Flood Hydraulic Analyses: A Case Study of Amik Plain, Turkey

: In recent years, signiﬁcant ﬂood events have occurred in various parts of the world. The most important reasons for these events are global warming and consequent imbalances in climate and rainfall regimes. Many studies are performed to prevent the loss of life and property caused by ﬂoods. Many methods have been developed to predict future ﬂoods and possible a ﬀ ected areas. Developing computer and numerical calculation methods gives opportunities to make simulations of ﬂood hazards. One of the a ﬀ ected areas, which is also one of the world’s ﬁrst residential districts at Hatay in Turkey, is the Amik Plain. In this study, the ﬂoods on the Amik Plain in Hatay province are analyzed. Hatay airport was also a ﬀ ected during ﬂoods since 2012 and serious material damage occurred. For this purpose, Google Earth Pro software was used to obtain maps of the basin where the airport is located and the rivers it contains. Afterwards, Hydrologic Engineering Center’s River Analysis System module (HEC-RAS) was used for the hydraulic and hydrological deﬁnitions of the river basin. The results of numerical models are presented as simulated maps.


Introduction
Flooding is a condition that causes water and soil environments to be disturbed in areas that are not normally submerged owing to increased flow and rising level in a stream. This situation may harm humans and animals and damage land and property near the river. In Turkey, 695 floods occurred from 1975 through 2010, causing great economic losses second only to earthquakes. A total of 634 people lost their lives in these floods, 810,000 hectares of land were left under water, and the total loss was 3.717 billion dollars. During this period, the distribution of flood losses by sectors was found to be 45% agricultural areas, 32% settlements and infrastructure, 7% portable goods and vehicles, 1% transportation, and 15% in other areas [1,2]. These stats show that flood is an extremely serious and important issue. In addition to saving many lives, the works to be done to predict floods and protect them from danger will contribute a lot to the national economy. The main purpose of this study is to investigate how vulnerable the Amik plain is against flood disasters using geographic information systems and HEC-RAS.
Flood precautions and recreation studies are performed by state hydraulic works (DSI) in Turkey. State hydraulic works (DSI) is an institution responsible for the planning, management, development, bridges, settlements, and Hatay Airport runway and apron were exposed to flood waters. According channel were involved in the floods [3,17]. The flood hydrograph measured at the current 122 observation stations during this flooding is given in Figure 3. The Amik Plain in the Mediterranean region of Hatay Province is a plain where floods have been experienced in various years. According to the data obtained from the Provincial Directorate of Agriculture, between 27 January and 14 March 2012, rainfall and sudden snow melting caused flooding over a 13,000 ha area on the Amik Plain. A total of 8465 ha of agricultural land, roads, bridges, settlements, and Hatay Airport runway and apron were exposed to flood waters. According to the reports of the Provincial Directorate of Agriculture, the flood damage observed in agricultural lands cost 17,453,161 TL. The Asi River, Karasu stream, Afrin creek, Muratpasa channel, and Comba channel were involved in the floods [3,17]. The flood hydrograph measured at the current observation stations during this flooding is given in Figure 3. Water 2020, 12, x FOR PEER REVIEW 5 of 30  According to this, the highest hydrograph values are observed in the Afrin and Karasu channels. The Amik Plain flood lake elevation hydrograph during this flooding is given in Figure 4. Accordingly, the highest level was measured as 81.13 m on 21 February 2012.

128
According to this, the highest hydrograph values are observed in the Afrin and Karasu channels.

129
The Amik Plain flood lake elevation hydrograph during this flooding is given in Figure 4.     The Amik Plain is flooded almost every year as a result of rainfall. Similarly, as a result of heavy rainfall in February 2019, the water level increased and caused the Amik Plain to be flooded. Figure 6 shows the situation of the Plain in 2019.

Model and Application
In this study, the HEC-RAS program was used to model flooding. HEC-RAS was developed in 1994 by USACE (United States Army Corps of Engineers). It uses the addition of a graphical user interface (GUI), which differs from the HEC-2 model in the 1970s, for modeling the water profile. The GUI is Windows-based, allowing the user to enter data, correct the entries, and display and get the output of the analysis in an easy format [18][19][20][21]. The flowchart showing the adopted methodology of the program is shown below (Figure 7).  The HEC-RAS package program can perform four different river analyses in one dimension. These include the following:

1.
Calculation of regular current water surface profiles; 2.
Moving solid boundary sediment transport modeling; 4.
For these four approaches, geometric data can be defined in the program and geometric and hydraulic calculations can be made. Regular uniform flow represents the current state in which the flow characteristics do not change depending on the location and time. The main characteristic of the slow changing flow is the slight changes in water depth and speed from one section to another. The HEC-RAS program can solve current problems with subcritical, supercritical, and both.
One-dimensional energy equation is used by HEC-RAS to calculate water surface profiles. With this equation, the following variables are used, the friction coefficient in the Manning equation for energy losses, the kinetic energy correction coefficient for the change in velocity height owing to shrinkage and expansion changes, and the momentum equation where the water surface changes suddenly. In transitions from the river regime to the flood regime (hydraulic jump), the changes in the flow properties caused by the bridges and the analysis of the flow properties formed in the cross-sections of the rivers are also solved using these equations.

Basic Equations for Profile Calculation
Water surface profiles are calculated from one section to another by the method named as the standard step method based on the repeated solution of the energy equation. The energy equation can be written as follows and given Figure 8: where z 1 , z 2 , main channel bottom height (m); Y 1 , Y 2 , sectional flow depth (m); V 1 , V 2 , mean velocity (m/s); α 1 , α 2 ,velocity coefficients; g, gravitational acceleration (m/s 2 ); and he, energy loss height (m). The energy loss between two sections (he) consists of friction loss and contraction or expansion loss. Energy head loss is defined by the following expression: where L, channel length; S f , friction slope between two sections; and C, expansion or contraction loss coefficient.   L channel length is calculated as follows: where L lob , L ch , L rob , flow way length, left flood plain, main channel, and right flood plain; Q lob , Q ch , Q rob , mean discharge between sections, left flood channel, main channel, and right flood channel.

Transmission Calculation of Section Subdivisions
In order to determine the total flow rate and velocity coefficient in a cross section, the cross section of the channel must be subdivided where the distributions are uniform. HEC-RAS considers the roughness (n) changes on main channel and flood channel to subdivide the cross sections as shown in Figure 9. The transmission capacity for each sub-region is calculated with the Manning formula given below.
where K, transmission capacity for each sub-region; n, Manning roughness coefficient for each subregion; A, bottom sectional area; and R, bottom section hydraulic radius.   The program collects the transmission capacities of subfields on the right and left coasts to calculate the total transmission capacity and the transmission capacities of the main channel, which are usually taken as a single part.

Head Loss Calculation
In HEC-RAS, friction loss is calculated with the help of the friction gradient S f and the length L, given by Equation (3). The friction slope (slope of the energy line) in each section is calculated from the Manning equation as follows.

Determination of Contraction and Expansion Loss
Contraction and expansion losses in HEC-RAS are calculated with the help of the following equation: where C is the contraction or expansion coefficient. The software accepts that, if the speed load at the downstream is greater than that at the upstream, the expansion occurs when the speed load at the upstream is greater than at the downstream. Sample values for the C coefficient are given in Table 1 below.
The program tries to determine the water surface profile using the above equations with repeated solutions and limits the maximum number of iterations to 20 to balance the water surface. While the program uses the energy equation for regular currents, it uses the momentum equation in varied flow. In many flow situations, there is a transition from supercritical to subcritical or to subcritical to supercritical. This situation occurs in sections, bridge structures, fall structures and sluices, and flow joints, which are important changes in channel slope. The momentum equation is derived from Newton's second law of motion as follows and given Figure 10: where F x , force; m, mass; and a, acceleration. The software accepts that, if the speed load at the downstream is greater than that at the 210 upstream, the expansion occurs when the speed load at the upstream is greater than at the 211 downstream. Sample values for the C coefficient are given in Table 1 below.   If we apply Newton's second law of motion to the flow segment between Sections 1 and 2, the momentum change in unit time can be written as follows:

212
where P, 1 and 2 hydrostatic pressure force in cross sections; Wx, the x component of the weight force; F f , friction loss between Sections 1 and 2; Q, discharge; ρ, specific mass of water; and ∆Vx, the change in velocity between points 1 and 2 in the x direction. After performing the necessary intermediate operations, the shape of the momentum equation used by the software can be obtained.
An important term used to determine the regime of the flow, that is, subcritical and supercritical flow in varied flows using the momentum equation, is the Froude number and is expressed as follows.
where V, flow velocity; g, gravity acceleration; and h, flow depth. If Fr = 1, it is defined as critical current, Fr < 1, it is defined as subcritical current, i.e., river regime, and if Fr > 1, it is defined as supercritical current, namely flood regime.

Manning Coefficient for Main Channel
If the main channel roughness does not change, the section is not subdivided and the calculation is made by considering a single n value. If the main channel roughness is variable, HEC-RAS checks the compatibility of the roughness of the main channel sub-parts, and if necessary, calculates a single equivalent roughness coefficient of the entire section.
It is important to determine the appropriate Manning coefficient for the accuracy of the calculated water surface profiles. Manning n value is highly variable and depends on many factors; surface roughness, plant properties, channel irregularity, channel horizontal slope, meandering, scouring and accumulation, obstacles, channel shape and size, flow, seasonal changes, temperature, hangers, and swabs.
In the case in which there is water surface information of a cross section, the Manning value (n) can be calibrated. If there is no such information, similar current conditions or value obtained by experimental studies can be used for the n value. There are many sources for determining the value of Manning n for different channel properties. The broadest information for the value of "n" in streams and flood channels is given in Chow's Open Channel Hydraulics (1959) [22]. The information from this book is given in Table 2. The model can be calibrated with the n values of streams that are presented in the book and presented with pictures.

HEC-RAS Software Buttons
When the program is first opened, the function of the buttons in the window is shown in Figure 11. For more information, the user's guide can be read.

Creating Geometric Data
In this study, the geometric model of the flooding, the maps of the Amik Plain, and the channels and rivers seen in Figure 12 were digitized as accurately as possible using the Google Earth software [23].Parts of the Küçükasi River (1050 m); Karasu River (8050 m); and the Comba, Muratpaşa, and Afrin channels (5100 m, 6550 m, and 6600 m, respectively) are modeled and a cross-section of every 50 m is formed. The drawing of the river network was created with the sections in the window that came in by clicking the Geometric Data icon in the main menu of HEC-RAS, shown in Figures 11 and 13. The HEC-RAS model of channels and cross-sections is presented in Figure 13.
The lowest elevation levels in the numerical model were measured on the Küçükasi River +81 m, Karasu River +79 m, Afrin channel +79 m, and the Muratpasa and Comba channels +78 m. Afrin and Karasu are formed from two parts, considering the intersection points where the flow network was created. The sections of the flow paths were formed by entering data into the program similar to the cross-section values specified by the DSI in the flood report. Figure 14 shows these sections. There is a slope of 0.1% in the formed current networks. On the basis of the authors' field observations, water organization reports, and Google Earth software, the authors saw that channels exist generally from natural soil, and that is why the Manning coefficient used for the study is considered as 0.042 for both sides of the channels and 0.032 for the bottom of channels. The Manning coefficient is selected in accordance with the regional conditions according to the information in Table 2.    Figure 14 shows these sections. There

Hydraulic Analysis
The steady flow icon is selected from the main bar and the window in Figure 15 appeared. After the geometric model was generated, the values measured at the flow observation stations during the flooding were entered into the program, as shown in Figure 15. These values were taken from the DSI flood report. On the basis of the available data, the model was run for steady flow conditions and the results were examined in accordance with this situation. the geometric model was generated, the values measured at the flow observation stations during the 289 flooding were entered into the program, as shown in Figure 15. These values were taken from the 290 DSI flood report. On the basis of the available data, the model was run for steady flow conditions and 291 the results were examined in accordance with this situation. The results of the flow network were analyzed separately for each branch, and the following 296 benefits were obtained. As seen in Figure 16a Figure 17a that the Froude number indicates that the flow is sub-critical across all sections, 302 that is, for the standard river regime, and there is a slow rise from upstream to downstream. Figure   303 17b indicates a slow increase in the flow rate as well as in the Froude number plot. Figure 17c shows 304 that the water depth decreases slowly as the flow rate increases from upstream to downstream.

Results
The results of the flow network were analyzed separately for each branch, and the following benefits were obtained. As seen in Figure 16a Figure 17a-c graphically show the change in the Froude number, speed, and water depth change along all sections for the second part of the Karasu river. When the graphs are examined, it is seen in Figure 17a that the Froude number indicates that the flow is sub-critical across all sections, that is, for the standard river regime, and there is a slow rise from upstream to downstream. Figure 17b indicates a slow increase in the flow rate as well as in the Froude number plot. Figure 17c shows that the water depth decreases slowly as the flow rate increases from upstream to downstream. Because the simulation operates in steady flow conditions, the HEC-RAS program keeps the discharge value constant, and thus water depth decreases with increasing velocity values.
As seen in Figure 18, water depth is one meter above the first part of the Karasu river section (first 50 m) and is two meters above the second part (last 5250 m) of the Karasu river section. Figure 19 shows perspective views of the last four sections of the parts between 5100 and 5250 m. Figure 20a-c show the changes in the Froude number, velocity, and water depth along all the sections in the first part of the Karasu River. A similar situation is also observed in the second part of Karasu River. Froude number and velocity increase slowly, but the water depth decreases. Figure 21a,b relates to the Afrin channel analyses. It is understood from these figures that, when the discharge is Q = 256 m 3 /s, the water level is approximately two meters above the channel banks for the first part of the Afrin channel (first 50 m), and it is approximately three meters above the banks for the last 1250 m. Figure 22 shows a perspective view of the flow for all sections.

314
As seen in Figure 18, water depth is one meter above the first part of the Karasu river section 315 (first 50 m) and is two meters above the second part (last 5250 m) of the Karasu river section. Figure   316 19 shows perspective views of the last four sections of the parts between 5100 and 5250 m.

314
As seen in Figure 18, water depth is one meter above the first part of the Karasu river section 315 (first 50 m) and is two meters above the second part (last 5250 m) of the Karasu river section. Figure   316 19 shows perspective views of the last four sections of the parts between 5100 and 5250 m.  Changes in Froude number, velocity, and water depth are shown for the first part of the Afrin channel and graphs relating to them are presented in Figure 23. When the graphs are examined, it is seen that the Froude number increases rapidly, from a value of about 0.3 upstream and approaching 1.            Figure 24 shows that, at discharge Q = 256 m 3 /s in the Afrin channel second part, flooding occurs three meters above the channel section top height in the first 100 m and five meters above the channel section top height in the last 5300 m. Figure 25 shows perspective views of the last five sections between 5100 and 5300 m. Figure 24 shows that, at discharge Q = 256 m 3 /s in the Afrin channel second part, flooding occurs 348 three meters above the channel section top height in the first 100m and five meters above the channel 349 section top height in the last 5300m. Figure 25 shows perspective views of the last five sections 350 between 5100 and 5300 m.    In Figure 27a,b, at discharge Q = 128 m 3 /s, flooding occurs about three meters above the channel top level for the Muratpaşa channel first 50 m sections, and it is about four meters above the top level of the channel for the Muratpaşa channel last sections (6550 m). Figure 28 shows perspective views of the last four sections between 6400 and 6550 m.

363
In Figure 27a Figure 30a,b shows that, at discharge Q = 128 m 3 /s, flooding occurs two meters above the Comba channel top level for the first 50 m section, and ten meters above the channel top level for the 5100 m section. Figure 31 shows perspective views of the last five sections between 4900 and 5100 m.  380 Figure 30a,b shows that, at discharge Q = 128m 3 /s, flooding occurs two meters above the Comba 381 channel top level for the first 50m section, and ten meters above the channel top level for the 5100 m 382 section. Figure 31 shows perspective views of the last five sections between 4900 and 5100 m.    When the flood water levels found by means of the HEC-RAS program are examined, it is seen that the values are much higher than the land elevation (78 m-81 m) of the region. The digital elevation model (DEM) map relating to Hatay region is used to mark the flood area, and it is presented in Figure 36 [24].

396
In Figure 33a

430
In this case study, in order to understand the water level of Hatay's Amik Plain during the flood, 431 all important flow networks were examined together, and the effect of the flood was investigated.

432
The results of the study are as follows.

433
By examining the hydraulic conditions of each channel and river branch, it is seen that the The potential danger zone is obtained with reference to the natural elevations of the plain. According to the ESRI results, it is possible to see that a huge part of the plain including Hatay airport and agricultural areas are at risk of flooding.

Conclusions
Amik plain is originally a region consisting of artificially dried natural lakes and swamps under the control of local government. This area was considered as an agricultural area and improved. At an advanced stage, an airport was built in part of the plain in line with demands and research. However, in recent years, both agricultural fields and the airport have faced flooding events under climatic effects. With this study, the flood capacity of the artificial channels and the affected area was analyzed using HEC-RAS software (developed by U.S. Army Corps of Engineers Hydrologic Engineering Center-HEC).
In this case study, in order to understand the water level of Hatay's Amik Plain during the flood, all important flow networks were examined together, and the effect of the flood was investigated. The results of the study are as follows.
By examining the hydraulic conditions of each channel and river branch, it is seen that the sections of Muratpaşa, Comba channels, and Afrin River should be increased. The presented hydraulic modeling method is sufficient to ensure flood safety, but it must be done before proposing the flood protection structure.
The highest flood value was found in the smallest channel Comba. As Hatay airport is close to this channel, necessary measures should be taken against high flood conditions in and around the airport. Considering this situation, it has been determined that it is appropriate to re-plan the Comba drainage channel.
According to the data obtained from the previous studies and model results of Amik Plain and changes in water level, it is seen that more measures should be taken during floods.
It is concluded that the capacities of measured excessive discharge values, Muratpaşa, Comba channels, and Afrin river sections are not sufficient to carry these discharge values directly to Küçükasi River.
According to the evaluation of the results obtained in this study and the similar studies presented here, the flood risk of the area is expected to increase day by day with the influence of changing climate conditions.
Depending on the changing climatic conditions, Amik Plain's research on the relationship between precipitation and discharge will help engineers to design optimum channels that represent measures against possible floods in the future.