URAT South Parallax Results

All raw and processed images have been bias corrected with dark and flat-field corrections applied to the 2-byte-integer FITS files using custom code. We use the same methods as in (Finch & Zacharias 2016) for detecting stellar images with a 4σ threshold above the background. Object centers have been determined using custom code to perform two-dimensional spherical symmetric Gaussian model profile fits of the observed stellar images using the processed pixel data. A significant contribution to the observed point-spread function (PSF) comes from seeing and also diffraction due to the small aperture, which leads to an observed image profile width of about 2 pixels full width at half maximum (FWHM). However, the PSF is uniform across the entire field of view allowing us to use one model function across the entire focal plane.

Grating images of order 1 and 2 were identified with custom code based on the expected brightness, location with respect to a central image and image elongation. Higher-order images are too elongated to make the list of acceptable images in the object detection code. A mean position from the first-order grating images is propagated to the astrometric reductions together with all fitted central image positions (all in pixel coordinate space). Then all individually detected grating images are removed from the observed objects list, before the R.A., decl. is calculated.

Although great care has been taken to make this step as accurate as possible, some spurious detections associated with grating images possibly, impacted by blended images will have made it into the astrometric reduction step. However, the parallax results presented here should be free of those contaminations because of conservative cuts applied to the data throughout this investigation and outlier rejections applied at various stages of the reduction process. Such "left-over" grating image contamination would be much more of a problem for a general URAT south catalog based on all available data.

A special reference star catalog was constructed for this project, the UCAC5 (Zacharias et al. 2017), instead of using the UCAC4, as we did earlier for the URAT north data. Gaia DR1 data (Gaia Collaboration et al. 2016) was used as the basis of this new reference star catalog providing accurate positions of stars in the about 5–21 G-band mag range at epoch 2015.0, close to our URAT south data observing epoch. The Tycho Gaia Astrometric Solution (TGAS; Gaia Collaboration et al. 2016) proper motions of DR1 were adopted for those just over 2 million stars. New proper motions were derived for about 100 million stars using UCAC4 early epoch data and Gaia DR1 positions for stars in common. Typical errors in proper motions are between 1 and 5 mas yr⁻¹ depending on brightness of the star. Thus, UCAC5 extends the TGAS data with similar precision in proper motions to about 14th magnitude and with somewhat lower precision to 16th magnitude.

An eight-parameter "plate" model, the same used in the north, has also been used here for the astrometric reductions (linear + tilt terms). However, for this investigation, we use the UCAC5 (see above) reference star catalog, restricted to a URAT magnitude of 8–15.

We use a conventional weighted least-squares adjustment with outlier rejection individually on each CCD exposure with typically several hundred to many thousand reference stars per astrometric solution (Figure 3). The weights are calculated based on the total, formal errors of individual observations. Exposures with less than 100 reference stars on a CCD were rejected as were exposures failing the observing quality control standards (see URAT1 paper, Zacharias et al. 2015). Using look-up tables from the preliminary reductions and residual analysis, we have corrected the data for both geometric field distortions (10–60 mas) and pixel phase errors (0–15 mas).

Zoom In Zoom Out Reset image size

The URAT data was split into groups of observations spanning about a month each. Preliminary astrometric solutions were obtained for each group and the residuals analyzed. Small (few mas) positional corrections as a function of magnitude were applied to central images, while the grating images required much larger corrections (up to about 80 mas). Astrometric reductions were iteratively repeated with updated correction models. The distribution of the astrometric solution errors (in units of mas and chi-square unit weight) are shown in Figure 4.

Zoom In Zoom Out Reset image size

Epoch positions (α, δ) of all stars have been obtained on the International Celestial Reference System (ICRS) via UCAC5. The errors of individual observations are typically 10–60 mas depending on brightness of the object and exposure time. This matches our typical errors in the north due to the shorter exposure times and signal-to-noise ratio. Individual epoch positions that have been matched to individual stars with mean data and indexing are stored in a separate large file allowing fast, direct access to the individual observations. These URAT data consist of about 400 million individual objects with about 11.3 billion observations. There are 254 million stars with three or more observations each. For more details about the instrument and astrometric reductions, the reader is referred to the URAT1 paper (Zacharias et al. 2015).

A sub-set of about 155 million stars was extracted from the astrometric solution mean position data, retaining only those stars that have at least 10 observations and an epoch span of at least 0.9 year (URAT observing). This set of stars was matched with the UCAC4 and SPM4 catalogs after updating the UCAC4 and SPM4 positions to the mean URAT south observation epoch of 2016.8 using the UCAC4 and SPM4 proper motions, respectively.

The position comparison of the URAT sub-set with UCAC4 revealed 83.4 million unique matches within 2 arcsec per coordinate. Similarly, the match with SPM4 gave 83.3 million uniquely matched objects. The URAT data as well as the UCAC4 and SPM4 matches were labeled with a common, running ID number, which allows easy retrieval of those data as well as the associated individual URAT observations of each star.

The same pipeline as for our northern hemisphere investigation (Finch & Zacharias 2016) was applied here utilizing routines from Jao (2004), the JPL DE405 ephemeris and making use of the parallax factor (Green 1985) for determining parallaxes. We use each URAT mid exposure to determine the location of the Earth. These rectangular coordinates X, Y, and Z at epoch are then used to calculate the parallax factors using the same formulae as our northern hemisphere investigation shown here:

$$\begin{eqnarray}&&{P}_{\alpha }=X\sin \alpha -Y\cos \alpha \end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{P}_{\delta }=X\cos \alpha \sin \delta +Y\sin \alpha \sin \delta \ -\ Z\cos \delta .\end{eqnarray} \tag{ 2 }$$

We then use each individual parallax factor corresponding to all individual data of a given target to simultaneously solve for each astrometric parameter (mean position, proper motion and parallax) using only "good" epoch data in a weighted least-squares adjustment with outlier rejection from the equations:

$$\begin{eqnarray}&&x(t)=x({t}_{0})+{\mu }_{x}(t-{t}_{0})+\pi {P}_{\alpha }\end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray}&&y(t)=y({t}_{0})+{\mu }_{y}(t-{t}_{0})+\pi {P}_{\delta }.\end{eqnarray} \tag{ 4 }$$

Here, as in Finch & Zacharias (2016), the x(t), y(t) are the positions of a given star on the tangential plane as function of time (t), π is the parallax, μ_x = μ_α cosδ and μ_y = μ_δ represent the proper motions in R.A. and decl., respectively. Here, we choose the initial instant of time (t₀ to be the first observing epoch as our zero point for x, y and t).

In Table 1, we show our initial adopted cuts to the URAT south epoch data of each individual star when solving for parallax. These limits have been imposed to not allow saturated stars, stars with too few photons, and stars with poorly determined positions to be used in the fits. We impose a number and epoch span of observation cut as well empirically after comparing the URAT parallaxes to the Hipparcos new reduction (van Leeuwen 2007) and TGAS catalogs.

As in Finch & Zacharias (2016), we add a 10 mas error floor to the random errors before calculating weights for individual observations. These weights, which were used in the least-square adjustments when solving for parallax, can vary largely due to the exposure time and amplitude of individual observations.

We imposed the same 3σ outlier rejection criteria as for the northern data where the largest residuals (typically a few percent) have been iteratively removed from the parallax fit solution of individual stars.

We use the same method here as in Finch & Zacharias (2016) to convert the relative parallaxes from the fit solutions to absolute parallaxes using photometric parallaxes for the same set of reference stars used in the reductions of the URAT positions. This method was used as opposed to the more reliable spectroscopic parallax method due to the lack of spectroscopic data for millions of stars in our survey.

We run each individual reference star in a given URAT frame through 16 photometric color–${M}_{{K}_{s}}$ relations (Finch et al. 2014). These relations make use of the AAVSO Photometric All-Sky Survey (APASS) BVgri and Two Micron All-Sky Survey (2MASS) JHK_s photometry, which are attached to the UCAC4 catalog. We typically have many hundred to several thousands of reference stars in the 2-by-2 square degree area of sky surrounding each target star. We use this data to obtain a mean absolute parallax correction for each target star with the assumption that all stars are main-sequence, due to the lack of information for each individual star.

The mean parallax correction for each target star in this investigation is 1.3 mas, which varies from 0.7 to 8.4 mas depending on the field. Fields with corrections larger than 5.0 mas tend to have fewer reference stars and larger errors on the corrections. This is in part due to the limited color range of our photometric color–${M}_{{K}_{s}}$ relations. For all stars having an unknown correction, we use the mean absolute correction. An unknown correction here is possible due to lack of photometry information in a given field and/or the strict limit of the color range for the relations. If a star has a correction that is greater than three times the average (≥3.9 mas), we adopt this as a cut-off and use the 3.9 mas for the conversion to absolute parallax. The original parallax correction is left in the tables for the readers to use as a flag to determine if an average was used.

We take the same approach as in Finch & Zacharias (2016) and do not apply any corrections to our parallaxes for the Lutz–Kelker bias (Lutz & Kelker 1973) because we do not draw any conclusions about completeness of a distance limited sample or interpret the results regarding absolute luminosity. The goal of this paper is to present the observed trigonometric parallaxes for a large number of stars.

We do however concede that our parallax results could be affected by a statistical bias due to the large number of stars. As explained in Finch & Zacharias (2016), if some observational errors of a distant star with small parallax happen to follow a much larger parallax path on the sky then a "significant parallax" with acceptable error will come out of the solution fit.

The chance of this happening will drastically decrease with increased epoch span (i.e., observations spanning 5–10 years are taken). However, for our stars, we do have only 1–2 years epoch span plus 1 or 2 observations at an earlier epoch for most stars. A single group of few observations accidentally offset a certain way can bias the result significantly. With a high number of stars with formal parallax solution like in this investigation, there will be a fraction of "fake" parallaxes just due to this statistical bias.

The goal of this paper is to discover new nearby stars within 25 pc. To determine how well we can detect nearby stars within 25 pc using our URAT south data, we use the Hipparcos new reduction (van Leeuwen 2007) and TGAS (Gaia Collaboration et al. 2016) catalogs using only the initial cuts from the URAT parallax pipeline summarized in Table 1.

The Hipparcos new reduction contains 514 stars with a parallax ≥40 mas, south of decl. + 20° where we have full coverage and with a Hipparcos magnitude between 9.0 and 16.0 which closely matches the URAT magnitude range. URAT recovers 469 (=91.2%) of this Hipparcos 25 pc sample. Most of the stars not recovered are likely due to the initial amplitude (S/N), parallax solution error and number of observation cuts, which are summarized in Table 1. Of the 469 stars recovered, 417 (=88.9%) have a URAT parallax ≥30 mas. We show a comparison of the Hipparcos 25 pc sample with the URAT results in Figure 5, where the center line indicates perfect agreement and the two outer lines indicate a 10 mas difference from the Hipparcos parallax values.

Zoom In Zoom Out Reset image size

The TGAS catalog contains 297 stars with a parallax ≥40 mas south of +20° decl. where URAT has full coverage having a Gaia magnitude between 9.0 and 16.0, which closely matches the URAT magnitude range. URAT recovers 282 (=94.9%) of these 297 TGAS stars with 276 (=92.9%) having a URAT parallax ≥30 mas. In Figure 5, we show a comparison of the TGAS 25 pc sample and URAT parallax solution.

In order to investigate how reliable the URAT parallax formal errors are, we compared them to TGAS data. Details are summarized in Table 2. A total of 31107 stars are in common between TGAS and our URAT south parallax solution table. After employing basic cuts for the number of observations, parallax errors, and epoch span, 13700 stars remain in our first sample (case 1 in Table 2). Case 2 uses the same cuts but furthermore limits the sample to stars with a TGAS parallax of 40 mas or larger and a URAT parallax of at least 32 mas. For each case, unweighted rms values were calculated for individual catalog parallax errors, the formal error of the parallax differences, and the observed spread of the parallax differences. Assuming the error estimates of the TGAS data are correct, both cases show that our URAT parallaxes formal errors are underestimated by 26%. However, Figure 5 shows that the URAT parallaxes themselves have no obvious bias when compared to the TGAS parallaxes (25 pc sample).

In the process of obtaining a clean new discovery sample for this paper as described in Section 4.2, we recover 623 stars (not in Hipparcos or TGAS) having a previously published parallax ≥40 mas. Of these recovered parallax stars, 64 have a published parallax ≥100 mas. In Figure 6, we show the comparison between the 623 known stars with published trigonometric parallaxes ≥40 mas and the URAT results. When searching for known parallax stars, a varying radius was used with a maximal search radius of 60 arcsec to allow match also with high proper motion targets. The closest star recovered from the URAT south epoch data is GJ 1061 with a URAT parallax of 276.5 ± 3.7 mas or 3.6 pc and a parallax of 270.86 ± 1.29 or 3.69 pc reported in Lurie et al. (2014).

In Figures 7 and 8, we show the relationship between the URAT parallax error with the epoch span of observation and number of observations, respectively, for the 1764 new plus known parallaxes from the URAT south data. These plots indicate as expected that the parallax error drops with more observations over a longer period of time. While the URAT solutions would benefit from this extra observing, a majority of the URAT parallax solutions have a parallax error less than 10 mas with our initial cuts.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We ran the entire URAT south central images observation data through the parallax pipeline, which gave us 24.7 million parallax solutions. The initial cuts applied to the epoch data are summarized in Table 1. Using the investigation described in Section 4.1, we adopted a more stringent set of additional cuts shown in Table 3 in order to obtain a more manageable and cleaner sample. After applying these cuts, we were left with 5556 new candidate nearby stars. We then removed the 701 Hipparcos+TGAS stars, leaving us with a list of 4855 candidate nearby stars. This sample was then matched with SIMBAD, Vizier online catalogs, RECONS 25 pc database, and other published papers to remove 625 stars having a previously published parallax, leaving us with 4230 nearby star candidates.

While this is a more reasonable sample to investigate, we do not take provisions for close doubles in the URAT reduction process, which may lead to many of these new candidates likely still having an erroneous parallax solution. A visual inspection of the parallax residual plots for the entire 4230 candidate nearby star sample was performed leaving us with 1229 nearby star candidates with high quality residuals. In Figures 9 and 10, we show an example of a high- and low-quality fit solution, respectively.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We then matched all stars with 2MASS and APASS using a 10 arcsec match radius to add J, H, K, B, V, g, r, i photometry data, so we could plot an HR diagrams for the 1229 nearby star candidates.

We then checked all stars not on the main sequence by eye using the Aladin sky atlas tool to pull up real sky images. We find 26 stars with incorrect APASS magnitudes due to the large 10 arcsec match radius. For these stars, no other APASS magnitude was available, so we removed the mismatched magnitudes, but they were left on the list as good parallax stars. We then decided to do a visual inspection of sky images for the remaining sample using the Aladin sky atlas tool. By doing this, we removed 313 stars that might have erroneous parallax solutions due to near-neighbor contamination. In Figure 11, we show the HR diagram for the remaining 916 nearby stars. Using these HR diagrams, we can identify six possible White Dwarf (WD) stars. We also identify 5 possible young stars and 46 possible subdwarf stars. These stars are flagged in the table with more information given in Section 5.

Zoom In Zoom Out Reset image size

During the visual inspection of the nearby star sample, we find 19 Common Proper Motion (CPM) and parallax pairs. Of these, 12 pairs did not have a double star designation in the SIMBAD database. All 19 pairs are presented in Table 4 where we give the names, parallaxes, proper motions, separation, and position angle of each pair along with notes. For the 12 new double stars, we picked the component with the brightest V magnitude to be the component "A." If a V magnitude was not present for any of the stars in the pair, we used 2MASS J magnitudes instead. In Figure 12, we show the relationship between the parallax and proper motion of component "A" and component "B" of the CPM pairs.

Zoom In Zoom Out Reset image size

The URAT north and URAT south have an overlapping region between +26° and −12° decl. We ran a match of the 916 new URAT south 25 pc sample with the entire revised 729 URAT north 25 pc sample discussed below resulting in 67 matches. In Figure 13, we show the comparison of the 67 stars in common between the URAT north and south data. Error bars have been added for the URAT north data, which, on average, has higher reported errors than the south data. While these stars have technically been previously found, we include them in both tables here because they were found again using different URAT data. All of these stars have been flagged in the tables.

Zoom In Zoom Out Reset image size

Of these 916 new nearby star candidates, 5 have a URAT parallax ≥100 mas. In Table 5, we present details for the 916 new URAT south nearby stars having an absolute parallax ≥40 mas (sorted by parallax). These stars were found to have no previously published trigonometric parallax. We include the names, R.A., and decl. coordinates (ICRS) of epoch J2015.5: derived from the mean URAT position at mean epoch along with the URAT proper motions, URAT magnitude, estimated spectral type, epoch span, number of observations, number of observations rejected, absolute parallax, parallax error, parallax correction, proper motion with associated errors along with notes. For the 151 entries in the URAT south sample having no previously reported identification in SIMBAD, we have given a USNO Proper Motion (UPM) name.

The URAT magnitude distribution of those 916 new stars within 25 pc is shown in Figure 14. The distribution of the number of URAT observations for those stars is shown in Figure 15. In Figure 16, we show a histogram of the parallax error for the 916 stars reported in Table 5, which peaks around 4 mas. We also show in Figure 17 a histogram of the proper motions for the same sample, which shows that many of the new nearby stars have slow proper motions (less than 400 mas yr⁻¹).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For completeness and quality control, we also ran the URAT north 25 pc parallax results through the same stringent cuts and visual inspections we used here for the URAT south discoveries. This includes cutting all stars with a parallax error greater than 14 mas, visually inspecting all parallax fits, and visually inspecting all-sky images. After applying these cuts, we remove 368 stars from the URAT north 25 pc sample leaving us with 729 having a quality fit and URAT parallax error less than or equal to 14 mas. All 729 URAT north results are presented in Table 6.

During the visual inspection of the URAT north nearby star sample, we find four CPM and parallax pairs. Of these, two pairs did not have a double star designation in the SIMBAD database. All four pairs are presented in Table 7 where we give the names, parallaxes, proper motions, separation, and position angle of each pair along with notes. For the two new double stars, we picked the component with the brightest V magnitude to be the component "A." If a V magnitude was not present for any of the stars in the pair, we used 2MASS J magnitudes instead. The URAT north 25 pc stars were then matched with TGAS, revealing 33 stars, all of which have a TGAS parallax ≥30 mas.

In Figure 18, we show an HRD for the north sample. From this diagram, we identify 10 possible WD stars. LSPM J0543+3637, one of the WD candidates, has not been previously identified as a WD. We also identify 60 possible subdwarf stars. These stars are flagged in the table with a spectral type of "WD" or "VI", respectively.

Zoom In Zoom Out Reset image size

We utilize the K–M spectral transition at V − K_s = 3.7 defined in Clements et al. (2017). If a star has no APASS V magnitude, we use the absolute magnitude at the K_s = 5.1, which is derived using known nearby dwarfs, to separate K and M main-sequence dwarfs.

The majority of targets (864 or 94%) presented in Table 5 are main-sequence dwarfs, and they have spectral types labeled as either "M" or "K." The large bias toward M dwarfs is expected because 70% of the population in the Galaxy is M dwarfs and brighter and earlier types of dwarfs are mostly been measured by the Hipparcos or Gaia in its first data release. All of these red dwarfs reported here have parallaxes greater than 40 mas shown in Table 5. The closest star among this southern sample is LHS 5264 (V − K_s = 5.17, ${M}_{{K}_{s}}=8.4$) at 9.1 pc or 110.3 ± 2.8 mas. It has a nearby X-ray source about 16'' away from its epoch 2000 coordinate, but it has not been discussed specifically in any literature other than several survey catalogs.

K and M subdwarfs are intrinsically fainter than their dwarf counterparts because of their low metallicity, so they are called "subdwarf" (Kuiper 1939; Jao et al. 2008). The K and M subdwarfs are basically scattered just below the main sequence, but their distribution merges with early K and late M dwarfs in the V − K_s versus M_Ks or V − I versus M_V plot. Besides, because of the high galactic velocities from billions of years of galactic heating (Gizis 1997), they usually have high tangential velocities. Hence, these two unique features of cool subdwarfs can be used to separate them from main-sequence dwarfs. Jao et al. (2017) demonstrated using the HRD and tangential velocity to select cool subdwarf candidates. We will use the same dwarf-subdwarf division line and the cut-off V_tan = 200 km s⁻¹ given in Jao et al. (2017) to select subdwarf candidates in this sample.

None of our southern sample exceeds the cut-off V_tan at 200 km s⁻¹. However, many of stars are below the dwarf-subdwarf division line and their locations make them promising subdwarf candidates. All of these subdwarf candidates are labeled with a luminosity class of "VI" in Table 5. For example, LP 764-14 (V − K_s = 4.96, M_Ks = 8.45), a known M3.5 type subdwarf (Reid et al. 2007), is right on the division line and its parallax is 42.1 ± 3.6 mas or 23.7 pc.

We expect the following few stars are extreme-subdwarfs because of their faintness at a given color, UPM 1010-2203, UPM 1215-0254, UPM 0057-3402, LSPM J1237+0804, and LP 651-57.

It is noteworthy that HD 13043B appears to be lower/fainter than a dwarf and could be a possible late K type subdwarf. Our parallax is about 43.9 ± 6.9 mas for this secondary star. HD 13043A is the CPM primary star with a spectral type of G2V (Bidelman 1985) and also has independent parallaxes in various catalogs. The original Hipparcos catalog (Perryman et al. 1997) has a parallax of 27.04 ± 0.86 and the revised Hipparcos catalog (van Leeuwen 2007) has 27.06 ± 0.58 mas. These two measurements seem consistent with each other but the revised parallax has a poor fit according to the goodness-of-fit given in (van Leeuwen 2007).

On the other hand, the Yale Parallax Catalog (YPC) has a weighted mean parallax of 39.6 ± 8.6 mas, calculated from three independent parallaxes of 47.8 ± 15, 38 ± 13.7, and 32 ± 16 mas. Apparently, the parallax of the primary star in the YPC is consistent with our measurement but differs from the Hipparcos data.

Both the Geneva-Copenhagen survey (Holmberg et al. 2009) and the PASTEL catalog of stellar parameters (Soubiran et al. 2010) showed this G2V dwarf has a [Fe/H] > 0 or is a metal-rich dwarf. We would need further spectroscopic observation of the secondary and as well as parallaxes for both components from Gaia to confirm the subdwarf feature of HD 13043B and their associations. We tentatively assign it as a subdwarf.

Like main-sequence dwarfs, WDs also form a sequence far below the main-sequence, but mixed with different types of WDs, i.e., DA, DB, DZ, etc. (Subasavage et al. 2017). In this work, we identify three new nearby WDs using their locations on the HRD and confirm three previously identified WDs.

GD 31 (WD 0231-054) is a known DA white dwarf with an estimated effective temperature of 17470 K (Gianninas et al. 2011). They also reported its spectroscopic distance as ≈29 pc, but our trigonometric parallax is 55.0 ± 10.4 mas or ≈18.2 pc.

UPM 0812-3529 (WD 0810-353) and UPM 0837-5017 (WD 0836-501) are two newly identified WDs. They have parallaxes of 104.4 ± 5.9 and 43.1 ± 5.2 mas, respectively. This makes UPM 0812-3529 the closest WD we have discovered in this survey.

EC 20173-3036 (WD 2017-306) was first detected by the Edinburgh-Cape Blue Object Survey as a blue object with an ultraviolet excess (O'Donoghue et al. 2013). Based on its location on the HRD, it is also a new nearby WD at ≈17.3 pc or 57.9 ± 4.1 mas.

LP 873-78 (WD 2118-261) was first identified by Ryan (1989) as a possible WD using UBVRI colors. Its location on the HRD confirms it is a WD.

GD 1192 (WD2333-165) is also a spectroscopic confirmed DA white dwarf and it has an estimated effective temperature of 13,790 K (Gianninas et al. 2011). They also reported its spectroscopic distance is about 31 pc, but our trigonometric parallax is 41.6 ± 4.1 mas or ≈24 pc.

Young stars are elevated above the main-sequence because they are still contracting along the Hayashi track onto the main-sequence (Hayashi 1961). These pre-main-sequence stars often have active atmospheres and emit strong Hα or can be detected in X-ray, but young stars older than 600 Myr may have saturated X-ray activity (Zuckerman & Song 2004). Thus, using activity alone may not be a good indicator for youth. Here, we use both of their locations on the HRD and atmospheric activities as good indicators to select a few possible young stars.

An unresolved binary can also been elevated on the HRD because of their combined luminosities and its close separation may enhance its atmospheric activity. Hence, the young star candidates we highlight here would need further spectroscopic observations to ensure their youth and high-resolution imaging techniques to resolve possible doubles.

GJ 3237 is a known M4.5 dwarf (Hawley et al. 1996) with an Hα emission (Newton et al. 2017). It also has an X-ray source shown in the ROSAT all-sky survey (Appenzeller et al. 1998).

ASAS J081742-8243.4 is classified as an M3.5 dwarf (Riaz et al. 2006) in the β Pic moving group (Malo et al. 2013), and our parallax confirms its youth and expected elevated position on the HRD.

1RXS J114738.0+050119 has an X-ray source in the ROSAT catalog (Beuermann et al. 1999) and it is clearly elevated on the HRD.

LP 795-38 is an M4V dwarf (Scholz et al. 2005), and Gaidos et al. (2014) reported an Hα emission during their spectroscopic observation. Winters et al. (2015) estimated its photometric distance of 13.3 pc and our parallax is 42.29 ± 3.5 mas or 23.6 pc. Its youth or unresolved binary could make its photometric distance closer than distance calculated from the parallax.

LHS 5273 has an estimated photometric distance at 16.8 pc (Winters et al. 2015) and our parallax is 44.0 ± 4.1 mas or 22.7 pc. The ROSAT all-sky bright source catalog has a X-ray source about 21'' away and no other stellar source, but LHS 5273 is within this radius (Voges et al. 1999). Because of its elevated location on the HRD, a possible X-ray source, and a mismatched photometric distance, we expect LHS 5273 is either a young star or an unresolved double.

LHS 6419 is an M3V dwarf (Newton et al. 2014). It has no X-ray source in the ROSAT catalog but has Hα emission (Newton et al. 2017).

LP 704-15 is a system with a total of four stars. We measure independent parallaxes of the wide CPM pair for the primary (LP 704-15) and secondary (LP 704-14). They have parallaxes of 53.0 ± 4.7 and 48.2 ± 4.4 mas, respectively. The LP 704-15 itself is an SB2 (Bowler et al. 2015) and also has an M8 co-moving companion with a separation about 047.

LP 704-15 was initially identified as a candidate of the Argus association by Malo et al. (2013). However, Bowler et al. (2015) argued that LP 704-15 may be an older dwarf because (1) its stellar activity at FUV and NUV could be produced by the SB2, not youth, and (2) the wide companion, LP 704-14 is lack of Hα emission and could be older than 4 Gyr. Their argument is consistent with LP 704-14's location on the HRD with V − K_s = 4.83 and ${M}_{{K}_{s}}=6.67$, which is not elevated. The elevated LP 704-15 is mainly caused by the combined optical and NIR photometry from SB2 and M8.

2MASS J01130536-7050378 is noted as a possible red subgiant based on the color–mag diagram by Boyer et al. (2011) in the direction of Small Magellanic Cloud (SMC). However, it has a proper motion about 76.9 mas yr⁻¹ from our parallax data and it shows visible proper motion while we blink multi-epoch archival images using Aladin. We think it is a foreground M dwarf at ≈17 pc (V − K_s = 5.91, ${M}_{{K}_{s}}=8.53$) in the direction of SMC.

Wolf 230 with a V magnitude of 11.80 is confirmed in both the URAT north (π = 101.9 ± 6.0 mas, pmra = 61.2 ± 6.0 mas yr⁻¹, pmdc = −265.8 ± 6.0 mas yr⁻¹) and URAT south (π = 105.4 ± 10.4 mas, pmra = 80.4 ± 4.8 mas yr⁻¹, pmdc = −275.5 ± 4.4 mas yr⁻¹) data to be within 10 pc with a mean distance of 9.6 pc.