Home Help GOHelp Tools Create Account Login   Not Logged in

Home

Section 2: Kepler GO Help for Kepler-GALEX Crossmatch

Quick links

Column Descriptions
Section 1: Extending requests beyond the Enhanced Kepler Target Search form

SQL Tutorials

Intro SQL Tutorial
Advanced SQL Tutorial
Note: For the Advanced SQL Tutorial, CASJobs is based upon SQL Server Syntax


Part 1: Introduction
Part 2: Downloads
Part 3: The Cookbook on CasJob Sample Queries
Part 4: Plots, correlations, and distributions
Cautions about matchings

Part 1: Introduction

This page is designed to let investigators conveniently use the MAST Kepler-GALEX cross match catalog. Since both missions use a catalog with their own coordinates, our catalog is meant to allow investigators to go to one place to find objects that have been observed by GALEX that have been or may be observed by Kepler. Kepler parameters in the catalog are those obtained from the ground, not from the satellite.

Part 2 gives links to the list of mutually closest matches (negligible ambiguity of match in either direction), though it misses a small fraction of matches where the positions for the same object in the Kepler catalog (and the Kepler field of view - FOV) and GALEX catalogs differs by more than 5 arcsec) or which may have other problems associated with 1 to 1 object by matching. Part 3 is a brief tutorial to find crossmatches with CasJobs means of customizable SQL queries dropped into a text box. It is probably best to open a second browser window to consult CasJobs pages other than the myDB or Query page, where you will spend most of your time running your queries. Part 4 itemizes links to various plots relevant to the quality of coordinates, separations on the sky, and colors and magnitudes in the two color systems.

The GALEX filter magnitudes (in the mag_AB system) are designated "FUV" and "NUV," and the Kepler ground-based magnitudes are given as griz, as in the SDSS (Sloan) filter system they emulate. The Kepler magnitudes are taken from the "KIC" (Kepler Input Catalog") while the GALEX ones are taken from the Medium Imaging, Guest Investigator surveys and All Sky Imaging (MIS, "GII", AIS) of the GR6 (General Release 6, published in 2010), respectively. Note that no Kepler fluxes measured by the satellite are represented. The addition of the FUV, NUV for those stars in common areas, some 35% of the Kepler FOV (only 24% in the deeper, non-AIS GALEX surveys), corrects for the lack of u magnitudes in the Sloan survey and those allow for the discovery of hot stars and the occasional extragalactic object. Moreover the crossmatch object list is limited by the ability of GALEX to detect faint objects. The Kepler KIC extends to at least 21st magnitude but is known to be incomplete fainter than 17th magnitude (g-band). For these reasons only a small fraction of KIC entries have crossmatches to a FUV and/or NUV magnitude.

Part 2: Downloads

In this section we offer users two quick download options. The first is to download a compressed comma separated value (CSV) file that lists the mutual closest neighbor matches. This list, which we call the "Gold Standard," has the following properties: (1) matches have 1:1 correspondences in both directions (Kepler to GALEX, GALEX to Kepler), (2) coordinates of matched objects in the two catalogs differ by less than 2.5 arcsecs), and (3) no secondary matches exist with separations larger than 5 arcsec. The Kepler-GALEX match page refers to these options as the "accurate" and "complete" searches, respectively.

Note that the Gold Standard list could miss a number of objects for which matches occur with coordinates differing by more than 2.5 arcsecs. For this reason we offer a second download option to make available a full Kepler-GALEX crossmatch table (called KGMatch) for separations up to 5.0 arcsecs.

Altogether, there are 190,220 matches in the Gold Standard table (called "KGGoldStandard") and 379,031 matches in KGMatch. It is also important to realize that about 85% of the matched Kepler objects were observed in both the GALEX Far- and Near-UV bands; only a small percentage of the remaining 15% were observed only in the Far-UV band.

Part 3: The Cookbook on CasJob Sample Queries

In this section we go through the steps to run and download results from the Kepler CasJobs site. Users of this site need to register once (some of you have already done so on the GALEX CasJobs page) and read the home and Help page texts. We recommend that you print out a hard copy of this page or open a second browser window in CasJobs when you are going through the Cookbook.

Let's begin with a bit of vocabulary to define the knobs needed to navigate the CasJobs site, both to stage a query and download results to your machine. The topmost panel of links is referred to as the (top) gutter. Appearing below the gutter are a horizontal row of buttons on certain pages such as the Query or MyDB page. On the left these are called Recent, Clear, and on the right Syntax, Plan, Quick, and Submit. Clicking on buttons in the button or gutter row takes the user to a new CasJobs page or to an extension of the current pagee - or initiates an action such as a query (request) to CasJobs. Drop down menus allow the user to select of several possible choices. Examples are provided by the Sample button. The Context drop down menu gives a list of tables that are available in the system or your own user-created database. Also , open boxes permit the user to enter text such as the Table or Task Name box and, in the large box, the SQL query itslf.

During the execution of these steps it is crucial to keep track of which table you are using: consulting the kepler crossmatch table, e.g. KGMatch, or your own mydb.myTable. This is done from the Context menu. Get used to flipping back and forth between these two context options. The most typical mistake running CasJobs queries is failing to reset the Context selection.


Step 1: Run a sample query as a dry run

(a) Set the Context menu to kepler.

(b) Select Query in the top gutter.

(c) Select your preferred item in the Samples menu. The easiest way to do this is to mouse over the descriptions of the various samples. Then select your favorite one. The English description and SQL syntax should appear in the open query box. Now click on the "Quick" button.

Notice incidentally that we've arbitarily decided to run Kepler GALEX Sample A" on the KGGoldStandard table and the rest on KGMatch. Also notice that angular measures are given generally in decimal degrees, e.g. for coordinates or "box searches" (Sample B), but object-coordinate separations are given in arcseconds.

(d) Take a sneak peek at the results of this query. Click on the Quick button to the right. Within a few seconds you should get a message in green: "Query complete!" If a red one appears, make sure you've really set the context to kepler or check the SQL syntax. At this stage you shouldn't have tampered with this syntax, so this is not likely to be the problem.

(e) Now add a phrase giving a new table that you will be creating with your query by typing into the query, e.g. "into mydb.myTable_listout", before the "from KGMatch" line in the query. You may enter the same name into the Table box above the query box for clarity, but this is optional.

(f) Rerun the query by clicking on Submit. This action will take you to a 'My Query Details" page with a tab in the upper right that is initially yellow stating Ready. These crossmatch queries run over a small database, so within several seconds this tab will convert to blue (Finished) or red (Failed). You can also see the status of a running query by clicking on History in the top gutter; generally this is the quickest way of discovering if a job run under Submit has been executed successfully. Please examine the returned column names. In most or all cases, these names are those used elsewhere on MAST pages. If you have questions about their meanings, contact us.


Step 2: Downloading the table from your dry run

(a) Go to your newly created myTable.. This is done by clicking myDB in the top gutter, selecting MyDB in the menu download. If you decide you want to delete this table for some reason click on the check box in front of your listed table and click on the "All Selected..." tab. This will allow you to make the deletion.

(b) Select your myTable for download. Click on the link to your table. This will take you to a "mytable_WHATEVER" page, where you can click on Download in the row of tabs under the page title.

c) On the new page select your table download format. Users typically download CSV as format so they can play with their table in Excel or convert it to ascii on their own machines.

d) Download the table. This is a two-step process. First, click on the Go button. The table may or may not be converted instantaneously. So, second, mark it for download either by refreshing the browser page or by clicking on Output in the gutter. Either action will cause a new row to appear under the Available Output listing. Click on the yellow Download link after the current timestamp and download the CSV table in the manner that your computer/browser lets you do (e.g., by clicking on the right mouse button to "Download Linked file"). When you download your table, several default fields from both the Kepler and GALEX files will be listed. CasJobs does not allow you to add fields beyond these shown in the listed tables, e.g. the formal errors on magnitudes shown in the Kepler-specific interface forms. The definitions of these fields and their ranges can be found on the results help pages located within the individual MAST/mission sites.


Step 3: Run the modified query you tailor

The execution of your own query runs largely as in Step 1 and identically as Step 2. However, in Step 1, section (c) the actions must now be expanded because after clicking in the boilerplate sample you will make modifications. As before, we suggest a dry run via the Quick submission first. If the execution time for your query is too long for Quick execution, you can add "top 100" after the opening SQL word "select." ("Top 100" means that the query will return the first 100 entries retrieved from the database. To sort them according to, say, KIC identifier, add "order by kic_kepler_id" at the end of your query.)

Omigosh: me execute new SQL commands?! It's not as hard as you think, and if you run into trouble ask us for help at archive@stsci.edu. The first step is to have knowledge of the table fields. You can get this in two ways. First, you will find a quick link to these fields at the top of our Kepler CasJobs home page. Second, you can click "myDB" on the gutter, and then set the Context menu selector to kepler. You will find both crossmatch tables in the list. Clicking on the table entry will bring up a new screen with all the table's fields listed in a horizontal row. If you find new qualifiers to add in a SQL-where clause, like "FUV - NUV" or "FUV - g", etc., make sure that their types (real, integer, string) are what you expect when you state the where condition.

It is our hope that the Samples B and C (which, respectively, filter requests on spatial (sky and magnitude) and physical stellar physical variables will be sufficient for you to walk your way to where you want to go. However, we do solicit the addition of other samples from creative users!


Run a query for an uploaded target list.

Suppose you want to run a simple query on a list of KIC object identifiers. Our Sample D enables this capability. The query assumes a single column list of KIC values. Note that the first line of the input list should give a column title; in our example we refer to your arbitrary (single word) name as "kic_kepler_id". Note that one can run this on the MAST/Kepler Target Search page as well, but the list returned through the CasJobs tool is more reliable.

Cautions about matchings.

None of these queries may give you exactly what you want, and some thought should be given to the concept of matching two pairs of coordinates for given objects. First, note that the KGGoldStandard table is conservative: it comes the closest to giving you an accurate set, but it may not be complete. In general most nearby object with separations greater than 2.5 arcsecs between mission-given coordinates are chance associations of " interloping objects", but not all of them. For this reason the KGMatch table may give you more than you want, but it is complete as the ground-based or satellite observations allow for Kepler and GALEX, respectively.

In practice a great number of things can occasionally go wrong in a matching exercise because of instrumental properties. Not all KIC or GALEX targets are bonafide astronomical objects - a few are image artifacts. Image artifacts have a higher chance of occurring near the rim (occurring at a value near glx_tleCentDist = 0.6 degrees), so objects matched for tile center distances above 0.5 degrees should be treated with caution.

Here is cute example of something that can go wrong. Try running the query: "select * from KGMatch where kic_kepler_id = 7434250 ". This query will return two apparently independent GALEX-detected objects, each falling within 2.5 arcsec of the coordinates for this KIC object, and they have distanceRank values of 1 and 2. In both cases, the glx_detectorOn column is marked "FN", signifying that both FUV and NUV camera detectors were turned on during the observations. However, each of the two apparently separate objects was recorded in only one or the other of these detectors. The objects have glx_tleCentDist values of 0.58 degrees, and the two object coordinates fall 2 arcsec from one another. Moreover, both objects have similar NUV and FUV magnitudes! This is because they are in fact the same astronomical object. In this case the GALEX pipeline processing software could not give the exact same coordinates to the two extracted images because of differential image distortions across the camera fields. Such distortions are largest near the camera rims. Also, note that sometimes phony GALEX images can be extracted because of artifact reflections off the rim. Our example is a good object lesson of what can happen even if no artifacts are present - so be sure and examine your results to see if they make sense.

In addition to GALEX's more modest faint magnitude limit noted in the Introduction, GALEX detectors are also limited at the bright end by saturation. This effect can be seen for objects brighter than 9-10th magnitude expecially for the FUV detector (see Fig. 2 of Morrissey et al. ApJ, 619, L7, 2005). Moreover, sometimes the pipeline processing system will not give a result at all for an object that has been recorded but is too bright to give useful results; in such cases a NULL may appear for the FUV or NUV magnitude even though glx_detectorOn equals "FN."

Part 4: Plots, correlations, and distributions

The Kepler commissioned D. Latham at Harvard/CfA to undertake photometric monitoring of all stars in the Kepler field (users should be aware that some of them are distributed across the whole sky). This monitoring is part of the Harvard group's Stellar Classification Program. It utilized copies of the SDSS project's griz filter set to observe the KIC stars. A preprint by Butner and Smith (soon to become available) demonstrates correlations between magnitude measurements for some 112 stars observed in common. These correlations have slopes of 1.0 and negligible zeropoint offsets. Some of these plots will be posted on MAST/Kepler pages when permissions can be granted.


Plots of useful distributions

GALEX Footprints on the Kepler Field of view: As a general reference, a skymap is shown of the Kepler FOV, superimposed with GALEX tiles (white circles) against it. The local sky density of KIC objects is indicated by shades of blue.

Distributions of positional offsets: We give links to a distribution of positional offsets (r) between the KIC positions, based on the USNO Catalog B, positions derived from the GALEX GR6 pipeline, and the 2MASS project. The separation at the half power point for the Kepler and GALEX posiitons is perhaps r = 0.6 arcsec for the Kepler positions. For the Kepler and 2MASS positions it is the plot between KIC and 2MASS catalog positions. The half power point is 0.2 arcsec or less.

Users should be aware that plots of the differences in Right Ascensions and Declinations for the KIC and GALEX positions show no shift in the centroid for RA. They do show a difference of 0.1 arcsec for Dec. At this writing we do know which of the two catalogs has a greater systematic error in the Dec values. We also point out that the distributions of matches extend out to our arbitrary separation limit of 10 arcsecs. This is based on a separate crossmatch search which we do not post and which can give rise to slightly different results for offsets of just less than 5 arcsecs. We have computed the 10 arcsec results for users who may want to study the "background" offsets of random association of interloping objects. Such users may contact MAST directly for the extended table.

We also point to match count distribution plots for KIC and GALEX colors and magnitudes of the full sample of primary matches out to 5 arcsecs. (For individual plots see tabs at bottom of page.) The following distributions are worth special mention:

NUV magnitude distribution:
This distribution peaks at m_NUV = 23 and drops quickly as we reach the GALEX NUV detector limit.

Kepler g magnitude distribution: This distribution is double-peaked. The primary (bright magnitude) peak reflects a comparatively the population of Galactic stars; at 17th magnitude one begins to sample a decreasing number of stars in the fringes of the Galactic disk. The fainter peak at g-magnitude = 19 and primarily represents faint galaxies.

Kepler g-r color distribution:
This plot gives the distribution of colors and is to first order not selection limited.

GALEX FUV-NUV color distribution:
The GALEX color distribution is double-peaked and reflects the galactic and extrgalactic populations referred to in "Kepler g magnitude distribution above. This was originally noted by Bianchi et al. 2007, Figure 7.

NUV vs. FUV magnitudes of Kepler objects
The NUV vs. FUV magnitude plot of matched Kepler objects is interesting in that it shows two separate distributions. This is a different way of looking at the bimodal distribution referred to in the plot immediately above. The lower distribution of objects, with a slope of 1, represents the primary population of stars given in the KIC catalog. The upper distribution exhibits a slope slightly above 1 and reflects a separate population of objects. The explanation for this bifurcation is not yet clear, but work by L. Bianchi et al. (priv. commun.) suggests that the upper distributions is composed of a mixture of M dwarfs and distant faint galaxies. The latter tend to dominate at about mNUV ˜ 20.

Hybrid NUV-g color distribution:
This GALEX color distribution is single-peaked and reflects the distribution of colors noted, as also noted by Bianchi et al. 2007, Figures 5 and 6.