Hugging Face's logo

Spaces:
nakas
/
ecmwf_wind_docker
Sleeping

App Files Community
ecmwf_wind_docker / README.md
nakas's picture
nakas
Fix README.md with proper Hugging Face Spaces configuration
7d033ba
preview code
|
raw
history blame contribute delete
2.28 kB
metadata
title: ECMWF Wind Particle Visualization
emoji: 🌪️
colorFrom: blue
colorTo: purple
sdk: docker
sdk_version: 4.44.0
app_file: app.py
pinned: false

🌪️ ECMWF Wind Particle Visualization

A Gradio application that downloads real ECMWF 10m wind data and creates windy-style particle animations using Folium maps with Leaflet-Velocity plugin.

Features

  • 🌍 Real ECMWF Data: Downloads operational 10m wind forecasts (U/V components)
  • 🎨 Windy-style Animation: Canvas-based particle system using Leaflet-Velocity
  • ⏱️ Time Controls: TimestampedGeoJson animation with interactive time slider
  • 🗺️ Interactive Maps: Multiple tile layers and regional views
  • 🎯 Particle Physics: RK4 integration for accurate wind field advection

How It Works

Data Processing

  1. Downloads ECMWF 10m wind data (U/V components) in GRIB2 format
  2. Converts GRIB to grib2json-compatible velocity JSON format
  3. Implements bilinear interpolation for smooth wind field sampling

Particle Animation

  1. Canvas Rendering: Uses Leaflet-Velocity for real-time particle advection
  2. TimestampedGeoJSON: Pre-computed particle trajectories with time controls
  3. RK4 Integration: Accurate particle movement through wind field
  4. Color Coding: Particles colored by wind speed magnitude

Technical Details

Data Source

  • ECMWF Open Data: Free operational forecasts updated every 6 hours
  • Parameters: 10m U-wind and V-wind components
  • Resolution: 0.25° (~25km spacing)
  • Format: GRIB2 converted to velocity JSON

Algorithms

  • Bilinear Interpolation: Smooth wind field sampling between grid points
  • RK4 Integration: 4th-order Runge-Kutta for particle advection
  • Coordinate Conversion: Proper handling of geographic projections

Libraries

  • Gradio: Web interface framework
  • Folium: Python-Leaflet bridge for interactive maps
  • Leaflet-Velocity: Canvas-based particle rendering
  • xarray/cfgrib: GRIB data processing
  • ecmwf-opendata: ECMWF data access

Attribution

Weather data provided by ECMWF under their Open Data initiative: