tonalli: an asexual genetic code to characterise APOGEE-2 stellar spectra. I. Validation with synthetic and solar spectra

Lucía Adame,Carlos Román-Zúñiga,Jesús Hernández,Ricardo López-Valdivia,Edilberto Sánchez

DOI: https://doi.org/10.1093/rasti/rzae052

2024-11-23

Abstract:We present tonalli, a spectroscopic analysis python code that efficiently predicts effective temperature, stellar surface gravity, metallicity, $\alpha$-element abundance, and rotational and radial velocities for stars with effective temperatures between 3200 and 6250 K, observed with the Apache Point Observatory Galactic Evolution Experiment 2 (APOGEE-2). tonalli implements an asexual genetic algorithm to optimise the finding of the best comparison between a target spectrum and the continuum-normalised synthetic spectra library from the Model Atmospheres with a Radiative and Convective Scheme (MARCS), which is interpolated in each generation. Using simulated observed spectra and the APOGEE-2 solar spectrum of Vesta, we study the performance, limitations, accuracy and precision of our tool. Finally, a Monte Carlo realisation was implemented to estimate the uncertainties of each derived stellar parameter. The ad hoc continuum-normalised library is publicly available on Zenodo (DOI <a class="link-https link-external" data-doi="10.5281/zenodo.12736546" href="https://doi.org/10.5281/zenodo.12736546" rel="external noopener nofollow">https://doi.org/10.5281/zenodo.12736546</a>).

Instrumentation and Methods for Astrophysics,Astrophysics of Galaxies,Solar and Stellar Astrophysics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to extract the basic parameters of stars (such as effective temperature, surface gravity, metallicity, α - element abundance, as well as rotational and radial velocities) from the near - infrared spectra of APOGEE - 2 efficiently and accurately, especially for pre - main sequence stars.

Specifically, the paper proposes a Python code tool named tonalli, which uses the asexual genetic algorithm (AGA) to optimize the best match between the observed spectra and the synthetic spectral library. This tool is designed to handle stars with effective temperatures between 3200 and 6250 K, and verifies its performance, limitations, accuracy, and precision by simulating the observed spectra and the APOGEE - 2 solar spectra. In addition, the uncertainty of each derived stellar parameter is estimated through Monte Carlo simulations.

### Key Problem Summary:

1. **Accurate Extraction of Stellar Parameters**:
- Extracting reliable basic stellar parameters from large - scale spectral survey data is a complex and important task.
- In particular, the study of pre - main sequence stars is challenging because these stars evolve rapidly and their spectra are affected by multiple factors (such as accretion disks, magnetic fields, etc.).

2. **Limitations of Existing Methods**:
- Existing atmospheric models have large differences for certain types of stars (such as early - type or late - type stars).
- The main pipeline of APOGEE - 2, ASPCAP, is mainly used for classifying red giants and is not suitable for classifying pre - main sequence stars.

3. **Need for New Methods**:
- A new and efficient algorithm is required to handle the spectra of different types of stars, especially pre - main sequence stars.
- tonalli achieves this goal through the asexual genetic algorithm, can fit multiple parameters simultaneously, and provides reliable parameter estimates and their uncertainties.

### Solutions:

- **Asexual Genetic Algorithm (AGA)**: Used to optimize the best match between the observed spectra and the synthetic spectral library, gradually narrowing the parameter space to find the optimal solution.
- **Synthetic Spectral Library**: Interpolates using the MARCS model atmosphere library to generate the synthetic spectra closest to the observed spectra.
- **Performance Verification**: Verifies the performance and precision of the tool by simulating the observed spectra and the APOGEE - 2 solar spectra.

### Formula Presentation:

The key formulas involved in the paper include:

1. **Chi - square Statistic (𝜒²)**:
\[
\chi^2=\sum_{\lambda}\frac{(F_{\lambda,o}-F_{\lambda,s})^2}{(\sigma_{\lambda,o})^2}
\]
where \( F_{\lambda,o} \) is the monochromatic flux of the observed spectrum, \( F_{\lambda,s} \) is the monochromatic flux of the synthetic spectrum, and \( \sigma_{\lambda,o} \) is the observation error.

2. **Decrease in Search Range**:
\[
\Delta x_{i,n}=\Delta x_{i,0}\cdot p^n
\]
where \( \Delta x_{i,0} \) is the initial search range, \( p \) is the convergence factor, and \( n \) is the number of generations.

Through these methods and formulas, tonalli can provide reliable parameter estimates in a wide range of star types, especially suitable for the study of pre - main sequence stars.