Difference between revisions of "Monitoring young stars"
m |
m |
||
Line 12: | Line 12: | ||
Time series analysis can be very powerful, and is used in many different fields of science, from tree rings to weather to sunspots to variations of young stars. There is a LOT of information out there on the web with information on analysis of time series data, but everything I found jumped into heavy-duty programming, math, and statistics without much explanation - e.g., the information is aimed at professional scientists. Can anyone else find some online resources with basic explanations, or do we need to write something? | Time series analysis can be very powerful, and is used in many different fields of science, from tree rings to weather to sunspots to variations of young stars. There is a LOT of information out there on the web with information on analysis of time series data, but everything I found jumped into heavy-duty programming, math, and statistics without much explanation - e.g., the information is aimed at professional scientists. Can anyone else find some online resources with basic explanations, or do we need to write something? | ||
+ | |||
+ | Note: [http://homepage.mac.com/dvhscience/SpaceAcademy/Projects/2004-2005/differentialphotometry.html Here is a page] from one of the other teacher teams on reducing their time-series photometry data. | ||
==What are you doing? and Why?== | ==What are you doing? and Why?== |
Revision as of 23:50, 19 November 2007
This page assumes you have read the Finding cluster members article.
The IC 2118 team has embarked on a campaign to monitor some of the candidate TTauri stars that we found using Spitzer data. They are using ground-based telescopes to see if the stars vary.
This page will be filled out more with general information about the project, but you can watch them work here.
Contents
Time series analysis
What is time series analysis?
The term "time series analysis" refers to the analysis of any data set where you have many measurements over some amount of time, for example, watching the same star many times per night over many nights.
Time series analysis can be very powerful, and is used in many different fields of science, from tree rings to weather to sunspots to variations of young stars. There is a LOT of information out there on the web with information on analysis of time series data, but everything I found jumped into heavy-duty programming, math, and statistics without much explanation - e.g., the information is aimed at professional scientists. Can anyone else find some online resources with basic explanations, or do we need to write something?
Note: Here is a page from one of the other teacher teams on reducing their time-series photometry data.
What are you doing? and Why?
What you're doing is pretty straightforward - you are looking for variations in the flux you measure from a star. However, many things can affect the flux - the weather (the amount of dust and humidity in the air), the stability of the atmosphere (seeing), the time of night (are you looking through a lot of air as the target rises or sets, or are you looking at it during its highest point during the night), where on the CCD chip you have placed the target (each pixel is calibrated differently). So, in order to limit the influence of these effects, you can take the average magnitude of every star in the frame, and look at how each star varies with respect to that mean. If the seeing gets much worse during one part of the night, that effect will be felt by every star in the frame. So, the mean will change for that frame. When you look at how the brightness of each star varies with respect to each mean for each frame, then you can take out some of those effects, and just look for variations that (you hope) are coming intrinsically from the target.
So what about the periodic ones?
People who know Luisa's research know that she is interested specifically in finding rotation rates in young stars. Finding periodic variations is a special case of finding variations in general. Finding variations in general is easier. To find periodic variations, you have to not only watch while the variation completes at least 2 cycles (to make it believable that you have found periodicity), but also do some relatively fancy math. (If you've heard of Fourier transforms, it's related to that - not quite the same, mind you, but close -- see the scary words here.) You can't just look at the light curves and say a-HAH, I see something periodic! (Try it in the light curves below, and I bet you that no two of you will find the same periodicity, much less find the same ones in the same stars that the math finds.) Right now, we don't know of any generally-available tools that will let you do this analysis, but we are looking. If you know of any, please let us know.
Data to play with
raw ground-based data to practice with (data taken by Mr. Spuck and students) - images on which you can measure photometry.
luisa's light curves (data taken by R. Makidon and used by L. Rebull for part of her thesis - just the reduced photometry time series from many 100s of images, for ~5000 stars. Read the "readme.txt" file in order to understand what you are looking at.
raw Spitzer data to play with can be found in the data that comes as part of program 20079. The final paper from these data is here; note that we did not find any variations that we could say for sure came from the object and not the instrument! Note too that many of the programs looking for planetary transits use the same basic observing strategy, so you can go get and reduce those data too! If NASA gives us funding to keep going in the Warm Mission (after Spitzer runs out of cryogen), there will be many opportunities for more time series analysis using Spitzer data.
Where are you looking?
Spot has a feature (as does Leopard for that matter) where you can overlay outlines of the Spitzer instrument fields of view on a larger image, and that way you can see where in the sky you are pointing. If you would like to see where in the sky you are looking with your ground-based telescope for your monitoring project, you can also do this in Spot!
- Download, install, and start Spot.
- Create a target. Click on the bullseye icon, or pick "new target" from the "targets" menu. Enter in a name, and RA/Dec for a part of sky near your target. Tip: if you are using a well-known name, just enter the name, and ask Spot to use Simbad or NED to obtain the RA/Dec for you (="resolve the name").
- Download a background image. Go to the "images" menu and pick an archive to use. Tips: ISSA is the IRAS all-sky survey, and will let you download an image up to 5 degrees on a side. 2MASS is the 2-micron all-sky survey, and will only let you download relatively tiny pieces of sky at one time. You can also load a fits image from your computer.
- Ask Spot to overlay the field of view. After you have an image in place, go to the "overlays" menu and pick "generic focal plane". You will get a dialog box asking you the focal plane type (e.g., do you want it to draw a circle or a square?) and the size ("side/diameter") in arcseconds for the circle or square. You can also change the color it uses.
Voila! You can do this again and again on the same image - just keep creating new targets and asking it to overlay the focal plane at the location of each of your targets.