Setting Up SWAN for Coastal Wave Modeling

Introduction to SWAN

SWAN (Simulating WAves Nearshore) is a third-generation wave model designed for coastal regions and inland waters. This tutorial will guide you through:

1. Installation and compilation
2. Grid setup
3. Input file configuration
4. Running simulations

Installation

Prerequisites

# Ubuntu/Debian
sudo apt-get install gfortran make

# macOS with Homebrew
brew install gcc make

Compiling SWAN

# Download SWAN
wget http://swanmodel.sourceforge.net/download/swan4145.tar.gz
tar -xzf swan4145.tar.gz
cd swan4145

# Compile
make config
make install

Creating a Computational Grid

SWAN uses a structured grid. Here’s a simple example:

CGRID REGULAR -90.0 25.0 0.0 1000 800 0.01 0.01

This creates a grid:

• Starting at longitude -90°, latitude 25°
• With 1000 x 800 points
• Grid spacing of 0.01° (~1 km)

Input File Structure

A basic SWAN input file (INPUT) looks like:

PROJ 'CoastalWaves' 'v1.0'

SET LEVEL=0.0
MODE NONSTATIONARY

CGRID REGULAR -90.0 25.0 0.0 1000 800 0.01 0.01

INPGRID BOTTOM REGULAR -90.0 25.0 0.0 1000 800 0.01 0.01
READINP BOTTOM 1.0 'bathymetry.dat' 4 0 FREE

BOUND SHAPESPEC JONSWAP PEAK DSPR DEGREES
BOUNDSPEC SIDE WEST CCW CON PAR 2.0 10.0 270.0 20.0

GEN3 KOMEN
FRICTION JONSWAP
BREAKING

OUTPUT OPTIONS TABLE
BLOCK 'COMPGRID' NOHEADER 'output.dat' LAY 3 HSIGN TPS DIR DSPR

COMPUTE 20240101.000000 1 HR 20240131.000000
STOP

Boundary Conditions

Wind Input

INPGRID WIND REGULAR -90.0 25.0 0.0 100 80 0.1 0.1
READINP WIND 1.0 'wind.dat' 4 0 FREE

Wave Spectrum at Boundaries

BOUNDSPEC SEGMENT IJ 1 1 1 800 VARIABLE FILE 'boundary_spectra.txt'

Running SWAN

# Serial execution
./swan.exe INPUT

# Parallel execution with MPI
mpirun -np 4 ./swanrun -input INPUT

Post-Processing

Use Python to analyze SWAN output:

import numpy as np
import matplotlib.pyplot as plt

# Read SWAN table output
data = np.loadtxt('output.dat', skiprows=5)
x, y, hs, tp, dir, dspr = data.T

# Plot significant wave height
plt.scatter(x, y, c=hs, cmap='viridis')
plt.colorbar(label='Hs (m)')
plt.title('Significant Wave Height')
plt.show()

Tips and Best Practices

1. Grid Resolution: Balance between accuracy and computational cost
2. Time Steps: Use adaptive time stepping for efficiency
3. Validation: Always validate against observations
4. Boundary Conditions: Ensure proper offshore boundary placement

Common Issues

Convergence Problems

If SWAN doesn’t converge:

• Check bathymetry for unrealistic values
• Verify boundary conditions
• Adjust numerical parameters (ACCUR, DREL)

High Computational Time

• Use coarser grids for initial testing
• Enable parallel processing
• Optimize output frequency

Conclusion

SWAN is a powerful tool for coastal wave modeling. With proper setup and validation, it can provide accurate wave predictions for engineering and research applications.

Resources

• SWAN User Manual
• SWAN Examples