred in the conduits which significantly reduced its carrying capacity. Best practices of stormwater network management will help to reduce the blockage and improve the conduit efficiency. The questionnaire survey results also reflected the lack of proper maintenance of stormwater network system and the necessity of a stormwater drain network management plan in the study areas.
Figure 18

Change in maximum discharge of conduits at 2050s in ward 116.

Figure 18

Change in maximum discharge of conduits at 2050s in ward 116.

Close modal

Ward 152

The pollutants load estimated for ward 152 in the 2030s (2021–2040) and 2050s (2041–2060) scenario are presented in Table 13. For both the 2030s and 2050s scenario, the maximum washed off pollutants load was observed for the MIROC-MIROC5 model. Similar to ward 116 the washed off pollutants load is directly proportional to the volume of rainfall.

Table 13

Pollutant load estimated from the model for 2030s and 2050s for the ward 152

ScenarioTSS (kg)TP (kg)TN (kg)BOD (kg)
Model results for 2030s 
CNRM-CM5 3,093,002 22,463 25,256 2,339,386 
ICHEC-EC-EARTH 2,279,369 16,562 18,504 1,724,037 
MIROC-MIROC5 5,687,947 41,311 46,371 4,302,806 
MPI-M-MPI-ESM-LR 3,304,837 24,000 26,989 2,499,693 
Model results for 2050s 
CNRM-CM5 4,074,599 29,586 33,296 4,074,599 
ICHEC-EC-EARTH 2,727,995 19,820 22,170 2,063,411 
MIROC-MIROC5 5,753,183 41,786 46,894 4,351,951 
MPI-M-MPI-ESM-LR 3,105,002 22,550 25,350 2,348,399 
ScenarioTSS (kg)TP (kg)TN (kg)BOD (kg)
Model results for 2030s 
CNRM-CM5 3,093,002 22,463 25,256 2,339,386 
ICHEC-EC-EARTH 2,279,369 16,562 18,504 1,724,037 
MIROC-MIROC5 5,687,947 41,311 46,371 4,302,806 
MPI-M-MPI-ESM-LR 3,304,837 24,000 26,989 2,499,693 
Model results for 2050s 
CNRM-CM5 4,074,599 29,586 33,296 4,074,599 
ICHEC-EC-EARTH 2,727,995 19,820 22,170 2,063,411 
MIROC-MIROC5 5,753,183 41,786 46,894 4,351,951 
MPI-M-MPI-ESM-LR 3,105,002 22,550 25,350 2,348,399 

The maximum discharge and the reduction in carrying capacity of the conduits for the 2030s scenario are presented in Figure 19. Conduits with maximum change in discharge are C24 (18.06 and 18.22%), C25 (19.26 and 19.02%), and C50 (13.57 and 13.65%) for CNRM-CM5 and MPI-M-MPI-ESM-LR. For ICHEC-EC-EARTH, C24 (16.26%), and C25 (16.79) showed a maximum change in discharge. For MIROC-MIROC5 data conduits, C22 (16.77%), C23 (20.37%), C24 (33.18%), C25 (34.18%), C45 (20.95), C48 (16.75%), and C50 (24.46%) showed maximum change in discharge. Among all RCMs, the MIROC-MIROC5 scenario had most no. of conduits with a reduction in their carrying capacity and the ICHEC-EC-EARTH scenario had least no. of conduits with a reduction in their carrying capacity.
Figure 19

Change in maximum discharge of conduits at 2030s in ward 152.

Figure 19

Change in maximum discharge of conduits at 2030s in ward 152.

Close modal
Similarly for the 2050s, the maximum change in discharge and reduction in carrying capacity of the conduits are presented in Figure 20. Conduits with noticeable change in carrying capacity for ICHEC-EC-EARTH and MPI-M-MPI-ESM-LR scenarios are C24 (18.04 and 16.55%) C25 (18.67 and 17.32%), and C50 (12.50 and 12.74%), In the CNRM-CM5 scenario conduits C22 (12.11%), C23 (14.69%), C24 (14.69%), C25 (24.91%), C45 (15.16%), and C50 (17.73%) had noticeable reduction in the carrying capacity. As witnessed for the 2030s scenario, the MIROC-MIROC5 scenario had the maximum reduction in the carrying capacity of the conduits which are as follows, C22 (19.37%), C23 (23.39%) C45 (21.64%), C48 (16%), C50 (26.54%), C24 (37.98%), and C25 (39.34%). Among all the GCMs scenarios, the MIROC-MIROC5 scenario had the most no. of conduits with a reduction in their carrying capacity, whereas the MPI-M-MPI-ESM-LR scenario had the least no. of conduits with a reduction in their carrying capacity.
Figure 20

Change in maximum discharge of conduits at 2050s in ward 152.

Figure 20

Change in maximum discharge of conduits at 2050s in ward 152.

Close modal

The summary of reduction in the carrying capacity of the conduits for various scenarios are presented in Table 14.

Table 14

Reduction in carrying capacity of the conduits for various scenarios

ScenarioWard 116 – reduction in carrying capacity of conduits
Ward 152 – reduction in carrying capacity of conduits
2030s
2050s
2030s
2050s
Min value (%)Max value (%)Min value (%)Max value (%)Min value (%)Max value (%)Min value (%)Max value (%)
CNRM-CM5 0.046 25.95 0.052 21.65 0.09 19.26 0.12 24.91 
ICHEC-EC-EARTH 0.024 17.30 0.039 18.96 0.07 16.79 0.08 18.67 
MIROC-MIROC5 0.080 35.28 0.068 40.03 0.17 34.48 0.18 39.34 
MPI-M-MPI-ESM-LR 0.054 16.54 0.051 15.28 0.10 19.02 0.10 17.32 
ScenarioWard 116 – reduction in carrying capacity of conduits
Ward 152 – reduction in carrying capacity of conduits
2030s
2050s
2030s
2050s
Min value (%)Max value (%)Min value (%)Max value (%)Min value (%)Max value (%)Min value (%)Max value (%)
CNRM-CM5 0.046 25.95 0.052 21.65 0.09 19.26 0.12 24.91 
ICHEC-EC-EARTH 0.024 17.30 0.039 18.96 0.07 16.79 0.08 18.67 
MIROC-MIROC5 0.080 35.28 0.068 40.03 0.17 34.48 0.18 39.34 
MPI-M-MPI-ESM-LR 0.054 16.54 0.051 15.28 0.10 19.02 0.10 17.32 

Among all the scenarios the MIROC-MIROC5 had the maximum volume of rainfall and maximum amount of washed off pollutants loads. The high volume of rainfall and pollutants load caused a domino effect, as a result, there was maximum blockage in the conduits and significant reduction in their carrying capacity. As expected, the maximum reduction in carrying capacity of the conduits occurred in the 2050s scenario for most of the climatic models. In some cases, the 2030s scenario had maximum reduction in carrying capacity. This is due to the trend in the rainfall volume forecasted by respective climate models. As discussed earlier, the rainfall volume is directly proportional to the runoff volume and pollutants load and indirectly proportional to the carrying capacity of the conduits. This underlines the importance of using the appropriate climatic model to study the impact of rainfall in the design and management of conduits.

The conclusions and suggestions from this study are as follows:

  • The highly impervious nature of the study area wards has significantly reduced the infiltration rate to about 3–7%. The runoff in all the scenarios was more than 90%, which is mainly because of the low infiltration rate. So, this study suggests that the installation of infiltration galleries and rain water harvesting systems to improve the infiltration rate.

  • Evaluation of conduits based on 100 years return period flood from climate models could provide more insight in managing flood events. Outcomes from such evaluation can be used to propose additional stormwater drain networks wherever necessary or improving the existing drainage network with appropriate slope and size.

  • In this study we have noticed, the single event method cannot be used to study the impact of pollutants load on the conduit efficiency, and it was observed the continuous simulation method is the appropriate way.

  • The highest blockage and maximum reduction in carrying capacity of conduits were about 65 and 40%, respectively. The outcomes from these kinds of studies can be used to devise proper stormwater drain network management plans.

  • This study also observes that for an effective design and management of an urban stormwater drain network, it is important to study the combined effect of rainfall conditions based on climate data, surface imperviousness, and pollutants load.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Agarwal
S.
&
Kumar
S.
2019
Applicability of SWMM for semi urban catchment flood modeling using extreme rainfall events
.
International Journal of Recent Technology and Engineering (IJRTE)
8
(
2
),
245
251
.
Ahamed
S. M. F.
&
Agarwal
S.
2019
Urban flood modelling and management using SWMM for new RR pet region, Vijayawada, India
.
International Journal of Recent Technology and Engineering (IJRTE)
7
(
6C2
),
317
322
.
Andimuthu
R.
,
Kandasamy
P.
,
Mudgal
B. V.
,
Jeganathan
A.
,
Balu
A.
&
Sankar
G.
2019
Performance of urban storm drainage network under changing climate scenarios: flood mitigation in Indian coastal city
.
Scientific Reports
9
,
7783
.
https://doi.org/10.1038/s41598-019-43859-3
.
Athira
T.
&
Agilan
V.
2020
Impact of climate change on rainfall over Chennai
.
IOP Conference Series: Earth and Environmental Science
491
,
012008
.
Bhadran
D.
,
Sunil
A.
,
Nair
S.
,
Mirajkar
A.
&
Sankaran
A.
2020
Ranking of cmip5-based general circulation models using compromise programming and topsis for precipitation: a case study of upper Godavari basin, India
.
International Journal of Big Data Mining for Global Warming
02
,
2050007
.
doi:10.1142/S2630534820500072
.
Bosshard
T.
,
Kotlarski
S.
,
Ewen
T.
&
Schär
C.
2011
Spectral representation of the annual cycle in the climate change signal
.
Hydrology and Earth System Sciences Discussions
8
,
1161
1192
.
http://dx.doi.org/10.5194/hessd-8-1161-2011
.
Brendel
C. E.
,
Dymond
R. L.
&
Aguilar
M. F.
2021
Modeling storm sewer networks and urban flooding in Roanoke, Virginia, with SWMM and GSSHA
.
Journal of Hydrological Engineering
26
(
1
),
1943
5548
.
Choi
J.
,
Park
B.
,
Kim
J.
,
Lee
S.
,
Ryu
J.
,
Kim
K.
&
Kim
Y.
2021
Determination of NPS pollutant unit loads from different land uses
.
Sustainability
13
(
7193
).
https:// doi.org/10.3390/su13137193
.
Fang
G. H.
,
Yang
J.
,
Chen
Y. N.
&
Zammit
C.
2015
Comparing bias correction methods in downscaling meteorological variables for a hydrologic impact study in an arid area in China
.
Hydrology and Earth System Sciences
19
,
2547
2559
.
https://doi.org/10.5194/hess-19-2547-2015
.
Fathy
I.
,
Abdel-Aal
G. M.
,
Fahmy
M. R.
,
Fathy
A.
&
Zeleňáková
M.
2020
The negative impact of blockage on storm water drainage network
.
Water
12
(
7
),
1974
.
https://doi.org/10.3390/w12071974
.
Gupta
K. A.
&
Nair
S. S.
2010
Flood risk and context of land-uses, Chennai city case
.
Journal of Geography and Regional Planning
3
(
12
),
365
372
.
Hussain
N. S.
,
Zwain
M. H.
,
Basim
K.
&
Nile
K. B.
2022
Modeling the effects of land-use and climate change on the performance of stormwater sewer system using SWMM simulation: case study
.
Journal of Water and Climate Change
13
(
1
),
125
138
.
IPCC
2014
Climate Change 2014: Synthesis Report
.
Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change Geneva
,
Switzerland
.
IPCC Climate Change
2013
The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
.
Cambridge University Press, Cambridge
,
UK. New York, USA
.
Janga
S.
,
Chob
M.
,
Yoonc
J.
,
Yoo
y.
,
Kime
S.
,
Kimf
G.
,
Kimg
l.
&
Aksoyh
H.
2007
Using SWMM as a tool for hydrologic impact assessment
.
Desalination
212
,
344
356
.
https://doi.org/10.1016/j.desal.2007.05.005
.
Karmegam
D.
,
Ramamoorthy
S.
&
Mappillairaju
B.
2021
Near real time flood inundation mapping using social media data as an information source: a case study of 2015 Chennai flood
.
Geo Environ Disasters
8
,
25
.
https://doi.org/10.1186/s40677-021-00195-x
.
Komaragiri
S. R.
&
Kumar
D. N.
2014
Ranking of global climate models for India using multicriterion analysis
.
Climate Research
60
,
103
117
.
10.3354/cr01222
.
Kumar
N. K.
,
Thota
S. M.
,
Ashrit
R.
,
Mitra
K. A.
&
Rajeevan
N. M.
2022
Quantile mapping bias correction methods to IMDAA reanalysis for calibrating NCMRWF unified model operational forecasts
.
Hydrological Sciences Journal
67
(
6
),
870
885
.
Leboutillier
D. W.
,
Kells
J. A.
&
Putz
G. J.
2000
Prediction of pollutant load in stormwater runoff from an urban residential area
.
Canadian Water Resources Journal
25
(
4
),
343
359
.
Lee
J. G.
&
Heaney
J. P.
2003
Estimation of urban imperviousness and its impacts on storm water systems
.
Journal of Water Resources Planning Management
129
(
5
),
419
426
.
Li
D.
,
Wan
J.
,
Ma
Y.
,
Wang
Y.
,
Huang
M.
&
Chen
Y.
2015
Stormwater runoff pollutant loading distributions and their correlation with rainfall and catchment characteristics in a rapidly industrialized city
.
PLoS ONE
10
(
3
),
e0118776
.
https://doi.org/10.1371/journal.pone.0118776
.
Mendez
M.
,
Maathuis
B.
,
Hein-Griggs
D.
&
Alvarado-Gamboa
L.-F.
2020
Performance evaluation of bias correction methods for climate change monthly precipitation projections over Costa Rica
.
Water
12
(
2
),
482
.
https://doi.org/10.3390/w12020482
.
Miller
J. D.
&
Hutchins
M.
2017
The impacts of urbanization and climate change on urban flooding and urban water quality: a review of the evidence concerning the United Kingdom
.
Journal of Hydrology: Regional Studies
12
,
345
362
.
Moriasi
N. D.
,
Arnold
J. G.
,
Van Liew
W. M.
,
Michael Bingner
L. N. R.
,
Harmel
R. D.
&
Veith
L. T.
2007
Model evaluation guidelines for systematic quantification of accuracy in watershed simulations
.
Transactions of the ASABE
50
(
3
),
885
900
.
Onen
F.
&
Bagatur
T.
2017
Prediction of flood frequency factor for gumbel distribution using regression and GEP model
.
Arabian Journal for Science and Engineering
42
,
3895
3906
.
Pickett
S. T. A.
,
Cadenasso
M. L.
,
Grove
J. M.
,
Nilon
C. H.
,
Pouyat
R. V.
,
Zipper
W. C.
&
Constanza
R.
2001
Urban ecological systems: linking terrestrial, ecological, physical, and socio-economic components of metropolitan areas
.
Annual Review of Ecology, Evolution, and Systematics
32
,
127
157
.
Ponce
V. M.
2016
Q & A on the return period to be used for design. Available from: http://returnperiod.sdsu.edu
.
Ramanujam
S.
2011
Temple tales of Triplicane. The Times of India. Chennai. The Times Group. Retrieved 9 September 2012
.
Rathnayake
U.
2019
Comparison of statistical methods to graphical methods in rainfall trend analysis: Case studies from Tropical catchments
.
Advances in Meteorology, Hindawi
2019
,
1687
9309
.
Rossman
L. A.
2005
Storm Water Management Model User's Manual Version 5.0
.
U.S. Environmental Protection Agency, Office of Research and Development Cincinnati
,
OH
.
EPA
/600/R-05/040
.
Rossman
L. A.
2010
Storm Water Management Model User's Manual Version 5.0
.
U.S Water Supply and Water Resources Division National Risk Management Research Laboratory
,
Cincinnati, OH
.
45268-EPA/600/R-05/040
.
Ryu
J.
,
Jang
S. N.
,
Kim
J.
,
Jung
Y.
,
Engle
A. B.
&
Lim
J. K.
2016
Development of field pollutant load estimation module and linkage of QUAL2E with watershed-scale L-THIA CAN model
.
Water
8
(
292
),
1
17
.
Schmidli
J.
,
Frei
C.
&
Vidale
P. L.
2006
Downscaling from GC precipitation: a benchmark for dynamical and statistical downscaling methods
.
International Journal of Climatology
26
,
679
689
.
Shahid
S.
,
Wang
X. J.
,
Harun
S. B.
,
Shamsudin
S. B.
,
Ismail
T.
&
Minhans
A.
2016
Climate variability and changes in the major cities of Bangladesh: observations, possible impacts and adaptation
.
Regional Environmental Change
16
(
2
),
459
471
.
Shashikanth
K.
&
Sukumar
P.
2017
Indian monsoon rainfall projections for future using GCM model outputs under climate change, advances in computational sciences and technology
.
Research India Publications
10
(
5
),
1501
1516
.
Sonavane
N.
,
Rangari
V. A.
,
Waikar
M. L.
&
Patil
M.
2020
Urban storm-water modelling using EPA SWMM – a case study of Pune city
. In
IEEE Bangalore Humanitarian Technology Conference (B-HTC)
, pp.
1
5
.
Takele
G. S.
,
Gebre
G. S.
,
Gebremariam
A. G.
&
Engida
A. N.
2021
Future climate change and impacts on water resources in the Upper Blue Nile basin
.
Journal of Water and Climate Change
13
(
2
),
908
925
.
Tu
M. C.
&
Smith
P.
2018
Modeling pollutant build up and washoff parameters for SWMM based on land use in a semiarid urban watershed
.
Water, Air, and Soil Pollutant
229
,
121
.
https://doi.org/10.1007/s11270-018-3777-2
.
Tuomela
C.
,
Sillanpää
N.
&
Koivusalo
H.
2019
Assessment of stormwater pollutant loads and source area contributions with stormwater management model (SWMM)
.
Journal of Environmental Management
233
,
719
727
.
Vishwakarma
A.
,
Choudhary
K. M.
&
Chauhan
S. M.
2022
Assessment of CMIP5 and CORDEX-SA experiments in representing multiscale temperature climatology over central India
.
Water Practice & Technology
1
,
1
13
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).