DIWASP-Python

DIrectional WAve SPectrum analysis - Python port of the DIWASP Matlab toolbox

DIWASP is a comprehensive toolbox for estimating directional wave spectra from measurements of water surface elevation, pressure, velocity, or acceleration. This Python implementation provides a modern, user-friendly interface built on top of industry-standard scientific Python libraries.

Features

• Multiple estimation methods: DFTM, EMLM, IMLM, EMEP, and BDM
• Flexible input formats: pandas DataFrame or xarray Dataset
• Seamless integration with wavespectra for advanced wave analysis
• Windowed analysis: Process continuous time series with configurable window length and overlap
• Multiple sensor types: Pressure, velocity, acceleration, surface elevation, and more
• Array configurations: Support for single sensors or multi-sensor arrays
• Modern output: Returns wavespectra-compatible xarray Datasets with comprehensive metadata

Quick Start

Installation

pip install diwasp

Basic Usage

import pandas as pd
from diwasp import diwasp

# Load your wave data (pressure and velocity measurements)
df = pd.read_csv('wave_data.csv', index_col='time', parse_dates=True)

# Run directional wave analysis
result = diwasp(
    df,
    sensor_mapping={
        'pressure': 'pres',      # Map column names to sensor types
        'u_velocity': 'velx',
        'v_velocity': 'vely',
    },
    window_length=1800,          # 30-minute analysis windows
    window_overlap=900,           # 15-minute overlap
    depth=15.0,                  # Water depth in meters
    z=0.5,                       # Sensor height above seabed
    method='imlm',               # Iterative Maximum Likelihood Method
)

# Access wave statistics
print(f"Significant wave height: {result.hsig.values} m")
print(f"Peak period: {result.tp.values} s")
print(f"Peak direction: {result.dp.values} degrees")

# Save results
result.to_netcdf('wave_spectra.nc')

Estimation Methods

Method	Description	Speed	Accuracy
dftm	Direct Fourier Transform	Fastest	Lowest
emlm	Extended Maximum Likelihood	Fast	Good for narrow spectra
imlm	Iterative Maximum Likelihood	Medium	Good balance (default)
emep	Extended Maximum Entropy	Medium	Robust to noise
bdm	Bayesian Directional	Slowest	Highest

Supported Sensor Types

• Surface elevation (elev)
• Pressure (pres)
• Velocity components (velx, vely, velz, vels)
• Acceleration components (accx, accy, accz, accs)
• Surface slopes (slpx, slpy)
• Displacement (dspx, dspy)

Output Format

The function returns a wavespectra-compatible xarray Dataset containing:

Dimensions:

• time: Center time of each analysis window
• freq: Frequency bins (Hz)
• dir: Direction bins (degrees, nautical convention)

Variables:

• efth: Spectral energy density (m²/Hz/degree)
• hsig: Significant wave height (m)
• tp: Peak period (s)
• fp: Peak frequency (Hz)
• dp: Peak direction (degrees)
• dm: Mean direction (degrees)
• spread: Directional spread (degrees)

Examples

Pressure-Velocity-Velocity (PUV) Sensor

from diwasp import diwasp
import pandas as pd

# Load data
df = pd.read_csv('puv_data.csv', index_col='time', parse_dates=True)

# Analyze
result = diwasp(
    df,
    sensor_mapping={'p': 'pres', 'u': 'velx', 'v': 'vely'},
    window_length=1800,
    window_overlap=900,
    depth=20.0,
    z=0.5,
    method='imlm',
)

# Plot time series
result['hsig'].plot()

Pressure Gauge Array

# Multiple pressure sensors in triangular configuration
result = diwasp(
    df,
    sensor_mapping={'p1': 'pres', 'p2': 'pres', 'p3': 'pres'},
    window_length=1800,
    window_overlap=900,
    depth=10.0,
    x={'p1': 0, 'p2': 5, 'p3': -5},
    y={'p1': 0, 'p2': 5, 'p3': 5},
    z=0.0,
    method='emep',
)

Using xarray Dataset

import xarray as xr

ds = xr.Dataset({
    'pres': (['time'], pressure_data),
    'velx': (['time'], u_data),
    'vely': (['time'], v_data),
}, coords={
    'time': time_values,
})

result = diwasp(
    ds,
    sensor_mapping={'pres': 'pres', 'velx': 'velx', 'vely': 'vely'},
    window_length=1800,
    window_overlap=900,
    depth=15.0,
)

Documentation

Full documentation is available at https://wavespectra.github.io/diwasp-python

• Installation Guide
• High-Level Wrapper Function
• Estimation Methods
• Data Structures
• API Reference

Requirements

• Python >= 3.9
• numpy >= 1.20
• scipy >= 1.7
• xarray >= 0.19
• wavespectra >= 4.0
• pandas >= 1.3

Development

Install for Development

git clone https://github.com/wavespectra/diwasp-python.git
cd diwasp-python
pip install -e ".[dev]"

Run Tests

pytest

Code Formatting

black diwasp tests
ruff check diwasp tests

Citation

If you use DIWASP in your research, please cite the original DIWASP toolbox:

Johnson, D. (2002). DIWASP, a directional wave spectra toolbox for MATLAB: User Manual. Research Report WP-1601-DJ (V1.1), Centre for Water Research, University of Western Australia.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Credits

• Original DIWASP Matlab toolbox: David Johnson, University of Western Australia
• Python port: Developed with support from the oceanographic community
• Built on wavespectra, xarray, and pandas

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Related Projects

• wavespectra - Wave spectra analysis library
• Original DIWASP (Matlab) - The original Matlab implementation