Weather Data
Spacetime uses weather data when estimating link attenuation between any candidate pair of compatible transceivers as part of its continuous forward analysis of candidate links and their utility. The weather data is input to several ITU-R link attenuation models, each of which has specific requirements for the types of data necessary to perform its attenuation calculations.
Sources
Spacetime pulls weather data on-demand, caching for subsequent access to the same meteorological variables. It can pull from any service that generates GRIB2 files (and indexes thereof), including NOAA (GFS) and ECMWF.
Weather-based Attenuation Models
Several ITU-R attenuation models are available, and when paired with live or forecast weather data are used in estimating link attenuation. A survey of attenuation models and input data they require follows.
ITU-R P.676
This “Attenuation by atmospheric gases and related effects” model needs at least the following data types:
- Pressure (Pascals)
- Temperature (Kelvin)
- Water vapor pressure (Pascals)
at several heights above local orography, distributed across all latitudes and longitudes through which a link may pass.
ITU-R P.838
The “Specific attenuation model for rain for use in prediction methods” model needs at least the following data types:
- Rain height (meters)
- Rain rate (meters/second or equivalent like mm/hour)
The local rain rate is the key datum, but a surface level rain rate can be combined with an approximate rain height to estimate the volume of space impacted. Data for all latitudes and longitudes through which a link may pass should be supplied.
ITU-R P.840
The “Attenuation due to clouds and fog” model needs at least the following data types:
- Cloud ceiling (meters)
- Cloud layer thickness (meters)
- Cloud liquid water density (grams/meter3)
- Cloud temperature (Kelvin)
distributed across all latitudes and longitudes through which a link may pass. This suffices for a single layer of clouds; multiple layers of clouds would require the same data for each layer, also distributed across all latitudes and longitudes of interest.
ITU-R P.1814
From “Prediction methods required for the design of terrestrial free-space optical links”, Spacetime estimates scintillation effects. Specifically, it uses the Hufnagel-Valley (H-V) model to estimate the refractive index structure parameter, derived from wind speed:
- East and North components of wind (meters/second)
Weather data structure
Weather forecast products from ECMWF and NOAA have many data products useful for link attenuation modeling. Their delivery format may be broadly categorized as:
- layered (3D, i.e. values vary by height) and
- flat (values vary only by geographic location and do not vary with height)
ECMWF
A summary of ECMWF weather variables which can form a useful input data set:
| Quantity | Layered / Flat | ITU-R Models | ECMWF Variable | Notes |
|---|---|---|---|---|
| pressure | layered | P.676 | n/a | all layered data is keyed by pressure levels; no explicit variable required |
| height | layered | P.676, P.840 | Z | requires additional conversion from geopotential to geometric height in meters above the WGS84 ellipsoid |
| temperature | layered | P.676, P.840 | T | cloud temperature modeled as approximately equal to atmospheric temperature |
| water vapor pressure | layered | P.676 | Q | converted via epsilon (gas constant for dry air / gas constant for water vapor) |
| rain height | flat | P.838 | deg0l | modeled as equal to the height of the zero-degree isotherm |
| rain rate | flat | P.838 | TP | |
| cloud ceiling | flat | P.840 | CC | source variable is layered; derived by walking from the ground layer up, inspecting the fractional cloud cover and noting when it crosses a density threshold (currently 50%) |
| cloud layer thickness | flat | P.840 | CC | source variable is layered; delta between the bottom (ceiling, above) and the top of an assumed single layer of clouds (i.e. assuming no open air gaps between multiple layers of clouds). It is derived by walking down from the highest altitude, inspecting the fractional cloud cover (“CC”) and noting when it crosses a density threshold (currently 50%) |
| Cloud liquid water density | layered in principle; flat | P.840 | TCLW | convert total cloud liquid water to density by dividing by cloud layer thickness (assuming one layer of clouds) |
| U component of wind | layered | P.1814 | U | eastward component of the wind |
| V component of wind | layered | P.1814 | V | northward component of the wind |
NOAA
NOAA products have similar data in a variety of formats. Many products are limited to U.S. or North American territories.
Of interest for a future ITU-R P.531 implementation might be the Global Total Electron Content data product.