Release Note for LAMOST DR4_v2

The DR4 was published to domestic users on December, 2016, but we found a few of bugs after this data release, such as a 2.7 km/s systematic difference of the stellar radial velocities.

DR5 has been released to domestic users at the end of 2017, and has modified the problems found in DR4. Thus, we produced a new version of DR4, which include the data of pilot survey and the first four year regular survey in DR5. We abbreviate this new DR4 data as “dr4_v2” which revised the problems in old version, and published it on the website http://dr4.lamost.org/. The old version of DR4 data is still retained which is abbreviated as “dr4_v1”, and was published on the website http://dr4.lamost.org/v1/.

We compare the parameters of “dr4_v2” and “dr4_v1”, and the results are shown as follows.

1. The Parameter Comparisons between dr4_v2 and dr4_v1

We select targets which have both dr4_v2 and dr4_v1 parameters, and compare their atmospheric parameters (Teff, logg, [Fe/H]) and rv of the two versions in following figure. The average values and standard deviations of differences of the two versions are also shown in the figure.

From the upper left panel of the figure, we can see that there is a vertical substructure around Teff_dr4_v1 = 4500 K and a horizontal substructure around Teff_dr4_v2 = 4500 K. The number of stars in two substructures is less than 500, and most of their sng are less than 10. We manually checked those spectra in dr4_v1 and dr4_v2, and found that almost all of them have low quality, and relatively worse in dr4_v1. Most of those low sng spectra have similar continuum shapes and noise pattern, which might be the reason of the two substructures.

The histograms of the differences between dr4_v2 and dr4_v1 for atmospherical parameters (Teff, logg, [Fe/H]) and rv are shown as follows. The differences of parameters are divided into four groups according to the snr, presented by different colors. We can see that the scatters of atmospherical parameters in the bin of 0<snr<30 are obviously larger than other bins.

2. The Parameter Comparisons between dr4_v2 and DR3

We select targets which have both dr4_v2 and DR3 parameters, and compare their atmospheric parameters (Teff, logg, [Fe/H]) and rv of the two versions in following figures. The average values and standard deviations of differences of the two versions are also shown in the figures.

The histograms of the differences between dr4_v2 and DR3 for atmospherical parameters (Teff, logg, [Fe/H]) and rv are shown as follows. The differences of parameters are divided into four groups according to the snr, represented by different colors.

3. The Parameter Comparisons between dr4_v2 and SDSS

The difference of parameters between LAMOST dr4_v2 and SDSS (DR14 and APOGEE) are shown in this section.

(1). The following figures show the comparisons of atmospherical parameters (Teff, logg, [Fe/H]) as well as rv between LAMOST dr4_v2 and SDSS DR14. For the parameters of logg, a branch is seen in the bottom-right of the top-right panel. It could be explained by the inhomogeneous distribution of the atmospherical parameters for the ELODIE library. There are holes for giants in the ELODIE library, making the estimations of the initial guess for atmospherical parameters in the giant holes inaccurate.

(2). The histograms of the differences of atmospherical parameters (Teff, logg, [Fe/H]) as well as rv between LAMOST dr4_v2 and APOGEE are provided as follows. In comparison with APOGEE, the differences of parameters are divided into four groups according to snr. We also plot the mean values and the standard deviations corresponding to each snr bin in the figures.

Note that, the equatorial coordinates of 291 .fits files are not correct in the DR4_v1 data, and their plate numbers are the same. We have already revised these coordinates in this version, and provide the update coordinate list here.