UCAC4 NEARBY STAR SURVEY: A SEARCH FOR OUR STELLAR NEIGHBORS

A comprehensive census of the nearby stars is required to determine accurate stellar luminosity and mass functions in the solar neighborhood, and is vital to our understanding of the distribution of stellar mass among different types of stars throughout the Galaxy. In addition, because of their proximity to the Sun, the nearest stars are the most accessible for surveys of stellar activity, ages, multiplicity, and exoplanets. A volume-limited sample of stars is dominated by red dwarfs that make up more than 70% of stellar objects in our Galaxy (Henry et al. 2006). Still, red dwarfs are currently under-sampled by volume because of their low luminosities, coupled with the magnitude limits of current all-sky surveys and the astrometric limitations of scanned plates. Ground-based surveys like the U.S. Naval Observatory Robotic Astrometric Telescope (URAT) (Zacharias 2005) currently underway (Finch et al. 2012a, 2014) will pick up stars down to 18th magnitude, while space missions like Gaia (launched in 2013 December) will reach to 20th magnitude.

Significant discoveries of nearby stars have been made in the past decade by searching the Digitized Sky Survey in the northern sky (Lépine 2005, 2008), the SuperCOSMOS Sky Survey in the southern sky (Hambly et al. 2004; Henry et al. 2004; Subasavage et al. 2005a, 2005b; Finch et al. 2007; Boyd et al. 2011a, 2011b; Scholz et al. 2000, 2002), and the USNO CCD Astrograph Catalog (UCAC), also in the southern sky (Finch et al. 2010, 2012b). All of these studies rely primarily on searching for stellar systems with detectable proper motions. Nonetheless, currently available data sets have yet to be completely searched for nearby stars, particularly those with small proper motions.

In this paper we search the more than 100 million sources in the U.S. Naval Observatory fourth CCD Astrograph Catalog (UCAC4; Zacharias et al. 2013), which reaches to 16th magnitude, to reveal nearby stars. This all-sky survey takes advantage of the newly released astrometric results from the UCAC4 along with merged optical photometry from the American Association of Variable Star Observers (AAVSO) Photometric All Sky Survey (APASS) and infrared photometry from the Two Micron All-Sky Survey (2MASS). With the addition of the APASS photometry in the UCAC4 release, we are now able to create a suite of photometric color–${{{\rm M}}_{{{K}_{s}}}}$ relations that provide distance estimates to nearby red stars with uncertainties of 15%. Using this suite of relations, we have combed the UCAC4 catalog for candidate nearby stars to reveal new candidates that were missed in previous searches, most of which required that the stars have detectable proper motions. In Section 2 we describe the data sets and creation of the photometric color–${{{\rm M}}_{{{K}_{s}}}}$ relations that incorporate BVgriJHK_s photometry and high-quality trigonometric parallaxes. We outline the search for nearby star candidates in UCAC4 in Section 3, describe the results in Section 4, and provide concluding remarks in Section 5.

2.1. UCAC4 and Wide-field Infrared Survey Explorer (ALLWISE)

2.2. Photometric Distances

To obtain photometric distances for UCAC4 sources, we generated a new set of 16 photometric color–${{{\rm M}}_{{{K}_{s}}}}$ relations using (a) $BVgri$ optical photometry from APASS, (b) JHK_s near-infrared photometry from 2MASS, (c) nearby red dwarfs with high-quality trigonometric parallaxes (defined as having errors of less than 5 milliarcseconds from the Research Consortium On Nearby Stars (RECONS)⁶ group, and (d) a set of M dwarfs with spectral types M6.0–M9.5 V within 25 pc referred to as the supplemental sample in Henry et al. (2004, hereafter TSNX). The method used is similar to that in TSNX, in which the ${{{\rm M}}_{{{K}_{s}}}}$ band is used with UBVRIJHK_s photometry to obtain 12 relations. The ${{{\rm M}}_{{{K}_{s}}}}$ band is selected here to boost the number of very red stars having photometry in the band used for luminosities (e.g., fewer very red stars have V photometry), and to mitigate the problem of interstellar reddening (although minimal within 25 pc). The input sample contains 110 stars from RECONS, a 64 star sample from Riedel et al. (2010, hereafter TSNXXII), and 31 stars from the very red supplemental list, bringing the total sample to 205 stars having accurate trigonometric parallaxes.

We matched the input list of 205 parallax stars with the UCAC4 catalog to obtain proper motions and BVgriJHK_s photometry. We filled in missing photometric data (photometry not found in UCAC4) using the APASS website, VizieR, or Aladin as needed. Aladin was also used to verify stars by eye to eliminate any source misidentifications, of which none were found. All parallaxes were updated with the most recent results from the RECONS database, and APASS photometry was updated using DR7 web access when available. All stars in close multiple systems, known subdwarfs, known white dwarfs, stars without enough photometric data for at least seven relations, stars with photometric errors greater than 0.10 magnitude, and stars brighter than the APASS saturation magnitude limits (defined in Table 1) were removed from the list, leaving a total of 168 stars for the color–${{M}_{{{K}_{s}}}}$ relations. This list was then given a final spot check by eye using Aladin and specific catalogs to verify photometry and parallaxes.

Table 1.  APASS Photometric Saturation Magnitude Limits^a

Bandpass	Saturation Limit
B	10.25
V	10.25
g	11.00
r	10.00
i	9.00

^aAs of Data Release Six (DR6).

Download table as:  ASCIITypeset image

Combinations of eight filter bands provide a total of 28 possible color–${{M}_{{{K}_{s}}}}$ relations. Of these, 16 were deemed to be reliable, in that there was at least a one magnitude range in color using the sample of stars available for a fit, and there was a reasonable correlation with ${{M}_{{{K}_{s}}}}$; an example is shown in Figure 1 for ($V-{{K}_{s}}$). A second-order fit was used for all relations, with higher orders not showing any meaningful improvements. In Table 2, we give the applicable color range, number of stars used, the fit coefficients and the root mean square (rms) values for each of the 16 color relations used in this paper. Some stars do not have the complete set of eight-band photometry available or might lie outside of the applicable color range; thus, the number of stars used for each fit is less than the sample of 168 stars. The equations used for the relations have the following format, shown here for the ($V-{{K}_{s}}$) relation:

$$\begin{eqnarray}&&{{M}_{{{K}_{s}}}}=-0.1239{{(V-{{K}_{s}})}^{2}}+2.523(V-{{K}_{s}})-2.472.\end{eqnarray} \tag{ 1 }$$

These relations are similar to those in TSNX in the number of stars used (∼100–140) and rms values (∼0.4 mag). Both suites of relations would benefit by increasing the number of very red stars to extend the applicable ranges of each fit, such as those reported in (Dieterich et al. 2014).

Zoom In Zoom Out Reset image size

Table 2.  Details of the 16 Photometric Distance Relations

	Color Range	Stars Used	Coeff. 1	Coeff. 2	Coeff. 3	rms
Color	(mag)	(number)	(× color2)	(× color)	(constant)	(mag)
V–i	1.4–3.9	113	−0.33800	+3.663	+0.9571	0.38
V–J	3.1–7.0	118	−0.14260	+2.551	−1.0870	0.40
V–H	3.5–7.7	118	−0.13910	+2.657	−2.3950	0.41
V–K	4.0–8.0	118	−0.12390	+2.523	−2.4720	0.42
B–i	2.8–6.0	102	−0.13570	+3.091	−1.6910	0.38
B–J	4.2–9.0	140	−0.08928	+2.188	−2.6330	0.37
B–H	4.9–10.0	141	−0.09372	+2.347	−4.0930	0.39
B–K	5.0–10.0	141	−0.08031	+2.174	−3.8580	0.39
g–i	2.2–4.5	105	−0.13340	+3.619	−1.0110	0.39
g–J	3.9–7.8	113	−0.13290	+2.693	−2.8800	0.39
g–H	4.2–8.4	113	−0.11760	+2.796	−4.2560	0.41
g–K	4.5–8.8	108	−0.13480	+2.637	−4.2120	0.41
r–i	1.0–3.0	105	−0.26210	+3.237	+2.9860	0.41
r–J	2.9–6.2	107	−0.11530	+2.383	+0.2988	0.41
r–H	3.4–6.8	107	−0.28630	+2.502	−0.9758	0.42
r–K	3.5–7.1	102	−0.19210	+2.353	−1.0590	0.42

Download table as:  ASCIITypeset image

The accuracy of the distance estimates has been evaluated by running the sample of 168 stars used to construct the fits back through the final relations. The resulting average distance errors of the 16 estimates, i.e., the differences between the photometric distance estimates and the distances from the trigonometric parallaxes, is 15.5%. This is virtually identical to the results of TSNX, in which the average error for the 12 relations was found to be 15.3%. We show a comparison of the photometric distance estimates and the trigonometric distances in Figure 2. Of the 168 stars used to construct the relations, 101 (60.1%) lie within the adopted error of 15.5% and 149 (88.7%) lie within two times the adopted error of 15.5%. All distance estimate errors reported in this paper include the 15.5% external error from the fits and the standard deviation from the up to 16 distance estimates for a given star, added in quadrature.

Zoom In Zoom Out Reset image size

We used a series of four cuts of the UCAC4+APASS+2MASS databases to extract a candidate list of stars within 25 pc. In the first cut, sources were extracted from UCAC4 that met the following criteria.

• 1.  
  Target must have photometry in at least two filters in the APASS catalog, with photometry errors (apase) ≲0.10 mag.
• 2.  
  Target must have photometry in all three JHK_s filters in the 2MASS catalog, with photometry errors (e2mpho) ≲0.10 mag.
• 3.  
  Target must not have a UCAC object flag (objt) = 1 or 2, indicating it is near an overexposed star or a streaked object.
• 4.  
  Target must have a valid, non-zero proper motion in UCAC4.
• 5.  
  Target must have a Lyon–Meudon Extragalactic DAtabase (LEDA) galaxy flag (leda) = 0 and a 2MASS extended source flag (2mx) = 0—both indicate a point source.

There were 25,865,591 sources output from this query. Those were then run through the suite of photometric distance relations and a second cut was performed that extracted only those for which (a) at least 7 of the 16 relations could be applied, and (b) distances were estimated to be within 25 pc, yielding a list of 381,054 candidate nearby stars.

To build a data set for further investigation, this list was then cross-matched using a 30 arcsecond radius with the RECONS 25 Parsec Database, the Hipparcos catalog via VizieR, SIMBAD, and selected journal papers to gather names, spectral types, parallaxes, distance estimates, and other useful information about these stars. During this process, we noticed that many candidates were either associated with or within 30 arcseconds of X-ray sources; these candidates are noted as such in the tables. A list of journal publications and online catalogs used during this search to help build the data set is given in Table 3.

Removing background giants from this large candidate nearby star sample was accomplished using two more cuts. The third cut identifies giants using $J-{{K}_{s}}$ and $V-{{K}_{s}}$ colors, similar to the techniques outlined in Riedel (2012) for candidates extracted using SuperCOSMOS and 2MASS data. The fourth cut uses the UCAC4 proper motions in a Reduced Proper Motion diagram (RPM), as seen in Lépine & Gaidos (2011) and similar to the RPM diagrams in Subasavage et al. (2005a, 2005b), Finch et al. (2007), and Boyd et al. (2011a, 2011b) to further eliminate any contamination.

In the top panel of Figure 3 we show a sample of known giants (dots) and M dwarfs (open circles) pulled from the 381,054 candidate nearby star list, supplemented with a known sample of dwarfs from RECONS. Using $V-{{K}_{s}}$ and $J-{{K}_{s}}$ colors, we constructed two boxes similar to those in Riedel (2012) that likely contain M dwarfs (vertices are given in Table 4). Box 1 contains the most likely M dwarf candidates, while box 2 is more likely to be contaminated by giants. The bottom panel in Figure 3 shows the entire 381,054 sample of nearby star candidates, and after the third cut, we reduce the sample to 4424 candidates that are found in the two boxes.

Zoom In Zoom Out Reset image size

To further eliminate background giants, the fourth cut utilizes proper motions from UCAC4, combined with photometry from APASS and 2MASS, to plot an RPM diagram. We used a modified distance modulus Equation, where μ in arcsec yr⁻¹ is substituted for distance:

$$\begin{eqnarray}&&{{H}_{v}}=V+5{\rm log} \mu +5.\end{eqnarray} \tag{ 2 }$$

An RPM diagram is useful in separating dwarfs and giants because distant giants tend to have very slow proper motions compared to nearby dwarfs. In the top plot of Figure 4, we show an RPM diagram of the same known sample of giants (dots) and M dwarfs (open circles) used in Figure 3. A delimiting line is drawn to separate the two classes of stars, with vertices at (2, 8) and (8.5, 16). In the bottom plot of Figure 4, we show the entire initial sample of 381,054 candidates, and use the RPM cut to eliminate another 2408 presumed giants from the 4424 candidates in the two boxes outlined in Figure 3 for the color cut, yielding 2016 candidate nearby stars. After removing 255 duplicate entries, which we found to be the only duplicates in the entire initial 381,054 sample, we are left with a manageable list of 1761 candidate nearby stars.

Zoom In Zoom Out Reset image size

After the color and RPM cuts, all stars having known parallaxes (669), distance estimates in a published paper (749), or known to be giants (4) were extracted, leaving 339 new candidate nearby stars within 25 pc. Details for all 1761 nearby candidates are given in Tables 5 (339 new discoveries) and 6 (1422 recovered stars), where we give the system names, R.A. and decl. coordinates, proper motions in R.A. and decl. with associated errors, $BVgriJH{{K}_{s}}W1W2W3W4$ photometry, distance estimates with associated errors, the number of color relations used for the estimates (typically all 16) and relevant notes. In Table 6, we provide additional columns for published distances and the type of distance (P = photometric distance; T = trigonometric distance). The first 28 lines of Table 5 are provided in the printed journal sorted by distance and representing all new candidate nearby star systems out to 15 pc. The first 15 lines of Table 6 are provided in the printed journal sorted by RA. The full versions of Tables 5 and 6 can be found in the electronic version of this paper. All positions, proper motions, and BVgriJHK_s come directly from the UCAC4 with some APASS photometry updated to the recent DR7 release for those stars with missing photometry (current release data can be found at www.aavso.org/apass). If no photometry was found or the photometry between 2MASS and APASS was not in agreement in the UCAC4 then the APASS and 2MASS databases were searched manually using either VizeiR or the APASS websites to check for consistency and to fill in gaps.

As a final check, the 339 new nearby candidates were evaluated by eye using Aladin to reveal false detections and to verify proper motions. Of these, 101 did not show any detectable proper motion (noted in Table 5). Further investigation showed that all but six of the 101 objects had fewer than one arcsecond of total UCAC4 proper motion given the epoch spread of the digitized plates used for visual inspection.

Finally, it is evident from looking at Figures 3 and 4 that a few bona fide M dwarfs are seen outside the boxes or above the delimiting line. As a result of the cuts, eight known M dwarfs were omitted from a sample of 181 known M dwarfs, the same sample shown in the top panels of Figures 3 and 4, which indicates that we have omitted ∼4% of nearby stars by our search method. Ultimately, given that our goal is a clean sample of dwarfs, we have omitted a few in favor of minimizing contamination by giants.

The 339 new nearby star candidates are all estimated to be within 25 pc of the Sun, including 5 predicted to be within 10 pc. Of the new 339 nearby star candidates, 218 are previously unknown proper motion stars and have been given a USNO Proper Motion (UPM) name designation for this paper, consistent with previous discoveries from our UCAC efforts (Finch et al. 2010, 2012b).

In Figure 5, we show a color–magnitude plot of 479 known stars with trigonometric parallaxes in Table 6 having available ALLWISE photometry (top) and the same color–magnitude plot of the 339 new nearby star candidates from Table 5 (bottom) for comparison. In Figure 6, we show a color–color plot of 479 known stars with trigonometric parallaxes in Table 6 having available ALLWISE photometry (top) and the same color–color plot of the 339 new nearby star candidates from Table 5 (bottom) for comparison. The known sample (top) in Figures 5 and 6 show some outliers from the bulk of the sample which have been labeled on the plot. The new nearby star candidates reported in this paper compare well with the bulk of the known sample.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In Figure 7, we show the sky distribution of the 339 new nearby star candidates. This plot shows that during this all-sky survey, more new nearby star candidates were discovered in the southern sky than in the north. This is because the southern hemisphere has historically been less rigorously searched for stars with proper motions less than $\sim 0_{\cdot }^{{\prime\prime} }18\;{{{\rm yr}}^{-1}}$, the cutoff of Luytenʼs all-sky survey (Luyten 1980). In Figure 8, we show a proper motion histogram of the newly discovered nearby star candidates in $0_{\cdot }^{^{\prime\prime} }02\;{\rm y}{{{\rm r}}^{-1}}$ bins, highlighting with dark bars those having distance estimates within 15 pc. This illustrates that most of the new nearby stars we have revealed are found in the slower proper motion regimes, yet each is estimated to be closer than 25 pc.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We used the RECONS 25 pc Database to construct a catalog of 320 stars with accurate trigonometric parallaxes placing them within 10 pc. Many of these objects are very bright (AFGK stars) or very faint (brown dwarfs), and are therefore outside the limits of our current search, which targets red dwarfs. A detailed analysis indicates that 75 did not match up to any UCAC4 entry, 34 did not contain any APASS photometry in UCAC4, 54 did not have any 2MASS photometry listed in UCAC4, 75 did not contain photometry with errors less than or equal to 0.10 mag, 10 did not meet the color range for the relations, 9 did not meet the criteria of using at least 7 relations for the distance estimate, and 1 was rejected because of an object flag. Of the remaining 62 stars that were recovered successfully, 51 (82%) are estimated to be within 10 pc using the relations presented here, and only 1 is estimated to be beyond 15 pc. We conclude that our search is successful in revealing nearby red dwarfs, and that our new relations provide reasonable distance estimates.

4.1. Local Statistics

We have identified 339 new nearby star candidates within 25 pc via this search, of which 218 were previously unidentified in any search or catalog. Most of these new discoveries are nearby stars with proper motions less than 0.2 arcsec yr⁻¹, indicating that many of the Sunʼs nearest neighbors have been missed to date because of the emphasis on larger proper motions.

The 16 relations reported here can be used to estimate distances to nearby red dwarfs within the applicable color ranges given in Table 2. The relations are quite reliable, as evidenced by the number of nearby stars recovered during the search. The primary limitations of the method are a lack of APASS sky coverage, and that the relations focus on red dwarfs only. This search used the UCAC4 catalog as a starting point, but a new search could be done once the APASS and upcoming URAT catalogs are complete, providing improved photometric sky coverage and depth.

Followup trigonometric parallax measurements are required to confirm which of the nearby star candidates are, in fact, closer than 25 pc. Given continued observations by the RECONS team from the ground, and Gaiaʼs anticipated astrometric accuracy and limiting magnitude, most or all of these targets should have accurate parallaxes in the coming years.

We thank the entire UCAC team for making this nearby star search possible. Special thanks go to members of the RECONS team for their support and use of the RECONS 25 pc database. This work has made use of the SIMBAD, VizieR, and Aladin databases operated at the CDS in Strasbourg, France. We have also made use of data from 2MASS, APASS, ALLWISE, and the ADS service.