Skip to content

Spectroscopic lookup tables (LUTs)

JURASSIC uses precomputed spectroscopic lookup tables (LUTs) to represent infrared absorption and emission efficiently. These tables are a central component of the model and enable fast radiative transfer calculations without performing line-by-line spectroscopy at runtime.

This page describes the role, structure, and configuration of LUTs in JURASSIC.

Purpose of lookup tables

Infrared radiative transfer depends on highly structured molecular spectra with millions of individual absorption lines. Evaluating these lines explicitly is computationally expensive and impractical for large-scale applications.

JURASSIC avoids this cost by using lookup tables that:

  • are generated offline using detailed line-by-line models,
  • store band-averaged emissivities or transmittances,
  • are interpolated efficiently during runtime.

This approach preserves high spectroscopic fidelity while enabling orders-of-magnitude speedups compared to line-by-line calculations.

Spectral approximations

The lookup tables are designed to support the spectral approximations used by JURASSIC, in particular:

  • the Emissivity Growth Approximation (EGA),
  • the Curtis–Godson Approximation (CGA).

The tables provide emissivity values for homogeneous atmospheric layers as functions of pressure, temperature, and absorber amount. These values are then combined along an inhomogeneous ray path using EGA.

Details of the underlying theory are provided in the Theory section.

Table dimensions and variables

Each lookup table is defined on a multi-dimensional grid. Typical dimensions include:

  • pressure
  • temperature
  • absorber column density (or equivalent quantity)

Separate tables are usually provided for each:

  • molecular species (emitter),
  • spectral window or band.

The exact grid definition depends on how the tables were generated and must be compatible with the atmospheric conditions encountered during runtime.

Table organization and naming

Lookup tables are located using the control parameter:

  • TBLBASE

TBLBASE defines a base path or prefix from which the model derives the full table file names.

A typical setup consists of:

  • one directory per spectral window,
  • one table per gas/emitter within each window.

The exact naming convention is flexible but must match the expectations of the table-reading routines and the selected table format.

Table formats

JURASSIC supports multiple lookup table formats. The format is selected via the control keyword:

  • TBLFMT

Common formats include:

  • ASCII tables (human-readable, larger, slower to read),
  • binary tables (compact, fast I/O, preferred for production use).

The example projects included with JURASSIC demonstrate working table layouts and formats and are the recommended starting point for new users.

Runtime interpolation

During execution, JURASSIC performs fast interpolation within the lookup tables to obtain emissivities for the local atmospheric state.

Interpolation is typically:

  • linear in pressure, temperature, and absorber amount,
  • performed independently for each gas and spectral window.

The interpolation routines are optimized for performance and are called frequently during ray-path integration.

Validity and extrapolation

Lookup tables are valid only within the parameter ranges for which they were generated. If runtime atmospheric conditions fall outside these ranges, interpolation may result in:

  • extrapolated values,
  • reduced accuracy,
  • or, in extreme cases, invalid results.

Best practice:
Ensure that lookup tables cover the full range of pressure, temperature, and gas concentrations expected in your application.

Generating lookup tables

Lookup tables are generated outside the core JURASSIC workflow, typically using a line-by-line radiative transfer model and a chosen spectroscopic database.

While JURASSIC is agnostic to the specific tool used for table generation, commonly used approaches include:

  • line-by-line reference models,
  • consistent spectroscopic databases (e.g. HITRAN-based).

Table generation workflows are often highly application-specific and are therefore not part of the core JURASSIC distribution.

Performance considerations

  • Binary lookup tables significantly reduce I/O overhead.
  • Keeping tables on fast local or parallel file systems improves performance for large runs.
  • Table reuse across many simulations amortizes the offline generation cost.

For HPC workflows, careful table management is essential for achieving good scalability.

Summary

Spectroscopic lookup tables are a fundamental building block of JURASSIC. They enable fast infrared radiative transfer simulations by decoupling expensive spectroscopic calculations from runtime execution.

Correct configuration and careful generation of lookup tables are crucial for both accuracy and performance.