[Paper Review] High Resolution Irradiance Spectrum from 300 to 1000 nm

Robert L. Kurucz|arXiv (Cornell University)|May 1, 2006

Atmospheric Ozone and ClimateEarth and Planetary Sciences3 references20 citations

TL;DR

This paper presents a high-resolution solar irradiance spectrum from 300 to 1000 nm, reprocessed from Kitt Peak FTS scans using improved atmospheric modeling, telluric line correction via HITRAN data, and iterative continuum fitting. The key contribution is a revised, artifact-reduced spectrum with monochromatic errors of 0.1–1.0%, recommended as a high-resolution reference for solar system and exoplanet studies.

ABSTRACT

The FTS scans that made up the Kitt Peak Solar Flux Atlas by Kurucz, Furenlid, Brault, and Testerman (1984) have been re-reduced. An approximate telluric atmospheric model was determined for each FTS scan. Large-scale features produced by O3 and O2 dimer were computed and divided out. The solar continuum level was found by fitting a smooth curve to high points in each scan. The scans were normalized to the fitted continuum to produce a residual flux spectrum for each FTS scan. The telluric line spectrum was computed using HITRAN and other line data for H2O, O2, and CO2. The line parameters were adjusted for an approximate match to the observed spectra. The scans were divided by the computed telluric spectra to produce residual irradiance spectra. Artifacts from wavelength mismatches, deep lines, etc, were removed by hand and replaced by linear interpolation. Overlapping scans were fitted together to make a continuous spectrum from 300 to 1000 nm. All the above steps were iterative. The monochromatic error varies from 0.1 to 1.0 percent. The residual spectrum was calibrated two different ways: First by normalizing it to the continuum of theoretical solar model ASUN (Kurucz 1992), and second, by degrading the spectrum to the resolution of the observed irradiance (Thuillier et al. 2004) to determine a normalization function that was then applied to the high resolution spectrum.

Motivation & Objective

• To produce a high-resolution solar irradiance spectrum from 300 to 1000 nm using reprocessed FTS scans from the Kitt Peak Solar Flux Atlas.
• To correct for atmospheric absorption features (O₃, O₂ dimer) and telluric lines (H₂O, O₂, CO₂) using updated line data and atmospheric models.
• To improve spectral continuity and accuracy by iteratively aligning overlapping scans and removing artifacts from wavelength mismatches and deep lines.
• To provide a reliable, high-resolution reference spectrum for solar system and exoplanet research, particularly for modeling irradiance on extra-solar planets.
• To secure funding for extending the spectrum to 5400 nm and improving the visible range with new line data or better spectra.

Proposed method

• Re-reduced original FTS scans from the Kitt Peak Solar Flux Atlas using a 3-point Gaussian smoothing to simplify continuum placement.
• Assigned each scan to an approximate atmospheric model and computed O₃ and [O₂]₂ transmission using models available at kurucz.harvard.edu, then divided these out.
• Computed telluric line transmission using HITRAN and other line data, adjusting parameters to match observed spectra through iterative fitting.
• Redetermined the wavelength scale and applied gravitational redshift correction to align scans in the solar laboratory frame in vacuum.
• Fitted a smooth continuum to high points in the spectrum, divided the observed spectrum by this continuum to produce a residual flux spectrum.
• Divided residual spectra by the computed telluric transmission to derive residual irradiance spectra, with artifacts removed by hand and replaced via linear interpolation.

Experimental results

Research questions

• RQ1 How can the accuracy and resolution of the Kitt Peak Solar Flux Atlas be improved by reprocessing FTS scans with modern atmospheric and line data?
• RQ2 What is the impact of uncorrected telluric absorption and atmospheric features (O₃, [O₂]₂) on high-resolution solar irradiance measurements?
• RQ3 To what extent can iterative continuum fitting and spectral alignment reduce artifacts and improve spectral continuity across overlapping scans?
• RQ4 How does the revised high-resolution irradiance spectrum compare with existing low-resolution standards like Thuillier et al. (2004) in key spectral regions?
• RQ5 What are the limitations of current solar models (e.g., ASUN) in reproducing observed flux levels, particularly in the G band and below 300 nm?

Key findings

• The reprocessed spectrum covers 300–1000 nm with monochromatic errors ranging from 0.1% to 1.0%.
• Artifacts from wavelength mismatches and deep telluric lines were manually removed and replaced with linear interpolation, significantly improving spectral fidelity.
• The revised spectrum shows a probable overestimation of ozone absorption below 320 nm and around 600 nm, suggesting potential errors in the original atlas or model assumptions.
• Discrepancies in the G band indicate that the ASUN solar model underproduces flux, likely due to insufficient opacity below 300 nm and an incorrect temperature gradient.
• The final high-resolution irradiance spectrum was normalized to the Thuillier et al. (2004) reference spectrum, which the author recommends as the best available high-resolution standard.
• The author estimates the project cost at $161,000 for a CD and plots, and requests $300,000/year in grants to extend the spectrum to 5400 nm and improve the visible range with updated data.