Exoplanet Spectroscopy

Sheneman, A. C., Langeveld, A. B., & Bartholomew, A. (2024) “Assessing the Accuracy of Telluric Corrections.” Honors Program Thesis. Paper. http://hdl.handle.net/20.500.12648/15024Sheneman, A. C. & Langeveld, A. B. (2024). “How significantly does Earth's atmosphere affect our ability to investigate the atmospheric spectra of exoplanets?” 243rd American Astronomical Society Conference. Poster. https://aas243-aas.ipostersessions.com/Default.aspx?s=8D-42-F7-C0-67-F9-04-FF-A5-53-FB-BF-21-58-70-6CSheneman, A. C., Langeveld, A. B., & Bartholomew, A. (2024). “How significantly does Earth's atmosphere affect our ability to investigate the atmospheric spectra of exoplanets?” Astrobiology Graduate Conference (AbGradCon24) at Cornell University. Poster. 10.13140/RG.2.2.13155.16167. Sheneman, A. C. & Langeveld, A. B. (2023). “How significantly does Earth's atmosphere affect our ability to investigate the atmospheric spectra of exoplanets?” Summer Research Symposium at Cornell University. Poster.

PAPER

AAS POSTER - ASTRONOMY ACHIEVEMENT AWARD

Overview

Cornell University Research in Astrophysics and Planetary Science REU ('23)New Paltz University Physics Senior Project & Honors Thesis Research ('23-'24)Mentor: Dr. Adam Langeveld (P.I. Ray Jayawardhana), Cornell University & Johns Hopkins UniversityProf. Amy Bartholomew, Department Chair of Physics & Astronomy at New Paltz UniversityProf. Patricia Sullivan, Director of the Honors Program at New Paltz UniversitySummary: Developed new statistical assessment methods for telluric contamination corrections to improve the accuracy of exoplanet atmospheric spectral analysis. Results: Determined an ideal airmass range for observation proposals of exoplanet atmospheres.

Paper

(read full paper here)

Abstract

Observing transiting exoplanets with ground-based telescopes and high-resolution spectrographs enables the resolution of individual absorption lines in the exoplanet transmission spectra. However, observing from the ground inherently introduces telluric contamination: spectral contamination from absorption due to molecules in the Earth’s atmosphere. We take high-resolution observations of a transiting exoplanet around a bright A-type star as a case study and use synthetic telluric molecfit models to remove contamination from water and oxygen molecules in Earth’s atmosphere. The quality of the telluric corrections was statistically assessed for several different telluric regions based on absorption depth and molecular absorption species. We find that corrections for shallow telluric lines are more robust than deeper telluric lines, though both depend similarly on airmass. Corrections for different molecular bands varied by region. Some regions demonstrate a higher dependency on airmass, potentially due to the wavelength, depth, or quantity of telluric lines. Finally, we determine that the most accurate corrections are performed at observations with airmass under 1.07 corresponding to a zenith angle of approximately 20.84 degrees. Whilst this is a somewhat limited airmass range, these results highlight the need for improving telluric models for future searches of water and oxygen features in Earth-like exoplanet transmission spectra with 30m-class telescopes. These results may assist in optimizing observations for retrieving and preserving more data in an exoplanet transmission spectrum, especially when absorption features from the same molecules–and potential biosignatures–in Earth-like atmospheres fall in these highly contaminated regions.

Methods

The Transit Method

Decreases in the total light received from the host star due to the exoplanet orbital path. Absorption due to the molecules comprising the atmosphere of the exoplanet is visible during transit (Cowley et al., 2014)

What is Airmass?

Airmass is generally the amount of atmosphere an observation is made through. All airmass values are found using the equation in the figure, where an ideal airmass of 1 is found directly at zenith (AM = 1 cos(0) = 1).

Atmospheric Parameters

Extracted atmospheric parameters and observation related data from FITS file headers. These variables included temperature, pressure, humidity, airmass, and the signal to noise ratio.

Isolating Peak Regions and Defining a Data Set

Performed molecfit corrections (dividing out a molecular model of the Earth's atmosphere from the observations to retrieve a corrected flux). Then, individual telluric lines and groups of telluric lines were isolated based on absorption depth ranges as well as wavelength regions.

Statistically Assessing Quality

Using the Reduced Chi-Square and Residual Sum of Squares statistical assessment methods, we could compare the accuracy of the corrections done based on the peaks previously isolated. These also allowed us to compare correction quality and atmospheric parameters.

Reduced Chi-Square

Test The sum of the squared differences between the corrected flux (F) and the continuum (C) divided by the variance (sigma)/

Propagation of Uncertainty

Performed on the Reduced-Chi Square test.

Residual Sum of Squares

The sum of the squared difference between the observed (y) and predicted (x) data, where the function that represents x (f(x)) is the continuum.

Results

Reduced Chi-Square Results

The statistical assessment is performed over all wavelength regions and telluric depths. χ 2 ν values normalized based on lowest airmass file for plots. Ideal corrections = 1.

Residual Sum of Squares Results

The statistical assessment is performed over all wavelength regions and telluric depths. RSS values normalized based on lowest airmass file for plots. Ideal corrections = 0.

Both assessments show that the greatest dependency is on the airmass of the observations.

Let's dig deeper:

Telluric Depth by Wavelength Region

χ 2 ν (first and third rows) and Divided RSS (second and fourth rows) for different regions of the H2O band (columns). The wavelength increases for each region from the smallest (left) to the largest (right). One potential cause for this may be the quantity of lines in each region at each depth. Preliminary results shown below seem to demonstrate that this is likely a contributing factor.This suggests that for regions with many telluric lines and/or deeper telluric lines, airmass prioritization in observation is increasingly important, whereas in regions with fewer and/or more shallow telluric lines, there seems to be a greater effect from instrumentation or the Signal to Noise Ratio (SNR).This may also be attributed to distinct wavelength regions based on heightened sensitivity within those regions, though this is likely explained by the quantity and depth of tellurics as well.