### 1. Introduction

### 2. Numerical Model

### 2.1 Outline of the Model

### 2.2 Wave and Nearshore Current Models

### 2.3 Sediment Transport and Water Depth Change

*h*is the water depth,

*E*is the dimensionless coefficient.

_{s}*q*and

_{bx}*q*are the bed loads, which are estimated based on Watanabe et al.(1985). The bed loads are given by

_{by}*A*and

_{w}*A*are dimensionless coefficients,

_{c}*u*is the friction velocity, is the critical friction velocity,

_{*}*u*

_{*c}is the steady current vector with river flow. In case the Q3D mode use, the bed load due to neashore current velocities at sea bottom is determined. The coefficients

*A*and

_{w}*A*are given by a function of the median diameter

_{c}*d*

_{50}(Shimizu et al., 1996).

*Q*

*is the difference between the upward sediment flux*

_{s}*F*

*and the down ward flux*

_{z}*w*, which is proposed by Sawaragi et al. (1985), and given by

*C*_{f}*F*is determined by

_{z}*a*(0≤

*a*≤1.0) is the dimensionless coefficient,

*C*

_{0}is the concentration at reference point (Sawaragi et al., 1983). The concentration of suspended load

*C*is determined by solving the depth-averaged advection diffusion equation, as given by

*U*and

*V*are the depth-averaged current velocities. ε

_{x}and ε

_{y}are the diffusion coefficients, given by .

### 3. Model Tests for Formation of Sand Bar

### 3.1 Topographic Features around River Mouth

### 3.1 Model Setup

*ε*,

_{s}*C*and

_{w}*α*, associated with the sediment transport. In CASE 5, Q3D mode, which is three-dimensional computational mode, was used in order to perform formation of offshore sand bar as shown in Fig. 2 (d). The final bathymetry after 2 days was simulated. The computations of wave and nearshore current models were carried out twice a day.

### 3.2 Computed Results

#### 3.2.1 Wave Dominating Type without River Flow

#### 3.2.2 Wave Dominating Type with River Flow

*C*= 0.00035 and

*C*= 0.001, respectively. The sediment concentration

*C*at the up-stream boundary (x = 700m) was set. We found that the contour line of 2m in the case of

*C*= 0.001 differs from that in the case of

*C*= 0.00035. The contour line of 2m in Fig. 7(b) was advanced, and then the terrace for

*C*= 0.001 is larger than that for

*C*= 0.00035. These computed results are due to the effect of suspended sediment transport computed by advection and diffusion model.

#### 3.2.3 Wave Dominating Type under Stormy Wave

### 4. Field Verification

### 4.1 Field Site

### 4.2 Model Setup

### 4.3 Computed Results

### 5. Conclusions

• Wave dominating type without river flow

2DH model was used under normal wave condition, which reproduced remarkable river-mouth bar. In the case of oblique incident wave, a sand spit was grown and then the blockage of river mouth was occurred. On the contrary, Q3D hydrodynamic model was used under stormy wave condition. Not only sand spit at river mouth but also offshore sand bar was developed. We confirmed that the offshore sand bar was formed by undertow effect.

• Wave dominating type with river flow

The blockage did not occur under normal wave condition although river-mouth bar was reproduced. Instead, a terrace in front of the river mouth was reproduced. Furthermore, by considering the discharged sediment from river, the deposition of the discharged sediment became larger and then the depth contour in front of the river mouth was advanced.