Aethera Logo

Methodology & Data Sources

At Aethera, we strive to make weather forecasts transparent, scientifically sound, and highly accurate. This page details the models, post-processing calculations, and data pipelines used to power our forecasts, trekking alerts, and air quality indexes.

1. Numerical Weather Prediction (NWP) Models

We do not rely on a single forecast model. Instead, we query, aggregate, and average data from leading global and regional meteorological organizations to compute the most likely local forecast:

  • ECMWF IFS (European Centre for Medium-Range Weather Forecasts): Considered the gold standard for global medium-range forecasting. It runs at a native horizontal resolution of ~9 km, providing unparalleled precision for temperature and rainfall trends.
  • NOAA GFS (Global Forecast System): The primary global model of the United States. Runs at ~22 km resolution, offering high-fidelity wind tracking and pressure readings.
  • DWD ICON (German Meteorological Service): A non-hydrostatic global model with a grid resolution of ~13 km. ICON is highly effective for simulating atmospheric circulation in rugged terrains like the Himalayas.
  • MET Norway: Excellent high-resolution forecasts particularly calibrated for high-latitude and mountainous regions.

2. Mountain Terrain Elevation Correction

Nepal's geography is characterized by extreme altitude changes (e.g. from the lowlands of Terai to Everest at 8,848m). Standard global models smooth out these mountains into a coarse grid average. To prevent inaccurate temperature readings at high altitudes, Aethera applies a real-time **terrain-correction algorithm**:

  • Elevation Mapping: We retrieve the precise elevation of your queried coordinates using high-resolution Shuttle Radar Topography Mission (SRTM) topography datasets.
  • Lapse Rate Adjustments: If the model's smoothed grid altitude differs from the actual physical altitude, we adjust the temperature using the environmental lapse rate. For dry air, temperature decreases by approximately 9.8°C per 1,000 meters of elevation gain; for humid air, it decreases by approximately 6.5°C per 1,000 meters.
  • Wind Shear Modeling: Mountain pass winds are dynamically scaled using a logarithmic wind profile model, adjusting for frictional drag at different elevations.

3. Air Quality Index (AQI) Calculation

Our air quality tracking measures the US Air Quality Index (AQI) standard on a scale of 0 to 500. The index is calculated based on the highest individual pollutant value from the following sensors:

  • Fine Particulates (PM2.5): The most critical health indicator, tracking microscopic dust and soot.
  • Coarse Particulates (PM10): Tracks larger dust and smoke particles.
  • Nitrogen Dioxide (NO2): Emitted primarily by traffic and power plants.
  • Ozone (O3): Ground-level ozone formed by reactions of solar radiation and pollutants.

Data is retrieved from the **Copernicus Atmosphere Monitoring Service (CAMS)** and the **Open-Meteo Air Quality API**, ensuring hourly updates.

4. Data Cache & Update Cadence

  • Current Weather & AQI: Refreshed every 60 minutes to maintain live-tracking integrity.
  • 14-Day & Hourly Forecasts: Recalculated 4 times daily (every 6 hours) immediately following new NWP model runs.

For research queries or raw data access, contact our systems engineering desk at data@omprasadghimire.com.np.