Flood has been recognized as the number one disaster in many parts of the world mainly affecting Asia and South America and cause tremendous damages to the properties, environment, and losses of life (Adikari et al. 2010; Marengo et al. 2013). Due to global warming, future rainfall is predicted to be more intense, resulting in increased flood peak, volume, and duration (Westra et al. 2014; Wang et al. 2014). In addition, the rate of sea-level rise is expected to be higher in the future which further amplifies the impacts of the flood (Nijland 2005). Significant damage from floods also happens due to the people living in high flood risk areas which are unsuitable for settlement (Elsheikh et al. 2015; Ghorbani et al. 2016).
Throughout history, Malaysia has faced several major floods. In particular, during the 2006/07 flood event, Kota Tinggi town in the State of Johor has recorded the highest, worst and costliest flood in history of the Johor River. The estimated total loss due to these disasters was RM 1.5 billion of which, RM 237.1 million was for damaged infrastructure alone (Hamzah et al. 2012; Tam et al. 2014). A total of 11,724 victims in the Dec 2006 flood event and 7,915 victims in Jan 2007 were evacuated (Hamzah et al. 2012; Tam et al. 2014; Karki 2019). In some cases, flood victims had to move to another relief center as the evacuation centers were also flooded.
It was found that the main cause of the flood was due to a large amount of rainfall, geographically low laying area, rapid land-use changes, and tidal effect (Tam et al. 2014; Karki 2019). The flood started on Dec 19, 2006, until Jan 16, 2007, where the first wave had inundated most of the Kota Tinggi town. Malaysia Metrological Department (MMD) reported high rainfall intensities from Dec 19, 2006, to Jan 12, 2007. The second wave continued on Jan 11, 2007, and this caused a huge disaster within a short period. The rainfall total during the first wave double compared to the average monthly rainfall, and during the second wave, the rainfall was about 5 times higher compared to the average monthly rainfall in Dec and Jan at Kota Tinggi. It has been reported that the water level was up to 5.45 m during the second wave compared to 4.9 m during the first wave. The disaster was compounded by the low infiltration rate of the soils in the area. At the same time, the Department of Irrigation and Drainage (DID) reported that a high tide of about 2 m occurred at the river mouth causing a backwater phenomenon (Tam et al. 2014; Karki 2019). Consequently, the outflow of stormwater in tidal detach, creeks, and rivers was blocked by the increasing sea water level. This problem is expected to recur year after year.
Flood modeling is an important tool in predicting the consequences of floods which provides useful information in managing the potential risks caused by flooding (Nkwunonwo et al. 2020). To cater to the needs to further understand the short- and long-term flood events, there is an increasing interest to develop a new hydrologic and hydrodynamic model in order to achieve better flood modeling results in terms of visualization and characterization. Typically, a model needs to be chosen based on appropriate catchment characteristics, model input parameters, and boundary conditions. HEC-HMS is one example of a semi-distributed and integrated hydrological modeling system. It models precipitation runoff with hydrological channel routing methods. It is applicable in a wide range of geographic areas for solving the widest possible range of water availability, flow forecasting, urban drainage, future urbanization impact, reservoir spillway design, flood damage reduction, floodplain regulation, and systems operation. In Malaysia, HEC-HMS was frequently used by consultancy projects and most preferred by university students because it is free and easy to download from the US Army Corps website (Howe Lim and Melvin Lye 2003). In addition, HEC-HMS was used by DID Malaysia for the Kelantan River Flood Forecasting program (Ramachandra Rao and Hamed 2019). For hydrological simulation, HEC-HMS provides an event-based on fully hydrological and metrological parameters. It has many references to be a guide for the simulation process.
In flood modeling, the tide level became one of the important elements especially when the area affected is close to the sea. HEC-RAS is one of the most widely used models, offered a known water surface as model input for tide level (Fan et al. 2012; Romali et al. 2018; Muñoz et al. 2021). This model computes water surface profiles and energy grade lines in 1D, steady-state, and gradually varied flow analyses. It has been applied extensively in studying the hydraulic characteristic of rivers (Thakur et al. 2017; Ogras and Onen 2020). Natale et al. (Natale et al. 2007) and El-Naqa and Jaber (El-Naqa and Jaber 2018) used HEC-RAS and HEC-GeoRAS as hydraulic model software to run the flood modeling with the contribution of tidal effect. Their analysis of floodplain showed the effect of floodplain modeling is much better in terms of flood extent and depth. A previous study by Shabri et al. (Shabri et al. 2011), showed high accuracy of the HEC-RAS model in simulating unsteady tidal flow under natural conditions. Therefore, flood modeling will be more significant with consideration taken on the tidal effect at downstream boundary conditions. In addition, the integration of HEC-GeoRAS into the hydrologic and hydraulic modeling provides detailing on flood mapping in the form of flood depth, the velocity of flow, and flood duration.
To complement the flood model, flood frequency analysis has been primarily used to analyze annual peak flow for large and mid-size catchments (Shiau and Shen 2001). Flood frequency analysis is usually applied for model validation of the observed and simulated hydro-climatic variables such as rainfall and streamflow (Ramachandra Rao and Hamed 2019). Various types of statistical distribution methods were applied and developed by researchers all over the world to determine the most appropriate data distribution for flood frequency analysis and to model the long-term flood characteristics (Sraj et al. 2015; Machado et al. 2015; Gizaw and Gan 2016; Serago and Vogel 2018). In general, a distribution with a larger number of flexible parameters such as GEV will be able to model the input data more (Ramachandra Rao and Hamed 2019). However, the present study will examine the performance of five probability distribution models, namely GEV, Lognormal, Pearson 5, Weibull, and Gamma for modeling the annual flood of the Johor River basin (JRB). These models were chosen because they are commonly recommended by many researchers (Kim et al. 2017; Langat et al. 2019).
As mentioned above, flood mapping at Kota Tinggi town is important to predict the possibility of flood occurrences and formulate an emergency action plan, insurance policy, and development planning. 2D numerical hydraulic models are considered advanced enough for the prediction of flood extent, depth, and flow velocities (Romali et al. 2018; Muñoz et al. 2021). Thus, the present analysis develops 2D flood mapping using the hydrodynamics model in order to estimate the present and future floods. One of the common practices in hydrology is estimating the Annual Exceedance Probability (AEP) and Average Recurrent Interval (ARI) (Ramachandra Rao and Hamed 2019). ARI refers to the return period in time between the events that have the same magnitude, volume, and duration. Information on ARI derived from frequency analysis is crucial for hydrologic analysis and designing hydraulic structures. This study concerned with developing an appropriate method for flood mapping by estimating the ARI of an annual flood using flood frequency analysis, examining the effect of the tide on flood modeling results, and mapping the 2006/07 flood and the simulated floods for 25, 50, 100 and 200 year return periods.