Simulation-based modeling of urban waterlogging in Khulna City

Khulna City is extremely vulnerable to the effects of climate change. The city experiences frequent waterlogging during extreme rainfall events. This research prepared a drainage simulation model considering climate change issues for investigating the extent of waterlogging in the city. Watershed and precipitation were analyzed to examine the existing scenario of the study area. Finally, a Mike Urban-based hydrological simulation model was formulated to investigate the severity of waterlogging. As found by precipitation analysis, 180 mm, 346 mm, and 396 mm rainfall might occur for the 5-year, 50-year, and 100-year return period, respectively. This research identified that volume of overland flow might be affected by climate change-induced rainfall. According to the simulation, 62.52% of the study area was waterlogged with different inundation depths. It was found that traffic movements were severely disrupted and structures were hampered due to waterlogging. 45% paved, 77% brick soling, and 65% unpaved roads were inundated by different inundation levels. On the other hand, 61.1% of structures of the study area were affected by waterlogging. The findings of the research might help the concerned authority in decision-making, especially for the drainage and water-related issues, to solve the waterlogging. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). doi: 10.2166/wcc.2020.256 ://iwaponline.com/jwcc/article-pdf/doi/10.2166/wcc.2020.256/688523/jwc2020256.pdf Showmitra Kumar Sarkar (corresponding author) Md. Esraz-Ul-Zannat Department of Urban and Regional Planning, Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh E-mail: mail4dhrubo@gmail.com Md. Atikur Rahman Institute of Water Modelling (IWM), Dhaka, Bangladesh Md. Feroz Islam Faculty of Geosciences, Utrecht University, Utrecht, The Netherlands


INTRODUCTION
Bangladesh (BD) is considered the world's most climate vulnerable country due to its setting as a vast low-lying land (Karim & Mimura ; The Asian Development Bank ; The World Bank ; USAID ). The adverse impacts of climate change are expected to be severe in BD and may lead to substantial economic loss (Mirza ). The impacts of waterlogging result in traffic congestion and damage to urban infrastructures (i.e., buildings, roads and walkways, underground service lines, etc.) (Wu et al. ).
Khulna City is one of the major coastal cities in Bangladesh and extremely vulnerable to climate change due to the influence of the tides of the Bay of Bengal (Ahmed & Ghosh ). It is also the third largest city as well as the industrial Khulna City's rain water used to drain out through the canals, and sometimes ponds were used to function as retention catchments for additional storm water. However, in recent years, waterlogging has become a severe problem for the city, especially during late monsoon, because of the unabated encroachment of perennial waterbodies and filling up of ponds for built-up areas due to rapid urbanization. In the southern part of KCC, more than 90% of households are affected by waterlogging and 38% of inhabitants experience short-term waterlogging frequently (Murtaza ).
Annually, 1,808.5 mm (nearly 80% during the monsoon) rainfall is experienced by the city (Weatherbase ). The city dwellers are more vulnerable to waterlogging which results from a number of issues (i.e., unpredictable and intense rainfall, storm water stagnation, poor drainage capacity, and river water overflow). The KCC has spent USD 20 million over the last five years on the re-excavation of rivers and canals to improve the drainage system of KCC (The Daily Star ). However, waterlogging has remained a serious problem for the residents of the city. The methods of identifying affected areas and selecting priority projects by KCC have not yet been clarified, so it is not obvious which areas are actually affected by waterlogging. Spatial modeling of the drainage system has also not been studied and the impacts of waterlogging are not properly identified.
As a result, KCC is spending huge amounts of government funds on implementing random projects but the condition of waterlogging remains the same.
In this study, we investigated the waterlogging scenario and identified the actual affected areas based on the Mike Urban simulation model. A simulation study showed the future scenario and found cost-effective solutions, discarding the necessity of implementing large-scale projects. This study also identified the possible impacts of waterlogging, which might help to set projects' priorities. A raster-based drainage simulation model was formulated using climate change-induced rainfall events and existing physical conditions (i.e., land use, drainage network, road network, etc.). The model identified waterlogged areas and the gap in drainage capacity within KCC, which in turn helped identify the impacts of waterlogging on various infrastructures. The findings of the research will help KCC in effective decision-making for sites which require specific solutions. Incorporating the methods of this study with a drainage improvement plan for urban areas will add a new dimension and increase the efficiency of the plan. The objective of the study is to prepare a drainage simulation model for identifying waterlogged areas and to investigate the impacts of waterlogging considering existing drainage capacity.

Description of study area and materials
The south-west coastal region of Bangladesh ( Figure 1) is vulnerable to the impacts of climate-induced disasters (i.e., intense rainfall, sea level rise, coastal inundation, erosion, cyclones, saline water intrusion, etc.) (Alam et al. ).
The KCC is located in this zone (Ali et al. ) and is highly vulnerable to a number of disasters. Four wards (i.e., ward-27, ward-29, ward-30, and ward-31)  KDA, BMD, and ward councillors). Twelve focus group discussions (FGDs) (three with only females, three with only males, and six with both males and females) were also conducted to investigate the previous and existing condition of waterlogging.

Digital elevation model (DEM) and watershed analysis
A DEM of Khulna City was prepared through spot height, existing natural barriers (i.e., rivers and canals) and field observations. The DEM was then used as a main input for watershed analysis (Rivas & Koleva-Lizama ). The Arc Hydro tool was used to describe natural drainage patterns of the catchments (Abdullah ). On the other hand, raster analysis is simpler, a faster process, quicker, and suitable for analyzing attributes of large areas (Ratti & Richens ).  determine the return period, yearly maximum precipitation data were ranked from highest to lowest. This research used yearly maximum precipitation instead of yearly total precipitation to identify waterlogged areas developed from extreme precipitation events. The probability of occurrence was calculated using Equation (1) according to Ward & Trimble (): where, F a ¼ probability of occurrence (%), n ¼ rank of each event, y ¼ total number of events, and return period (P r ) ¼ 100/F a .
Return periods, probabilities of occurrences (as independent variables), and log yearly maximum precipitation (in millimeters as dependent variable) were plotted on log-probability graph paper. A regression line was drawn to derive a model. Finally, three scenarios (i.e., for 5 year, 50 year, 100 year) were developed to understand the extent of extreme precipitation events. Catchment-wise volume of overland flow was calculated using the following concept according to Subramanya (): Volume of overland flow ¼ ( precipitation for return period Ã area of catchment) À (loss due to natural process þ existing drainage volume) This research assumed the volume of overland flow losses due to some natural processes (i.e., percolation, evaporation, evapotranspiration, infiltration, etc.). Finally, mapping was done to determine the spatial variation of overland flow volume in each catchment in three return periods.

Simulation of drainage system
This research formulated a mathematical simulation model of the drainage system of KCC using Mike Urban 11.
To generate the hydrological model, nodes were defined in the intersection and end points of a drainage system. This work used some parameters for each node such as unique name, ground levels, coordinates, diameters of drain, and invert level. Ground levels for nodes were extracted from DEM and diameters for each node were assumed as the width of connecting drains. The invert level of the upper stream was assumed as 2 meters below the ground level (according to KCC drainage design) and additional nodes of the network were calculated using slope, and distance between nodes along the drainage network. The invert level for downstream was calculated using the following formula according to Gupta et al. ():

RESULTS AND DISCUSSION
This work analyzed the different physical components of the study area associated with climate change and waterlogging.           City are located in ward-31 along the Rupsa River and waterlogging might hamper the production of these industries.
Waterlogging also might pose a danger to the longevity of educational, community, and service facility buildings. The relation between economic loss and inundation depth for major structures (i.e., residential, commercial, and mixed use) is shown in Figure 9. According to the findings, economic damage might increase with the increase in inundation depths.

RECOMMENDATION
The simulation model should be incorporated with the development plan. Policy-makers may then easily find the waterlogging affected areas as well as identify the site-specific solutions. Using the simulation model, future scenarios might be projected without construction, which will save both money and time.
According to the findings, drainage conditions, drainage facilities, and drainage system should be improved to tackle drainage problems to reduce the potential waterlogging risk.
On the other hand, without any interventions in the drainage system in future, more waterlogging risk was expressed as the simulation was done based on existing scenarios.

CONCLUSION
Waterlogging in the study area is increasing due to the possible impact of climate change on precipitation. Inundated areas in four wards of the KCC were identified in this work.
It was found that physical structures on the stream links might affect natural flows and inundate the area. According to the precipitation analysis, 396 mm, 346 mm, and 180 mm rainfall might occur for return periods of 100 years, 50 years, and 5 years, respectively. This research identified that the volume of overland flow, considering climate change, was increasing gradually with the increase of return period.
According to the 100-year return period, the inundation depths were between 5.95 cm and 32.22 cm. As found by the simulation model, 62.52% of areas (i.e., around 1,000 acres) are inundated with different inundation levels.
According to the hydrological model, it was found that drai-