Skip to content

canvod-auxiliary

Purpose

The canvod-auxiliary package provides auxiliary data management for GNSS Transmissometry (GNSS-T) analysis. It handles downloading, parsing, preprocessing, and interpolating SP3 ephemerides and CLK clock corrections to augment RINEX observation data with precise satellite positions and timing information.

The Dimension Alignment Problem

GNSS-T analysis requires combining two data sources with fundamentally different indexing:

RINEX Observations

  • High temporal resolution: 30 s, 15 s, or sub-second
  • Signal-level indexing by SID: "G01|L1|C"
  • Dimensions: (epoch: 2880, sid: 384)

SP3 / CLK Auxiliary

  • Low temporal resolution: 5–15 min
  • Satellite-level indexing by SV: "G01"
  • Dimensions: (epoch: 96, sv: 32)

Combining these requires:

  1. Dimension conversion — sv (32 satellites) → sid (384 signal IDs)
  2. Temporal interpolation — 15 min → 30 s sampling
  3. Coordinate transform — ECEF → geodetic → spherical (r, θ, φ)
  4. Physically correct interpolation per data type
graph TD
    A["`**SP3 File**
    sv dimension`"]
    A -->|Download| B["`**Raw Dataset**
    96 epochs, 32 svs`"]
    B -->|Preprocess| C["`**Preprocessed**
    96 epochs, 384 sids`"]
    C -->|Interpolate| D["`**Interpolated**
    2880 epochs, 384 sids`"]
    D -->|Augment| E["`**RINEX Dataset**
    2880 epochs, 384 sids`"]

Interpolation Strategies

Different data types require different methods based on the underlying physics:

  •   SP3 — Hermite Cubic Splines

    Satellite orbital motion is smooth and predictable. Hermite splines exploit available velocity data for C1-continuity and sub-millimetre interpolation accuracy.

    config = Sp3Config(use_velocities=True, fallback="linear")
strategy = Sp3InterpolationStrategy(config=config)

  •   CLK — Piecewise Linear

    Clock corrections may have discontinuities from satellite maneuvers and uploads. No derivative information is available. Piecewise linear interpolation handles these jumps safely.

    config = ClkConfig(window_size=9, jump_threshold=1e-6)
strategy = ClockInterpolationStrategy(config=config)

Key Components

  •   File Handlers

    Sp3File — SP3a/c/d format: positions (X, Y, Z) + velocities (VX, VY, VZ)

    ClkFile — RINEX clock format: satellite clock biases

    ProductSpec — Declarative config for URLs, latency, and auth

  •   Preprocessing

    prep_aux_ds() — Full 4-step pipeline: sv→sid mapping, pad to global SID, normalise dtype, strip _FillValue

    sp3_sid = prep_aux_ds(sp3_raw)
# {'epoch': 96, 'sid': 384}

  •   Coordinate Classes

    ECEFPosition — Earth-Centred Earth-Fixed (X, Y, Z)

    GeodeticPosition — WGS84 latitude, longitude, altitude

    Spherical coordinates (r, θ, φ) relative to receiver position

  •   FTP Download

    Primary: ESA/CODE FTP server

    Fallback: NASA CDDIS (requires Earthdata account)

    Per-DOY caching prevents redundant downloads

Usage

from canvod.auxiliary import (
    Sp3File, prep_aux_ds,
    Sp3InterpolationStrategy, Sp3Config,
    compute_spherical_coordinates, ECEFPosition,
    add_spherical_coords_to_dataset,
)

rinex_ds = Rnxv3Obs("station.25o").to_ds()
target_epochs = rinex_ds.epoch.values

# Download and parse
sp3_raw = Sp3File.from_url(date, "CODE", "final").to_dataset()

# Preprocess: sv → sid
sp3_sid = prep_aux_ds(sp3_raw)

# Interpolate to obs epochs
interp = Sp3InterpolationStrategy(config=Sp3Config(use_velocities=True))
sp3_at_obs = interp.interpolate(sp3_sid, target_epochs)

# Compute geometry
rx_pos = ECEFPosition.from_ds_metadata(rinex_ds)
r, theta, phi = compute_spherical_coordinates(
    sp3_at_obs["X"], sp3_at_obs["Y"], sp3_at_obs["Z"], rx_pos
)

# Augment dataset
augmented = add_spherical_coords_to_dataset(rinex_ds, r, theta, phi)

sequenceDiagram
    participant User
    participant Sp3File
    participant Preprocessor
    participant Interpolator
    participant Coordinates

    User->>Sp3File: from_url(date, agency, product_type)
    Sp3File->>Sp3File: FTP download + parse
    Sp3File-->>User: to_dataset() {epoch: 96, sv: 32}

    User->>Preprocessor: prep_aux_ds(sp3_raw)
    Preprocessor-->>User: {epoch: 96, sid: 384}

    User->>Interpolator: interpolate(sp3_sid, target_epochs)
    Interpolator->>Interpolator: Hermite splines
    Interpolator-->>User: {epoch: 2880, sid: 384}

    User->>Coordinates: compute_spherical_coordinates()
    Coordinates-->>User: (r, θ, φ) arrays

    User->>User: add_spherical_coords_to_dataset()