Space Weather¶

The astrojax.space_weather module provides solar and geomagnetic activity data for atmospheric density models. Space weather indices drive the NRLMSISE-00 model's predictions of thermospheric temperature and density, which directly affect atmospheric drag on LEO satellites.

Why Space Weather Matters¶

Solar EUV radiation heats the thermosphere, causing it to expand and increasing drag on satellites. Geomagnetic storms driven by solar wind further modulate upper-atmosphere density. Two families of indices capture these effects:

The NRLMSISE-00 model requires F10.7, the 81-day average F10.7a, and a structured 7-element Ap array describing recent geomagnetic history.

JIT-Compatible Design¶

Like the EOP module, astrojax stores space weather data as sorted JAX arrays in a NamedTuple. Lookups use jnp.searchsorted for O(log n) binary search, making everything compatible with jax.jit, jax.vmap, and jax.grad.

The SpaceWeatherData type is a JAX pytree automatically, so it can be passed through JIT boundaries and used with jax.lax control flow.

Loading Space Weather Data¶

Bundled Data (Recommended)¶

The simplest approach loads the bundled sw19571001.txt file:

from astrojax.space_weather import load_default_sw, get_sw_f107_obs

sw = load_default_sw()
f107 = get_sw_f107_obs(sw, 59569.0)

Custom Data File¶

Load from any CSSI-format space weather file:

from astrojax.space_weather import load_sw_from_file

sw = load_sw_from_file("/path/to/sw19571001.txt")

Cached Data with Auto-Refresh¶

For production use, load_cached_sw keeps a local copy and automatically downloads a fresh version from CelesTrak when the cached copy is older than a configurable threshold (default: 7 days):

from astrojax.space_weather import load_cached_sw, get_sw_f107_obs

# Uses default cache location (~/.cache/astrojax/space_weather/) and 7-day refresh
sw = load_cached_sw()
val = get_sw_f107_obs(sw, 59569.0)

# Custom cache path and 1-day refresh
sw = load_cached_sw("/tmp/my_sw/sw19571001.txt", max_age_days=1.0)

If the download fails (network unavailable, CelesTrak server down, etc.), the function falls back to the bundled data so it never raises on network issues.

The cache location defaults to ~/.cache/astrojax/space_weather/ and can be overridden with the ASTROJAX_CACHE environment variable.

Constant / Zero Space Weather¶

For testing or when specific constant values are known:

from astrojax.space_weather import static_space_weather, zero_space_weather

# Fixed known values
sw = static_space_weather(f107=150.0, ap=4.0, f107a=150.0)

# All-zero values
sw = zero_space_weather()

Querying Values¶

All query functions take a SpaceWeatherData instance and an MJD:

from astrojax.space_weather import (
    get_sw_f107_obs,
    get_sw_f107_adj,
    get_sw_f107_obs_ctr81,
    get_sw_f107_obs_lst81,
    get_sw_ap,
    get_sw_ap_daily,
    get_sw_kp,
    get_sw_ap_array,
)

# Solar flux
f107 = get_sw_f107_obs(sw, mjd)          # Observed F10.7 [sfu]
f107_adj = get_sw_f107_adj(sw, mjd)      # Adjusted F10.7 [sfu]
f107a = get_sw_f107_obs_ctr81(sw, mjd)   # 81-day centered average
f107a_lst = get_sw_f107_obs_lst81(sw, mjd)  # 81-day last average

# Geomagnetic indices
kp = get_sw_kp(sw, mjd)                  # 3-hourly Kp (0-9)
ap = get_sw_ap(sw, mjd)                  # 3-hourly Ap
ap_daily = get_sw_ap_daily(sw, mjd)      # Daily average Ap

# NRLMSISE-00 structured Ap array (7 elements)
ap_array = get_sw_ap_array(sw, mjd)

The NRLMSISE-00 Ap Array¶

The get_sw_ap_array function builds the 7-element magnetic activity array required by NRLMSISE-00:

This structured array captures the time history of geomagnetic activity that the NRLMSISE-00 model uses for thermospheric heating response.

Using with JIT and vmap¶

All query functions work inside JAX transformations:

import jax
import jax.numpy as jnp

# JIT compilation
jit_query = jax.jit(get_sw_f107_obs)
val = jit_query(sw, 59569.0)

# Vectorized over MJD batch
mjd_batch = jnp.array([59569.0, 59570.0, 59571.0])
vals = jax.vmap(lambda m: get_sw_f107_obs(sw, m))(mjd_batch)