Source: School of Physics and Astronomy
Written by: School of Physics and Astronomy
Edited by: Wang Dongmei

In this May, a paper “Non-parametric dark energy reconstruction using the to-mographic Alcock-Paczynski test”, led by undergraduates from the School of Physics and Astronomy, Sun Yat-sen University, was accepted for publication in The Astrophysical Journal (hereafter ApJ). The first three authors, Zhenyu Zhang, Gan Gu and Xiaoma Wang, are all grade 2016 undergraduates. The corresponding author is Xiaodong Li, an Associate Professor from the same School.

The undergraduates have spent tremendous effect on the project. Firstly, they studied the ‘tomographic Alcock-Paczynaski method’ (hereafter ‘tomographic AP’), and then utilized it to constrain dark energy parameters. Tomographic AP is a method for analyzing the cosmic large scale structure data. Once the survey data is ready, a specific set of cosmological parameters (e.g. dark matter ratio ?_m, dark energy equation of state w) have to be adopted, in order to reconstruct the 3-D distribution of matter. In case that the adopted parameters are different from the underlying true parameters of the Universe, one would find anomaly in the results. The anomaly manifests itself as ‘reshift-dependent anisotropic in the radial and angular directions’. Tomographic AP method utilizes this effect to constrain cosmology, based on the principle that the parameters which lead to minimal anomaly is mostly likely to be the truth. In 2016, scientists obtained tightest constraint on dark energy equation of state w by applying this method to the SDSS data.

To achieve reconstruction of w in a most model-independent manner, the students also used the ‘non-parametric method’. Generally speaking, when we regress or fit some physical quantity, we have to assume a particular form (e.g., liner, polynomial, exponential, ...) for it. These priori assumptions make the fitting model-dependent, i.e. limited by the specific model. Using the ‘non-parametric method’ we can overcome this limitation and achieve a model-independent reconstruction. In 2017, a group of researches applied this method to the dark energy reconstruction, and their work got published in Nature Astronomy (Gongbo Zhao et al., 2017).

Under the supervision of Professor Li, Zhang et al combined ‘tomographic AP’ together with ‘on-parametric method’ to get the state-of-art dark energy reconstruction. Professor Yunhe Li from Northeastern University also provided useful helps -- he developed a Fortran package for the non-parametric fitting. The undergraduates then conducted a series of tests on the programme after combining it to the program ‘COSMOMC’ (a cosmological MCMC package commonly used by the cosmology community). After carefully checking the accuracy and reliability, they reached the final conclusion: applying ‘tomographic AP’ in non-parametric dark energy reconstruction can improve the low redshift dark energy constraint by ~100%. The results also slightly favors a dynamical behavior of dark energy (see the figure). To get this results, the undergraduates have been intensively working on the project for a whole semester.

The ApJ referee thought the result being interesting, but raised five questions about the technical details of the paper, including details about the systematical errors. He/She also requested an ‘input-output’ test to demonstrate the reliability of the analysis. These questions take Zhang et al three months to get the required results. One week after they responding the referee’s comments, the paper was received for publication by ApJ.

Link to the paper: https://iopscience.iop.org/article/10.3847/1538-4357/ab1ea4