Flood Mapping Using HEC-RAS and HEC-HMS: A Case Study of Upper Citarum River at Dayeuhkolot District, Bandung Regency, West Java

Dayeuhkolot district is an area within Citarum River watershed that often experiences flood during the rainy season due to overflow discharge. This research identifies the flood area by using hydrodynamic modeling in Dayeuhkolot district. The model utilized satellite rainfall data. The hydrodynamic modeling of HEC-RAS 5.0.5 and HEC-HMS simulate the flow regimes and the hydrological characteristics in the study area, respectively. The results reveal that the increase in water level in two river tributaries within Citarum River watershed had caused the flooding at Dayeuhkolot district. Furthermore, the percentage of flood areas with low, medium, and high flood risk is 34.7%, 34.5%, and 30.8%, respectively. The flood map developed in this study could support the establishment of a strategy for managing the floodplain area and flood-affected areas.


Introduction
Rivers play an important role in human life, as evidenced by the various activities and activities of river utilization that are becoming increasingly widespread and complex, such as a source of energy for generating electricity, sources of irrigation or transportation, and transportation facilities.However, the river's function is impeded by changes in land use resulting from population growth, which is accelerating annually.The transition of land use functions in river areas in Indonesia results in flood disasters.Land use will continue to evolve to accommodate the requirements of urban residents, whose numbers continue to rise due to urbanization.Consequently, this activity frequently results in an over-exploitation of nature, including a decline in the environment's carrying capacity and uncontrolled land use changes.The tendency of growing disasters, particularly floods, in terms of quality and quantity, is a result of these activities, or it can be understood that the increase in land utilized for structures, as well as the decline in green open space, is a result of population growth [1,2].The Citarum River watershed is one of Indonesia's two strategic regions with the highest population growth.Therefore, this is the primary reason for the growth in land designated as critical.In addition, the upper Citarum watershed is shaped like a basin when viewed from a physiographic perspective.This basin is formed by the rugged and hilly terrain surrounding the watershed and the plain in its center.Physiologically speaking, the rain and flooding surface flows that occur in the upper Citarum Watershed are the outlet locations for the Upper Citarum River tributaries that are too close to each other, causing a parallel flow pattern.As with other rivers, the Upper Citarum River in the Bandung 1324 (2024) 012103 IOP Publishing doi:10.1088/1755-1315/1324/1/012103 2 region has simple bends, modest river gradients, and the possibility of river silting due to heavy sedimentation [3,4].During the rainy season, various watersheds of the Citarum River have experienced severe flooding.The Dayeuhkolot region and its environs are some of the locations that have been seriously affected and often inundated.This results from the accumulation of water volume from the Cikapundung and Cisangkuy tributaries, which causes the water level in the Dayeuhkolot region to rise.This demonstrates that the flooding has resulted in severe losses for the impacted areas [5].
Rainfall data is a crucial component of hydrological analysis because it is required for analyzing the transformation of rain, drought, flooding, and other needs.However, due to different constraints, there are frequently gaps in rainfall data or inadequate rainfall data [6].The number of surface rainfall observation stations established in the Upper Citarum watershed does not accurately reflect its topography.In addition, only seven rain stations record rain events continually in the Upper Citarum watershed, whereas the remaining 11 measurement stations do not typically collect rain data continuously.This is what decreases the dependability of rainfall observation data.The solution to the problem of observational rainfall data in the Upper Citarum watershed is to use simulated or estimated rainfall data from meteorological satellites, such as TRMM satellite rainfall data [7].
The rainfall-runoff transformation approach is utilized more frequently because data is easier to collect.The information consists of rain data in the form of short-and long-duration precipitation and the physical state of the existing watershed.In order to generate a flood hydrograph for use in this analysis, the rainfall-runoff transformation approach requires the use of three additional models: the loss method, the transformation method, and the base flow method.In the application, numerous computer models, including HEC-HMS, use the rainfall-runoff transformation method to perform flood analysis [8,9].HEC -HMS is a computer program that can model the complete hydrological cycle (rain transformation) and routing processes in watersheds.Runoff volume, direct runoff, baseflow, and channel flow are models that the HEC-HMS program may simulate or calculate.The US Army Corps of Engineers established the HEC-HMS program as one of its initiatives [10].HEC-RAS is one of the most popular models utilized by hydraulic engineers for channel flow analysis and floodplain delineation (United States Army Corps of Engineers [11].The HEC-RAS software will be utilized to analyze the flood inundation in the Dayeuhkolot District.HEC-RAS is a program for hydraulic modeling of the flow of streams through natural rivers and other channels.HEC-RAS requires the river discharge, channel, floodplain geometry, and channel resistance as inputs.HEC-RAS is intended to do one-dimensional and two-dimensional hydraulic computations for a whole network of natural or artificial channels.The 1D model assumes all water flows in the longitudinal direction, while the 2D model accounts for longitudinal and lateral water flow.In addition, in a 1D model, river bathymetry (bed topography) is represented by a series of cross-sections, which do not accurately depict the actual terrain.In a 2D model, however, this is represented by a continuous surface model, i.e., a finite element mesh, which more accurately depicts the actual topography [12,13].Geoinformatics provides numerous tools for the study and visualization of temporal and spatial data.It is feasible to simulate water surface profiles for progressively varying steady flow and the effects of various impediments, such as culverts, bridges, overbank constructions, and weirs, with HEC-RAS.ArcGIS is frequently used in conjunction with HEC-RAS ESRI [14].It consists of a collection of tools, utilities, and methods for managing geographic data [15].
Based on the description of the conditions in the Dayeuhkolot District, it is necessary to analyze the flood disaster to determine the amount of water level rise in the Upper Citarum River and the steps that must be taken to reduce risk by designing an appropriate flood control system.Flood inundation must be evaluated and analyzed before designing a flood control system in a given area.Flood mapping is required to determine the characteristics of floods, such as flow velocity, depth of flood inundation, and extent of flood inundation; this is a prerequisite for flood management planning.Flood maps are also important for land planners since they can be used to determine flood insurance and coverage.Using HEC-RAS version 5.0.5 and a case study of the Dayeuhkolot District, which is located in the Upper Citarum Watershed, an analysis of flood inundation will be conducted.

Methodology 2.1. Study Area
Upper Citarum River travels through Bandung District (passing through the subdistricts of Rancaekek, Sapan, Baleendah, and DayeuhKolot) to reach Saguling Dam.It is located between 6°43'21,8" and 7°19'38,1" south and 107°32'2" and 107°53'51,6" east.Long ago, the Upper Citarum watershed began to experience flooding.The capital city of Bandung was relocated from Krapyak (now: Dayeuhkolot) to Central Bandung in 1810 due to the annual flooding that occurs during the wet season.The area of interest in this study is shown in Figure 1.The prolonged flooding in this region results from the flat area in the centre of the basin and the surrounding hilly region.For instance, flooding in Dayeuhkolot is challenging to drain.It is believed to be induced by the topographical condition of the lowlands.Currently, the causes of flooding are more complex.In addition to heavy precipitation and insufficient channel capacity, population increase imposes an undue burden on the environment, particularly in terms of absorbing and storing precipitation.Also, how people manage the environment causes things like deforestation, land conversion, livestock breeding, unequal farming, household waste, irregular use of groundwater, and industry, all of which make floods like the one in the Upper Citarum Watershed even worse [15].

Data Collection
The national digital elevation model is constructed from multiple data sources, including IFSAR data (5 m resolution), TERRASAR-X data (5 m resolution), and ALOS PALSAR data (11.25 m resolution), with the addition of Masspoint stereo-plotting data.Using the EGM2008 vertical datum, the spatial resolution of the Digital Elevation Model and National Bathymetry (DEMNAS) is 0.27 arcsecond.Method of incorporating/adding mass point data to a Digital Surface Model/DSM (IFSAR, TERASAR-X, or ALOS-PALSAR) utilizing GMT-surface with tension 0.32 [16].Due to its reasonable precision and free availability, an 8 m × 8 m DEM from the shuttle radar topography mission was used to outline the watershed limit, identify sub-catchment boundaries and a stream network, and assess the terrain's drainage patterns.The river and floodplain geometry was produced by using the data from DEMNAS [17].Google Earth photos were utilized to classify land uses, and subsequently, these classes were used to estimate Manning's n values, which HEC-RAS required to complete hydraulic computations.The Tropical Rainfall Measuring Mission (TRMM) satellite from NASA collects worldwide average daily precipitation data with a grid of 0.25° × 0.25° or 28 × 28 km2.At two gauging sites, Dayeuhkolet and Nanjung, discharge data has been accessible for the year 2004 until 2013, which were utilized in this study.There are inconsistencies between the TRMM rainfall data and the actual rainfall data for the Upper Citarum Watershed, although the TRMM rainfall pattern still corresponds with the actual rainfall pattern.According to the analysis results, the correlation coefficient is 0.67.This correlation coefficient value suggests that these TRMM rainfall data can be used for research because the value is more significant than 0.67 [18].

Methods
The methodology used to generate floodplain maps included the following steps (Figure 2): (1) flood frequency analysis of the available observed discharge data to obtain floods with different return periods; (2) preparation of DEM based on DEMNAS data [17]; (3) delineation of watershed and drainage network using ArcGIS; (4) Hydrology analysis using HEC-HMS; (5) application of HEC-RAS to several potential flow scenarios corresponding to floods with different return periods; and (6) ArcGIS Pro floodplain mapping preparation.Peak discharges for various return periods are necessary to estimate water surface profiles and the amount of flooding under various flood intensities.The Dayeuhkolot District's Citarum River was subjected to a flood frequency analysis to determine flood peaks for various return periods.Frequency analysis employed the log-Pearson type III distribution [18,19], Pearson distribution, Gumbel's or extreme value distribution [20], log-normal distribution, and Normal distribution, the five most widely employed frequency distribution functions for estimating extreme floods.The Kolmogorov-Smirnov (KS) test with a 95% confidence interval was utilized to establish the optimal distribution for estimating flood peaks.

Hydrologic model HEC-HMS
The HEC-HMS model has been widely used in many hydrological studies.Figure 2 shows the process of this model to determine flood water discharge hydrograph.The hydrograph describes the conditions or characteristics of a watershed.Therefore, if the features of a watershed change, so will the hydrograph's form.Consequently, modeling the proposed flood hydrograph using the HEC -HMS software will require the sub-watershed area, the river's length, and the river's height to build a hydrograph that matches the characteristics of the Upper Citarum watershed.Figure 3   Due to the infiltration process, the loss rate method will occur using the SCS Curve Number method in this case study.In the HEC -HMS modeling, the SCS Curve number contains multiple parameters that must be provided, including initial loss, curve number, and imperviousness.Several parameters relating to land usage and soil layer porosity are assumed empirically by the parameter curve number.The land use and land change (LULC) of the Upper Citarum can be seen in Figure 4.After assigning n values to each land class and finishing all geometric data needs, the ArcGIS Export File was imported into HEC-RAS.This procedure required the transfer of geometric data from 6 ArcGIS to HEC-RAS.The next step was to provide the steady flow editor with steady flow data and boundary conditions.The peak flows are input into HEC-RAS to determine the anticipated flood levels along river reaches that extend through populated regions of the basin.Upon completion of data entry, the HEC-RAS model was performed to conduct a comprehensive flow analysis.The model created a complete analysis report that included the flow depth, discharge at each cross-section, and other information.After correcting all errors, the findings were exported as a RAS Mapper Export File to ArcGIS.After generating water surface and floodplain delineation, the RAS Mapper Export File was imported into ArcGIS to create a natural floodplain map.

HEC-HMS Flood Discharge Modeling
Flood modeling begins with the process of calibration and verification of certain flood events that occur in the study location.By using calibrated parameters, flood estimation can be done.The first step is to calculate the rainfall with return periods of 2, 5,10, 20, 25, 50, and 100 years at each available post with various probability distributions (Normal, Log Normal 2 and Parameter 3, Gumbel method, Pearson III, and Log Pearson III) followed by distribution analysis using the Smirnov method -Kolmogorov to get a distribution that corresponds to the smallest maximum deviation.The number of curves is also calculated and classified based on land use.After the rainfall and curve numbers are obtained, the two data are used to model the flood discharge hydrograph in the HEC-HMS.Flood calibration is an important process for making proper flood modeling.At this location, 2 flood events were recorded which occurred in February 2007 and March 2008.Use these 2 flood events for flood calibration and verification to find calibrated hydrological parameters.For calibration, the March 2008 events were used.After the flood hydrograph modeling process is shown in Figure 5a.From that figure also it can be seen that the modeling result and observed flow is also the same which means the model is calibrated.The next step is to validate the calibrated model.For the verification process, the February flood event is used.Using the same parameters from calibrated models, figure 6 shows the result of the simulation that has been validated, where the resulting peak discharge yields a 3% difference.This means the calibration and verification process is completed and these calibrated and verified parameters can be used for flood frequency analysis.The HEC -HMS Basin model is considered stable if the discharge is entered into the software and then computationally performed and produces the same peak discharge as the measured discharge data on the same day.The Citarum -Dayeuhkolot river discharge results using the SCS transform method show a significantly lower error rate which is shown in Figure 5, as determined by

Floodplain Map
Maps of floodplains were created utilizing elevation, land cover, geological, physiographic, and basin network information.Using the basin's water surface data and DEM, the flooded region during different return period floods was delineated.HEC-RAS water surface profile data were retrieved using RAS Mapper and included in an ArcGIS floodplain map.The data retrieved are listed in Table 1.The flood depth was present following the National Board of Disaster Management (BNPB) classification of flood risk [22].In regard to the total flood areas, the percentage area of flood risk for low, medium, and high risk is 34.7%, 34.5%, and 30.8%, respectively.Considering the data's compatibility with the flood inundation simulation and the Flood Control Guidelines, there are two primary choices available for utilizing the projected flood return period: 25 years and 50 years.The economic analysis for assessing flood-related losses necessitates the utilization of a planned flood with a return period of 25 years in its final phase.Hence, to evaluate the most effective flood management system, this study will employ a flood return time of 25 years.Dayeuhkolot District, situated in the Bandung Basin, falls within the Upper Citarum River Basin, an area prone to frequent flooding.Additionally, as depicted in Figure 7, the severity of flooding is more pronounced upstream compared to downstream of the river.This discrepancy may be attributed to the presence of numerous textile factories in the upstream area, as it is an industrial zone.Consequently, the discharge of liquid waste from these textile factories could potentially cause sedimentation in the upstream Citarum River that leads to a reduction of river capacity.The evaluation of a flood-prone area is a prior activity for mitigating flood damage [12].This map will support the establishment of a strategy for managing the floodplain area and flood-affected areas based on current land use patterns, taking into consideration potential and natural factors.The solution that can be applied to the planned flood discharge for the 25-year return period is to construct an embankment in the form of a retaining wall and conduct river normalization such as dredging to increase the river's cross-sectional capacity.However, based on the study of Paarlberg et al. [23], river normalization through dredging should be conducted with dredge-and-dump strategies targeted to harmonize natural river processes and to counteract unwanted riverbed levels.The river must be normalized so that flood control measures yield the best possible results.

Conclusions
The upper Citarum River is prone to a rise in water levels because of heavy rainfall and elevated discharge rates, resulting in the overflow of river water.This water flows toward the riverbank area and the neighboring areas close to the river, especially those with low elevation.Using the hydrological model HEC-HMS, the analysis of the hydrograph of the upper Citarum River at Dayeuhkolot district revealed that the maximum flow is ranging between 455 m 3 /s and 798 m3/s during different periods of 2,5,10, 20, 25, 50 and 100 years.HEC-RAS software was utilized to simulate floods and provided information on impacted areas and flood depth.According to the BNPB categorization of flood risk, this study shows low, medium, and high flood risk with a coverage area of 34.7%, 34.5%, and 30.8%, respectively.The study suggests building an embankment such as a retaining wall and removing sediment from the riverbed to prevent overflow during heavy rainfall.Further research should be conducted using HEC-HMS and HEC-RAS to simulate flood mitigation scenarios such as embankment structures on the side of the river.

Figure 1
Figure 1 Study area of Dayeuhkolot District, Bandung Regency, West Java Province below is the watershed basin model.

Figure 3 Figure 4
Figure 3 Upper Citarum Watershed map [21] (a) and model in this study (b)

a b Figure 5
HEC-HMS calibration result on March 2008 Event (a) and validated result on February 2007 event From Figure 5a there are 2-line colors, the red one shows the modeling result's hydrograph and the black one shows the observed peak flow.

Table 1 .
Peak discharge, area, flood depth, and flood risk in various return periods