Skip to content

README_bottle_simulation.md

Latest commit

 

History

History
161 lines (124 loc) · 6.04 KB

README_bottle_simulation.md

File metadata and controls

161 lines (124 loc) · 6.04 KB
Raw

Mediterranean Bottle Drift Simulation

This project models the drift of a 2-liter bottle half-filled with rice released from the Egyptian Mediterranean coast, comparing different OpenDrift models to predict landfall locations in the northeastern Mediterranean.

Overview

Object: 2-liter plastic bottle, half-filled with rice (~1.5kg total) Release Area: Egyptian Mediterranean coast (Alexandria, Port Said, Damietta, etc.) Target: Northeastern Mediterranean (Greek islands, Turkish coast, Cyprus) Models Compared: OceanDrift, PlastDrift, Leeway

Quick Start

  1. Test Installation:

    python quick_bottle_test.py

    This runs a simple 30-day simulation from Alexandria using synthetic data.

  2. Set Up Real Data (optional but recommended):

    python setup_mediterranean_data.py

    This guides you through setting up Copernicus Marine Service data access for realistic oceanographic conditions.

  3. Run Full Comparison:

    python bottle_drift_mediterranean.py

    This runs comprehensive simulations comparing all models across multiple release points and times.

Files Description

  • quick_bottle_test.py - Simple test simulation with constant environmental conditions
  • setup_mediterranean_data.py - Data setup guide and synthetic data generator
  • bottle_drift_mediterranean.py - Full comparative study with multiple models and scenarios
  • README_bottle_simulation.md - This documentation

Physical Properties

The simulation calculates realistic physical properties for the bottle:

  • Total mass: ~1.55 kg (1.5kg rice + 0.05kg bottle)
  • Buoyancy: Partially submerged, floating upright
  • Wind exposure: ~10% of bottle surface exposed to wind
  • Drag characteristics: Modeled as plastic debris

Release Scenarios

Release Locations (Egyptian Coast)

  • Alexandria (29.92°E, 31.20°N) - Major port city
  • Port Said (32.30°E, 31.27°N) - Suez Canal entrance
  • Damietta (31.82°E, 31.42°N) - Nile Delta coast
  • Marsa Matruh (27.24°E, 31.35°N) - Western coast
  • El Alamein (28.95°E, 30.84°N) - Western desert coast

Optimal Release Times

  • Spring (March): Moderate currents, stable weather
  • Early Summer (June): Predictable conditions, strong thermal circulation
  • Early Autumn (September): Good currents, post-summer stability
  • Winter (December): Strongest circulation, more variable weather

Expected Results

Based on Mediterranean circulation patterns:

Most Likely Trajectories

  1. Eastward along North African coast → Cyprus/Turkish coast (40-60% probability)
  2. Northeast towards Aegean Sea → Greek islands (20-30% probability)
  3. Circulate in central Mediterranean → Italian coast (10-20% probability)

Model Differences

  • PlastDrift: Most suitable for bottles, includes plastic-specific physics
  • OceanDrift: Basic drift, good baseline comparison
  • Leeway: Designed for search & rescue, may overestimate wind effects

Success Factors

  • Release location: Alexandria and Port Said show highest success rates
  • Season: Spring and autumn optimal for reaching northeastern destinations
  • Weather: Avoid winter storms, prefer stable summer conditions

Understanding the Results

Landfall Analysis

The simulation tracks:

  • Percentage of particles reaching land vs. staying at sea
  • Average landfall locations
  • Drift distances and travel times
  • Seasonal variations in success rates

Uncertainty Quantification

Each simulation releases 50-100 particles to represent:

  • Environmental data uncertainty
  • Model parameter uncertainty
  • Natural variability in ocean conditions

Customization Options

Modify Bottle Properties

Edit calculate_bottle_properties() in the main script to adjust:

  • Bottle size and material
  • Fill level and contents
  • Buoyancy characteristics

Change Release Strategy

Modify release parameters:

# Different release pattern
o.seed_elements(
    lon=np.linspace(28, 33, 10),  # Line of releases
    lat=np.linspace(31, 32, 10),
    time=start_time,
    number=20  # particles per release point
)

Environmental Data

  • Use real Copernicus Marine data for accuracy
  • Adjust synthetic data parameters in setup_mediterranean_data.py
  • Add additional environmental factors (waves, temperature)

Troubleshooting

Common Issues

  1. Import errors: Ensure OpenDrift is installed (pip install opendrift)
  2. Data access: Set up Copernicus Marine credentials for real data
  3. Memory issues: Reduce number of particles or simulation duration
  4. Plotting errors: Install matplotlib and cartopy dependencies

Performance Tips

  • Use shorter time steps (3600s) for accuracy vs longer (21600s) for speed
  • Reduce particle count for faster runs
  • Use constant readers for testing, real data for final runs

Scientific Context

Mediterranean Circulation

  • Atlantic Water: Enters at Gibraltar, flows eastward along North Africa
  • Levantine Intermediate Water: Dense water formation in eastern basin
  • Seasonal variations: Stronger circulation in winter, thermal forcing in summer
  • Mesoscale eddies: 50-200 km cyclonic and anticyclonic features

Drift Physics

  • Surface currents: Primary driving force for floating objects
  • Wind drift: Significant for partially submerged objects (1-3% of wind speed)
  • Stokes drift: Wave-induced transport (usually small for large objects)
  • Coastal effects: Stranding, reflection, local circulation

Next Steps

  1. Validate with observations: Compare predictions with known drift cases
  2. Sensitivity analysis: Test different bottle configurations
  3. Real-time simulation: Use current forecasts for time-sensitive releases
  4. Extended modeling: Include weathering, fragmentation, biofouling effects

References