ABSTRACT
The present study focused on the Kosasthalaiyar River basin in Chennai, Tamil Nadu, India. We analysed the 2015 peak flood and forecasted the feature data in the representative concentration pathways (RCP) 4.5 scenario for various years. We used the scientific data management system (SDSM) software to downscale the Geophysical Fluid Dynamics Laboratory (GFDL) 2.0 general circulation models (GCMs). According to hydrograph analysis, 142.7, 75.31, 461.73, and 248.22 mm runoff can occur in 2030, 2050, 2080, and 2100, respectively. The current study estimates probable peak flows by performing floodplain analysis on the Kosasthalaiyar River sub-basin using the Hydrologic Engineering Centre's Hydrologic Modelling System (HEC-HMS), the Hydrologic Engineering Centre's River Analysis System (HEC-RAS), and geographic information system (GIS) tools. It is possible to observe that the two major peak floods, measuring 581.6 and 110.7 m3/s, respectively, will occur on 28 November 2030 at 10:20 a.m. and 12 December 2050 at 9:20 a.m. Additionally, high floods of 997 and 1,438.4 m3/s can be recorded on 20 December 2080 at 9:50 a.m. and 29 November 2100 at 9:40 a.m., respectively.
HIGHLIGHTS
Significant peak flow timing differences are projected for the years 2030, 2050, 2080, and 2100.
2030 peak flow: 581.6 m3/s on 28 November 2050 and 110.7 m3/s on 12 December.
2080 peak flow: 997 m3/s on 20 December 2100 and 1,438.4 m3/s on 29 November.
Peak flow timing shifts to late November and early December.
Sholavaram, Tamil Nadu faces infrastructure and agriculture impacts; urbanisation worsens flood risks, advocating rainwater harvesting and afforestation.
INTRODUCTION
In today's fast-moving world, urbanisation is one of the global phenomena driven by factors such as industrialisation, economic opportunities, and social changes (Castro & Maidment 2020). It typically includes city growth, infrastructure expansion, and changes in land-use patterns (Srinivas et al. 2018a, b). Urbanisation, as one of the development factors, brings various benefits, but it also presents several changes and drawbacks, such as environmental degradation, public health challenges, and the loss of agricultural land (Khattak et al. 2016). One of the major consequences is drastic droughts and floods (Verma et al. 2022). Further, droughts are prolonged periods of abnormally low precipitation resulting from water shortages that can have significant social, economic, and environmental impacts (Mishra & Singh 2010; Miyan 2015). In addition, human activities such as deforestation and over-extraction of water resources primarily cause droughts (Miyan 2015). According to Şen (2018), floods are natural disasters characterised by the overflow of runoff water onto normally dry land. Ologunorisa et al. (2022) commented that, due to urbanisation, precipitated water is unable to infiltrate into the sub-surface of the earth, leading to water stagnation on the surface. Further, overtime runoff is increasing, leading to severe floods (Viglione & Rogger 2015; Şen 2018). Jain et al. (2007) and Prakash et al. (2023) commented that urbanised areas are well-developed locations, such as Mumbai, Bengaluru, Chennai, and so on; drastic floods may cause a huge amount of property loss. However, we cannot stop urbanisation as it is one of the economic factors contributing to a country's gross domestic product (GDP), but one can analyse the amount of flooding that will occur and do floodplain mapping, through which we can plot flood-prone areas of the particular basin (Chen et al. 2021). In addition, urban stormwater systems are crucial for transporting rainfall from these areas to lakes or rivers, which are the outlets for the water (Muis et al. 2015; Mazzoleni et al. 2021). However, managing stormwater in urban areas has become a big challenge since there is less space and land values are growing. Unfortunately, most stormwater drains in metropolitan areas are unable to handle excess runoff during high rainfall events due to inadequate maintenance and uncontrolled urbanisation (Prakash et al. 2023). These brief downpour events become more frequent and intense, and climate change alters the amount and peak of streamflow (Huntington 2010). Mujumdar et al. (2020) recommended that runoff measurement can solve numerous problems with watershed management. In conjunction with Mujumdar et al. (2020), Patz et al. (2014) commented that the primary cause of major and frequent flooding in a watershed is an excess of runoff volumes directed into the channels compared with the stream's carrying capacity.
Furthermore, as urbanisation and climate change raise the runoff peak, it is critical to evaluate the flood magnitude and intensity for flood control (Hettiarachchi et al. 2018). Climate change can significantly impact variations in peak flow (Patz et al. 2014; Guan et al. 2016; Hettiarachchi et al. 2018; Chen et al. 2021; Pasupuleti et al. 2023; Loganathan & Sathiyamoorthy 2024; Rajesh 2024). With early flood warnings, anticipating peak flows in urban areas can enhance flood preparedness. Kuller et al. (2021) conducted flow measurements with various instruments and a variety of sophisticated and conventional methods, particularly in flood areas. They found that the stage–discharge relationship, often known as a rating curve, is one of the oldest methods used by hydrographers worldwide (Stephens & Cloke 2014). However, a stream channel's floodplain and hydraulic characteristics are the primary determinants of its rating curve, which varies periodically at a specific stretch of the stream (Zhou et al. 2019). We use hydrologic modelling for a specific rainfall depth in the current study to assess how watersheds respond hydrologically, as Guan et al. (2016) used in the past. In addition, many studies used hydrological modelling to look at a wide range of topics, such as managing resources, figuring out the impacts of climate change and urbanisation, predicting and guessing the size of floods, and figuring out streamflow in an unmeasured basin (Huntington 2010; Patz et al. 2014; Demaria et al. 2016; Gangolu et al. 2023; Sharma et al. 2023; Al Khazaleh et al. 2024; Nandhini et al. 2024). In the current study, we simulated the HEC-HMS model for the 2008 flood event, collecting the no-service flood hydrograph at the basin's outlet from the Public Works Department of the Government of Tamil Nadu (Srinivas et al. 2018a, b). They demonstrate how to prepare using LIDAR digital elevation model (DEM)-based HEC-RAS and HEC-HMS models (Meresa 2019; Al-Areeq et al. 2023). The study indicated the usefulness of HEC-RAS software along with the Global Flood Monitoring System (GFMS) tool to identify the flood extension zone and return period of flood in the study area (Khattak et al. 2016). Kaya et al. (2022) used the HEC-RAS 5.0.7 open-source model and GFMS datasets to model floods and draw lines around different risk zones. However, research organisations, government agencies, and private agencies selected the HEC-RAS model due to its availability and widespread use. In addition, Malik & Pal (2020) used log-normal, Log-Pearson Type III (LPT-3), Gumbel's extreme value distribution (EV-I), and extreme value distribution-III (EV-III) models, as well as HEC-RAS software, to look at past and future flood frequencies and map out flood-prone areas. Subsequently, they collected historical gauge height data from the Dwarkeswar River near Arambag Town, Hooghly District, West Bengal, from 1978 to 2018 and computed the annual peak discharge for the same period using a rating curve. Derdour et al. (2018) developed an approach that utilised hydrological and hydraulic models, in conjunction with GIS techniques (Derdour et al. 2018), to examine the flooding patterns of the city of Ain Sefra during major events. This integrated methodology allows for a detailed assessment of flood risk, peak discharges, and floodplain inundation patterns, providing valuable insights for flood hazard assessment and urban planning in arid regions like Ain Sefra. Furthermore, Natarajan & Radhakrishnan (2020) noted that they demarcated the flood as a surplus runoff from a watercourse or river capable of submerging land areas. The medium-sized dendritic, ungauged urban catchment in Tiruchirappalli adopted an integrated modelling technique. HEC-HMS achieved a rainfall-runoff model, which is mandatory for flood risk assessment studies (Mandal & Chakrabarty 2016; Abdessamed & Abderrazak 2019; Guduru et al. 2023; Titus et al. 2024). In recent studies, Pradhan et al. (2022) studied the HEC-RAS model to find two flow variables, namely, water depth and flow velocity. They presented the flood inundation maps, which show that the depths of the 150-year return period are 14.01 m, respectively. Verma & Mirajkar (2024) combined HEC-RAS and GIS with the HEC-Geo-RAS companion for hydraulics modelling and flood inundation mapping. For the past few years, the application of remote sensing and GIS in terrain modelling and floodplain mapping has been relatively limited. Arc Map generated floodplain maps by importing HEC-RAS outputs and using the HEC Geo-RAS (Khattak et al. 2016). The aforementioned studies motivated the authors to study highly flood-prone regions, with Tamil Nadu being one of the most affected by North East monsoons annually (Srinivas et al. 2018a, b).
The study focusses on the Kosasthalaiyar River basin in Chennai, Tamil Nadu, India, and uses various tools such as the HEC-HMS, HEC-RAS, and GIS tools to perform floodplain analysis and forecast peak flood events for different years (2030, 2050, 2080, and 2100). These forecasts are significant for urban planning, flood mitigation, and climate adaptation, providing specific peak flow predictions that help in preparing for future extreme flooding scenarios.
The limited availability of a comprehensive study on floodplain analysis using HEC-HMS, HEC-RAS, and GIS on the Kosasthalaiyar river sub-basin motivated the authors to do the current study with the following objectives:
(a) to analyse floodplain characteristics along the Kosasthalaiyar River using HEC-HMS to understand precipitation-runoff processes and
(b) to simulate hydraulic behaviour and flood inundation using HEC-RAS to provide an accurate floodplain.
STUDY AREA
Data collection
We got information for this study area from a number of different places, such as India-WRIS rainfall data in csv format and the 30 m × 30 m resolution DEM map from the Shuttle Radar Topography Mission (SRTM) in dot-tif format (Goyal et al. 2018; Chakraborty 2019; Nagaveni et al. 2019; Kant et al. 2023). We obtained Landsat-8 maps for land use and land cover (LULC). We used the RCP 4.5 scenario to forecast the feature data for various years (Asadollah et al. 2022; Jahangir et al. 2022; Rahaman 2022; Zhang et al. 2022). We use the SDSM software to downscale the GFDL 2.0 GCM (Ahmed et al. 2013; Raju & Kumar 2014; Shakeri et al. 2021; Asadollah et al. 2022). Data gathered from various sources comes in two forms: spatial data, such as SRTM DEM, and non-spatial data, including details of the drainage network and rainfall data (Nagaveni et al. 2019; Verma & Mirajkar 2024). We use Arc Map 10.4 to convert the collected data into a GIS-compatible format. Obtained shape files of sub-catchments and drainage networks from Google Maps to construct a comprehensive watershed database. We extract elevation data from the SRTM DEM (Goyal et al. 2018), and we extract one of the major storm events (November–December 2015) from the India Meteorological Department (IMD) rainfall data used to run the model simulations (Singh et al. 2018). In November 2015, Chennai received 1,049.3 mm of rainfall, the highest amount since 1918. On 6 November 2015, Chennai received 246.5 mm of rainfall in a single day, the highest in the last 10 years for November at the time. On 1 December 2015, the city received 494 mm of rainfall in a 24-h period, causing severe flooding that killed more than 400 people and submerged lakhs of homes (Table 1).
S. No . | Data type . | Data source . | Sources . |
---|---|---|---|
1 | DEM | STRM | https://earthexplorer.usgs.gov/ |
2 | LULC | Landsat 8 | https://earthexplorer.usgs.gov/ |
3 | Precipitation data | IMD | https://imdpune.gov.in/cmpg/Griddata/Rainfall_25_NetCDF.html |
4 | Future precipitation data | CMIP 5 (GFDL 2.0) | https://esgf-node.ipsl.upmc.fr/projects/cmip5-ipsl/ |
5 | Hydrological data | CWC | https://indiawris.gov.in/wris/ |
S. No . | Data type . | Data source . | Sources . |
---|---|---|---|
1 | DEM | STRM | https://earthexplorer.usgs.gov/ |
2 | LULC | Landsat 8 | https://earthexplorer.usgs.gov/ |
3 | Precipitation data | IMD | https://imdpune.gov.in/cmpg/Griddata/Rainfall_25_NetCDF.html |
4 | Future precipitation data | CMIP 5 (GFDL 2.0) | https://esgf-node.ipsl.upmc.fr/projects/cmip5-ipsl/ |
5 | Hydrological data | CWC | https://indiawris.gov.in/wris/ |
METHODOLOGY
Digital elevation model
Land use and land cover
Stream networks
Arc Map commonly uses the Strahler stream order system, which merges two streams of the same order to form a stream one order higher. The process persists until it reaches the highest-order stream. Arc Map allows users to determine the flow direction and accumulation for each cell in a raster grid. This information is critical for delineating stream networks and analysing water flow patterns. Flow accumulation represents the number of cells that contribute flow to each cell in the raster grid, which helps identify stream channels. Arc Map provides tools for conducting network analysis on stream networks. This includes calculating distances along stream channels, identifying upstream and downstream locations, determining optimal routes for navigation or resource extraction, and performing network-based spatial analysis. Overall, stream networks in Arc Map serve as essential datasets for understanding and managing water resources, supporting a wide range of applications in environmental science, engineering, and spatial analysis. To make a flood model for the study area, we put together GIS, the HEC-HMS rainfall-runoff model, and the HEC-RAS hydraulic model, as shown in Figure 5 (Olayinka & Irivbogbe 2017; Meresa 2019; Nagarajan et al. 2022). The process of developing flood hazard cross-sections involves understanding model creation, input and output data, and interconnections. To prepare the input parameter, the dataset establishes a geodatabase in Arc Map (Kiran et al. 2022; Pradhan et al. 2022; Kant et al. 2023; Verma & Mirajkar 2024). We import and reproject the base data to the projected coordinate system WGS 1984/UTM zone 44 N datum before clipping it to the study region. We prepare the input dataset and then use HEC-HMS to create a rainfall-runoff model (Olayinka & Irivbogbe 2017). We use the results of the rainfall-runoff simulation to run the 1D flow routing model and create cross-sections for the specific strip (Kiran et al. 2022). We selected the 2015 flood, a major storm event in the Kosasthalaiyar basin's history, for model simulation. We use the RCP 4.5 scenario in SDSM software to downscale the GCM GFDL 2.0 data, obtaining future precipitation data for 2030, 2050, 2080, and 2100 years. We process the precipitation data using the same HEC-HMS rainfall-runoff model and the HEC-RAS hydraulic model.
RESULTS AND DISCUSSION
Rainfall-runoff model: HEC-HMS
Sub-basin . | Impervious (%) . | Sub-basin . | Impervious (%) . |
---|---|---|---|
Sub-basin 1 | 5 | Sub-basin 22 | 5 |
Sub-basin 2 | 5 | Sub-basin 23 | 5 |
Sub-basin 3 | 5 | Sub-basin 24 | 30 |
Sub-basin 4 | 5 | Sub-basin 25 | 5 |
Sub-basin 5 | 15 | Sub-basin 27 | 5 |
Sub-basin 7 | 5 | Sub-basin 28 | 5 |
Sub-basin 8 | 5 | Sub-basin 29 | 30 |
Sub-basin 9 | 5 | Sub-basin 30 | 5 |
Sub-basin 10 | 65 | Sub-basin 31 | 5 |
Sub-basin 11 | 5 | Sub-basin 32 | 5 |
Sub-basin 12 | 30 | Sub-basin 33 | 5 |
Sub-basin 12 | 15 | Sub-basin 35 | 30 |
Sub-basin 14 | 5 | Sub-basin 37 | 15 |
Sub-basin 15 | 15 | Sub-basin 38 | 100 |
Sub-basin 16 | 5 | Sub-basin 39 | 5 |
Sub-basin 17 | 5 | Sub-basin 42 | 30 |
Sub-basin 18 | 5 | ||
Sub-basin 19 | 30 | ||
Sub-basin 20 | 30 |
Sub-basin . | Impervious (%) . | Sub-basin . | Impervious (%) . |
---|---|---|---|
Sub-basin 1 | 5 | Sub-basin 22 | 5 |
Sub-basin 2 | 5 | Sub-basin 23 | 5 |
Sub-basin 3 | 5 | Sub-basin 24 | 30 |
Sub-basin 4 | 5 | Sub-basin 25 | 5 |
Sub-basin 5 | 15 | Sub-basin 27 | 5 |
Sub-basin 7 | 5 | Sub-basin 28 | 5 |
Sub-basin 8 | 5 | Sub-basin 29 | 30 |
Sub-basin 9 | 5 | Sub-basin 30 | 5 |
Sub-basin 10 | 65 | Sub-basin 31 | 5 |
Sub-basin 11 | 5 | Sub-basin 32 | 5 |
Sub-basin 12 | 30 | Sub-basin 33 | 5 |
Sub-basin 12 | 15 | Sub-basin 35 | 30 |
Sub-basin 14 | 5 | Sub-basin 37 | 15 |
Sub-basin 15 | 15 | Sub-basin 38 | 100 |
Sub-basin 16 | 5 | Sub-basin 39 | 5 |
Sub-basin 17 | 5 | Sub-basin 42 | 30 |
Sub-basin 18 | 5 | ||
Sub-basin 19 | 30 | ||
Sub-basin 20 | 30 |
The model was calibrated by hand at Sholavaram. In November 2015, performance metrics like Nash–Sutcliffe efficiency (NSE) and r2 were used to evaluate the calibration. Parameters like curve number (CN) and time of concentration (Tc) were changed until the simulated results were very close to the observed data. In calibration, NSe and r2 values are 0.81 and 0.78, respectively. After calibration, we validated the model for December 2015 by comparing the simulated results with the observed data for validation (NSE), resulting in r2 values of 0.77 and 0.73, respectively.
Hydrographs from the rainfall-runoff model are saved as time series data and directly input into the hydraulic model we developed using HEC-RAS. We used GFDL 2.0, a general circulation model (GCM), to predict future climate data under the RCP 4.5 scenario for the years 2030, 2050, 2080, and 2100, as illustrated in Figure 7(b)–7(e), respectively. On November 28, in the year 2030, the maximum precipitation observed was 45.5 mm. Two days, the 11th and 23rd of December, saw the maximum precipitation in 2050. We observed the maximum precipitation on November 20 and November 29, respectively, for 2080 and 2100. We will collect this data for November and December using the statistical downscaling model. We used the statistical downscaling model to reduce the coarse-resolution outputs of the GFDL 2.0 model to a finer spatial resolution suitable for regional analysis of India's southeast catchments. This downscaling process captures the detailed climate variations and trends necessary for accurate hydrological modelling. Further, Figure 7(b)–7(e) demonstrates the projected changes in precipitation during the critical months of November and December for the years 2030, 2050, 2080, and 2100, which are pivotal for understanding the hydrological responses and flood risks in the region. These projections are integral to the HEC-HMS model simulations, providing essential inputs for predicting future flood events and aiding in the development of robust flood management and mitigation strategies. For accurate hydrological modelling in the HEC-HMS setup, the basin parameters for each sub-basin (Table 3) are very important. These include CN, longest flow path length, basin slope, potential maximum retention, and time of concentration. The CN denotes the soil's infiltration characteristics, which directly influence runoff generation. The higher value for sub-basin 17 is 97.5; on the other hand, the lower value for sub-basin 12 is 55.5. It depicts that lower numbers result in low runoff, while higher numbers indicate rising runoff potential. The longest flow path length determines the distance water travels within the sub-basin; for sub-basin 7, the maximum distance is 57.08 km, and for sub-basin 15, the least distance is 8.71 km. This will influence the time it takes for runoff to reach the outlet. We observed a maximum slope of 5.531% for sub-basin 11, and a minimum slope of 2.776% for sub-basin 20. The steeper slopes generally lead to faster runoff. Potential maximum retention (S): sub-basin 13 records a higher value of 8.05 inches, while sub-basin 17 records a lower value of 0.25 inches. This value indicates the maximum water retention capacity of the basin before generating runoff, and it inversely correlates with the CN. Tc is the time required for water to travel from the most distant point in the sub-basin to the outlet, impacting the timing and peak of the flood hydrograph. For sub-basin 7, the highest Tc is 10.928 h, and for sub-basin 11, the lowest Tc is 1.686 h. We integrate these parameters, derived from topographic and soil data, into the HEC-HMS model to accurately simulate hydrological processes and predict flood events under various climate scenarios.
Kosasthalaiyar River sub-basin parameters . | ||||||
---|---|---|---|---|---|---|
Sub-basin . | CN . | Longest flow path length (km) . | Basin slope (Y) . | Potential maximum retention (S) = (1,000/CN) − 10 (in) . | Basin slope (Y%) . | Time of concentration (h) . |
Sub-basin 1 | 74.0 | 22.47 | 0.073 | 3.52 | 7.3 | 4.388 |
Sub-basin 2 | 94.8 | 26.11 | 0.045 | 0.54 | 4.478 | 2.978 |
Sub-basin 3 | 98.3 | 20.25 | 0.043 | 0.18 | 4.325 | 2.046 |
Sub-basin 4 | 78.4 | 12.47 | 0.029 | 2.75 | 2.852 | 3.846 |
Sub-basin 5 | 79.6 | 10.53 | 0.034 | 2.56 | 3.369 | 2.978 |
Sub-basin 7 | 79.1 | 57.08 | 0.039 | 2.64 | 3.856 | 10.928 |
Sub-basin 8 | 78.4 | 13.80 | 0.033 | 2.76 | 3.335 | 3.862 |
Sub-basin 9 | 72.5 | 19.45 | 0.032 | 3.79 | 3.184 | 6.169 |
Sub-basin 10 | 73.7 | 28.62 | 0.034 | 3.58 | 3.414 | 7.854 |
Sub-basin 11 | 96.2 | 16.01 | 0.055 | 0.39 | 5.531 | 1.686 |
Sub-basin 12 | 72.0 | 11.85 | 0.032 | 3.88 | 3.233 | 4.171 |
Sub-basin 13 | 55.4 | 11.51 | 0.032 | 8.05 | 3.189 | 6.322 |
Sub-basin 14 | 85.6 | 18.26 | 0.031 | 1.68 | 3.146 | 3.927 |
Sub-basin 15 | 60.3 | 8.71 | 0.032 | 6.58 | 3.2 | 4.460 |
Sub-basin 16 | 69.4 | 28.04 | 0.030 | 4.42 | 3.041 | 9.212 |
Sub-basin 17 | 97.5 | 17.81 | 0.045 | 0.25 | 4.523 | 1.884 |
Sub-basin 18 | 85.5 | 33.84 | 0.046 | 1.70 | 4.63 | 5.333 |
Sub-basin 19 | 65.3 | 11.95 | 0.032 | 5.32 | 3.221 | 5.044 |
Sub-basin 20 | 60.7 | 15.14 | 0.028 | 6.48 | 2.776 | 7.380 |
Sub-basin 22 | 80.6 | 14.18 | 0.037 | 2.41 | 3.72 | 3.491 |
Sub-basin 23 | 82.7 | 18.08 | 0.042 | 2.10 | 4.227 | 3.717 |
Sub-basin 24 | 77.6 | 21.06 | 0.032 | 2.89 | 3.159 | 5.704 |
Sub-basin 25 | 79.3 | 15.89 | 0.051 | 2.61 | 5.111 | 3.396 |
Sub-basin 27 | 93.1 | 20.36 | 0.036 | 0.74 | 3.566 | 2.976 |
Sub-basin 28 | 76.1 | 16.29 | 0.034 | 3.13 | 3.372 | 4.689 |
Sub-basin 29 | 72.1 | 18.64 | 0.036 | 3.86 | 3.552 | 5.699 |
Sub-basin 30 | 80.7 | 15.29 | 0.051 | 2.40 | 5.088 | 3.163 |
Sub-basin 31 | 66.6 | 16.04 | 0.032 | 5.02 | 3.201 | 6.181 |
Sub-basin 32 | 85.2 | 36.94 | 0.035 | 1.74 | 3.533 | 6.616 |
Sub-basin 33 | 78.5 | 19.43 | 0.037 | 2.74 | 3.658 | 4.829 |
Sub-basin 35 | 71.6 | 21.24 | 0.030 | 3.97 | 3.027 | 6.964 |
Sub-basin 37 | 89.2 | 27.42 | 0.041 | 1.21 | 4.085 | 4.170 |
Sub-basin 38 | 56.5 | 7.60 | 0.042 | 7.71 | 4.245 | 3.826 |
Sub-basin 39 | 66.2 | 20.87 | 0.031 | 5.11 | 3.081 | 7.863 |
Sub-basin 42 | 67.0 | 21.88 | 0.030 | 4.93 | 3.012 | 8.091 |
Kosasthalaiyar River sub-basin parameters . | ||||||
---|---|---|---|---|---|---|
Sub-basin . | CN . | Longest flow path length (km) . | Basin slope (Y) . | Potential maximum retention (S) = (1,000/CN) − 10 (in) . | Basin slope (Y%) . | Time of concentration (h) . |
Sub-basin 1 | 74.0 | 22.47 | 0.073 | 3.52 | 7.3 | 4.388 |
Sub-basin 2 | 94.8 | 26.11 | 0.045 | 0.54 | 4.478 | 2.978 |
Sub-basin 3 | 98.3 | 20.25 | 0.043 | 0.18 | 4.325 | 2.046 |
Sub-basin 4 | 78.4 | 12.47 | 0.029 | 2.75 | 2.852 | 3.846 |
Sub-basin 5 | 79.6 | 10.53 | 0.034 | 2.56 | 3.369 | 2.978 |
Sub-basin 7 | 79.1 | 57.08 | 0.039 | 2.64 | 3.856 | 10.928 |
Sub-basin 8 | 78.4 | 13.80 | 0.033 | 2.76 | 3.335 | 3.862 |
Sub-basin 9 | 72.5 | 19.45 | 0.032 | 3.79 | 3.184 | 6.169 |
Sub-basin 10 | 73.7 | 28.62 | 0.034 | 3.58 | 3.414 | 7.854 |
Sub-basin 11 | 96.2 | 16.01 | 0.055 | 0.39 | 5.531 | 1.686 |
Sub-basin 12 | 72.0 | 11.85 | 0.032 | 3.88 | 3.233 | 4.171 |
Sub-basin 13 | 55.4 | 11.51 | 0.032 | 8.05 | 3.189 | 6.322 |
Sub-basin 14 | 85.6 | 18.26 | 0.031 | 1.68 | 3.146 | 3.927 |
Sub-basin 15 | 60.3 | 8.71 | 0.032 | 6.58 | 3.2 | 4.460 |
Sub-basin 16 | 69.4 | 28.04 | 0.030 | 4.42 | 3.041 | 9.212 |
Sub-basin 17 | 97.5 | 17.81 | 0.045 | 0.25 | 4.523 | 1.884 |
Sub-basin 18 | 85.5 | 33.84 | 0.046 | 1.70 | 4.63 | 5.333 |
Sub-basin 19 | 65.3 | 11.95 | 0.032 | 5.32 | 3.221 | 5.044 |
Sub-basin 20 | 60.7 | 15.14 | 0.028 | 6.48 | 2.776 | 7.380 |
Sub-basin 22 | 80.6 | 14.18 | 0.037 | 2.41 | 3.72 | 3.491 |
Sub-basin 23 | 82.7 | 18.08 | 0.042 | 2.10 | 4.227 | 3.717 |
Sub-basin 24 | 77.6 | 21.06 | 0.032 | 2.89 | 3.159 | 5.704 |
Sub-basin 25 | 79.3 | 15.89 | 0.051 | 2.61 | 5.111 | 3.396 |
Sub-basin 27 | 93.1 | 20.36 | 0.036 | 0.74 | 3.566 | 2.976 |
Sub-basin 28 | 76.1 | 16.29 | 0.034 | 3.13 | 3.372 | 4.689 |
Sub-basin 29 | 72.1 | 18.64 | 0.036 | 3.86 | 3.552 | 5.699 |
Sub-basin 30 | 80.7 | 15.29 | 0.051 | 2.40 | 5.088 | 3.163 |
Sub-basin 31 | 66.6 | 16.04 | 0.032 | 5.02 | 3.201 | 6.181 |
Sub-basin 32 | 85.2 | 36.94 | 0.035 | 1.74 | 3.533 | 6.616 |
Sub-basin 33 | 78.5 | 19.43 | 0.037 | 2.74 | 3.658 | 4.829 |
Sub-basin 35 | 71.6 | 21.24 | 0.030 | 3.97 | 3.027 | 6.964 |
Sub-basin 37 | 89.2 | 27.42 | 0.041 | 1.21 | 4.085 | 4.170 |
Sub-basin 38 | 56.5 | 7.60 | 0.042 | 7.71 | 4.245 | 3.826 |
Sub-basin 39 | 66.2 | 20.87 | 0.031 | 5.11 | 3.081 | 7.863 |
Sub-basin 42 | 67.0 | 21.88 | 0.030 | 4.93 | 3.012 | 8.091 |
Hydraulic model: HEC-RAS
We calibrated the model by adjusting hydraulic parameter (Manning's n) to match simulated water levels with observed values collected from CWC (I-WRIS). Calibrate Manning's n using flow data from a specific storm in November 2015, and then validate the model using a different event in December 2015. We determined the average Manning's roughness coefficient to be 0.04524. We provide the peak discharge as input for HEC-RAS for the years 2030, 2050, 2080, and 2100. The observed cross-section is in Sholavaram, Tamil Nadu. The peak discharge of 2030 was 581.6 m3/s on 28 November 2030, at 10:20 a.m., the channel cross-section is shown in Supplementary Figure S1. On December 12 at 9:20 a.m., we will observe the peak discharge of 2050 at 10.7 m3/s and illustrate the channel's cross-section in Supplementary Figure S2. Similarly, Supplementary Figures S3 and S4 show that the peak discharges of 2080 and 2100 are 997 and 1,438.4 m3/s, on December 20 at 9:50 a.m. and November 29 at 9:40 a.m. respectively. We observe in the channel's cross-section that the water level in 2050 is lower than in previous years due to changes in precipitation patterns. These results underscore the critical role of HEC-RAS in understanding the hydraulic impacts of climate change and facilitating the development of effective flood management and mitigation strategies to safeguard communities and infrastructure against future flood risks. Further, in comparison with the present study, wherein, the use of HEC-HMS for modelling rainfall-runoff and HEC-RAS for modelling hydraulics, agrees with what other researchers have found about urban floods. For instance, Kamath et al. (2011) used the SCS unit hydrograph method for rainfall-runoff modelling, concluding that its simplicity and adaptability to diverse conditions make it ideal for flood prediction. This corresponds with the current study's use of the same process to calculate lag times and runoff patterns. Similarly, Olayinka & Irivbogbe (2017) highlighted the effectiveness of integrating GIS with HEC-HMS and HEC-RAS for detailed spatial analysis, particularly in urban flood modelling. Also, their research showed how important high-resolution DEMs and LULC data are for figuring out flow patterns and flood risks. This shows that adding GIS to models makes them much more accurate. Future climate scenarios, such as RCP 4.5, are critical components of flood prediction in climate change studies. In addition, Zhang et al. (2022) used future precipitation data to simulate flood events, revealing rising flood risks in urban areas under changing climate conditions. Our study's findings on the projected changes in peak flow for 2030, 2050, 2080, and 2100 echo these concerns, as the results indicate a shift in flood peak occurrences during November and December due to altered precipitation patterns. This is consistent with Nagarajan et al. (2022), who reported that changes in monsoon dynamics and urbanisation escalate flood hazards in rapidly developing regions, further validating the present study's conclusion on the influence of urbanisation and climate change. The comparison of percentage imperviousness across sub-basins in the Kosasthalaiyar River basin also finds parallels in other studies. Meresa (2019) and Rahaman (2022) both noted the significant role of impervious surfaces in increasing runoff and decreasing infiltration rates in urban basins. Our study, which showed higher imperviousness in sub-basins with increased urbanisation (such as sub-basin 10 with 65% imperviousness), aligns with their findings that urban development amplifies flood potential. These comparisons reinforce the broader consensus that urbanisation, when combined with changing climate patterns, requires advanced flood risk management and planning using reliable hydrological and hydraulic models.
CONCLUSIONS
Urban flooding is becoming increasingly frequent due to urbanisation and climate change. During monsoon seasons, drainage systems often surpass their capacity, leading to overland flooding. This study focused on the Kosasthalaiyar River basin in Chennai, Tamil Nadu, analysing peak floods from 2015 and forecasting future flood scenarios using climate models. The results indicate significant peak floods for 2030, 2050, 2080, and 2100, with projected flows increasing over time. Rainfall-runoff modelling and floodplain analysis highlight the growing flood risks, emphasising the need for better flood management in urban areas.
The present study conducts floodplain analysis on the Kosasthalaiyar River sub-basin using HEC-HMS, HEC-RAS, and Arc Map to estimate probable peak flows for the years 2030, 2050, 2080, and 2100. From the ongoing study, the following conclusions are drawn:
(a) From the SCS method, the significant peak floods, namely, 581.6 and 110.7 m3/s, can be observed on 28 November 2030 and 12 December 2050, respectively.
(b) Further, on 20 December 2080 and 29 November 2100, peak floods of 997 and 1,438.4 m3/s were observed, respectively.
(c) These peak floods are particularly visible in Sholavaram, Tamil Nadu, emphasising areas that may be more vulnerable in future flood scenarios.
(d) These floods may have a heavy impact on infrastructure, agriculture, and human lives. Apart from climate change, urbanisation significantly contributes to severe floods in urbanised areas by reducing the land's infiltration capacity and increasing the flow of precipitated water on the surface.
As urbanisation is one of the developing aspects of a country's growth, it provides various benefits but also has several drawbacks. Therefore, we need to follow various techniques to mitigate the flood risk, like roof-top rainwater harvesting and afforestation. The current study recommends the following actions to lessen the risk of flooding, particularly in metropolitan areas susceptible to urbanisation and climate change:
(a) Upgrade and expand drainage infrastructure: This will help to keep rainwater and sewage systems apart and lessen overflow during periods of intense precipitation. To avoid clogs, make sure your drains are cleaned and maintained on a regular basis.
(b) Green infrastructure: To improve water infiltration, lessen runoff, and ease the burden on drainage systems, install green infrastructure such as rain gardens, permeable pavements, green roofs, and urban wetlands.
(c) Restoration of floodplains: Preserve and restore natural wetlands and floodplains, which serve as buffers by soaking up extra water during periods of high precipitation, lowering the chance of flooding downstream.
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.
CONFLICT OF INTEREST
The authors declare there is no conflict.