Equivalent Widths -- Error calculations and more!

I worked on my equivalent width errors -- still really high, but consulted with Jan Marie who had a similar issue. I redid the calculation and they look much better. I went back to my equivalent widths and widened my area around the H alpha emission region to be 10 Angstroms. Plotted the spectra from 6500-6000 Angstroms and then overplotted the selected range around the H alpha region. At first it looked a little wonky (hence enlarging the region). But now it looks really good, and even for the spectra that look like they have a strong H alpha region, the calculation still seems to underestimate the EW. Currently fixing that problem.

After I fix that, on to plotting the the equivalent widths vs. spectral type. There are two plots that I hope to make by Wednesday afternoon. 1) Mean EW per spectral type vs. spectral type, and 2) fraction of active binaries vs. spectral type. In order to do this second plot (and also for just good measure), I need to determine sigma for my equivalent widths (...and then make a histogram and calculate sigma). I was having trouble with the plothist procedure -- it says that one of the called procedures doesn't exist in my library. I downloaded it and put it in the right folder, but it still isn't working so well. I'll check further into it by tomorrow.

BUT this means that I need to get my EW calculation figured out.

Red Sox game tonight!

review of the day

-Worked on error calculation for equivalent widths -- currently have it written up, and in the middle of implementing it into my code. Need to check on error calculation when dividing by constants.

-Fixed equivalent widths -- took into account dM radial velocity and thereby shifted H alpha region over

-Worked on separation calculation - check with Dylan on monday

...and I made possibly the most stupid mistake I've made in a very long time.

Binary separation should be the same for each pair (duh). My calculation does not show that. Not quite sure how this went by me at first, but it did.

I also corrected my equivalent width calculation.

Here are some plots that I made that show a possible correlation between equivalent width along with the velocities at each exposure.

separation and H alpha calculations

I worked on a separation calculation for the binaries today. In order to make sure that I was getting reasonable answers, I pulled out the separation values that Dylan calculated. I spent a while making sure that my values were reasonable - this included attempting to make a histogram. Because some of dylan's values were put in as INF, I ran into some issues, so I ended up just looking at medians of each. Median of calculated separation values = .28 AU, median of dylan's =.49 AU and median difference between Dylan and my calculation separation= .0037 in AU.

Also worked on calculating equivalent widths for H alpha. I chose to look at flux values between 6560 and 6565 Angstroms. Instead of calculating my own area underneath the spectra, I decided to go for the simpler option and use tsum. Looks a little high, but I don't have a real sense of high and low values. Looking through old West papers to get a better idea. Added equivalent width to my huge table.

I started the morning by reworking my system velocity calculation. Redid it and checked in with Dylan to see what he thought. He said it looked good, so I went on to make lots of plots.

The rest of the day was spent going through Dylan's sample and making plots of interesting looking binaries. Usually this included plotting up dM velocity-system velocity vs. WD velocity-system velocity and also time vs. both corrected velocities. Looking for fast samples, and good sig-to-noise ratios. Also re-ran sample to have a better formatted output table.

Example:

So I adjusted my velocity plots in order to take into account the system velocity. They look TOO good, so things that I have to do next:

-consult how Dylan calculate his system velocities

-plot time vs. M dwarf velocity

-check out spec types (earlier M dwarf types should be closer)

Great end to the week

Recap of this week:

-Calculated the radial velocities of the WDs in the first sample of dM+WD binaries

-Calculated system velocity of these binaries, and made sure that it was reasonable (after going back and being a little more stringent on my boundaries for signal-to-noise ratios and velocity limits, sigma ends up being a little less than 20 km/s)

-Applied this code to new sample of binaries from Dylan. I now have a file that has plate, mjd, fiber, dM rv, WD rv, system v, # expos, sig-to-noise and time (days/hours/min/sec).

Excited to go through the sample and have a closer look at some interesting pairs, especially considering that this sample has a much higher average number of exposures for each pair. I'd especially like to analyze some very fast binaries and plot out their radial velocities to see if we can get an idea of their orbits.

I started going through the data at the end of the day. Things I looked for: lots of exposures, good quality spectra (higher stns), relatively okay sounding velocities. I then plotted some up just for fun. Here's an example:

I wish I had more time in the day (have to catch a train in an hour), but really excited about continuing with this. What a wonderful way to end the week!

While I was waiting for the sample to run, I was reading Astrobites and and saw this quote by Ed Salpeter about finding the type of research that fits you best:

"For my scientific temperament I needed a field that was more controversial, more open-ended and new, where quick was useful and sloppy did not matter too much because it would all change soon anyway"

I think this is a really nice quote. One of the major reasons I was drawn to astronomy is that sense of urgency behind it all; right now we're at the time where our technology and scientific methods are catching up to each other, and that's extremely exciting place to be. It was nice to read this quote and see that sentiment reflected in someone who has experienced so much in astronomy.

All in all, a great way to end the week.

It's been a slightly upsetting afternoon. I've been working on applying my code to Dylan's new sample of binaries, but I keep running into issues. I carefully mimicked my previous code in order to avoid obvious errors (i.e. different file names, etc.) but I've having some more serious issues, such as reading some of my fits files. Dylan had to go through my mrdfits.pro program yesterday in order to update it - hopefully this isn't the source of my problems, but it is something I have to more closely investigate. I also just feel like I'm not moving through this as efficiently as I should be. I was hoping to get this done by early afternoon so I could move onto looking at interesting objects (i.e. fast binaries), but it looks like I will spend the rest of the afternoon debugging. Ah well.

After toying around with my histogram a little bit more, I went ahead and calculated sigma. Assuming that it was (relatively) gaussian, I was able to use the FWHM in order to calculate sigma. FWHM roughly equals 2.356 sigma, giving me a sigma of roughly 10.187. Which is reasonable!

I've spent the entire day working on this histogram. It looked pretty funky at first, with a lot of outliers. I've designed it so that I can easily put in different constraints. At the moment, I've been getting rid of spectra that have a signal-to-noise ratio that is smaller than 10, and if either the velocity of the white dwarf or M dwarf is above an absolute value of 600 km/s (also another indicator of crappy data). So far it has been kind of frustrating. Hopefully by the end of the day, my histogram will look a little better.

on to system velocities

After calculating my WD radial velocities, I moved on to calculating the system velocities of each of the binaries. The values seemed to range from reasonable for some of the binaries, to super weird for others. I compared my values to the ones that Dylan worked with, and some of them seem to be an average of the system velocities for the multiple exposures. In order to better gauge my error, I'm going to make a histogram of the differences of the system velocities between each exposure.

velocity shifts calculated for WDs!

Rather than using the system velocity that Dylan has already calculated, I instead applied the velocity shift calculation that I used for the M dwarfs to the white dwarfs in the binaries.

When I first approached this, I said to myself "This will be easy! I already did this for the M dwarfs." But reading in a different type of spectra file, and dealing with my bazillion indices in order for everything to align properly, along with formatting everything into a neat table caused it to ta

ke the better part of the day. It also took me a little to get back into the coding mindset -- but considering I was away for a couple of weeks, it was pretty easy getting back to it.

Applying my code to the WDs also got me prepared to apply it to new samples that Dylan will have for me shortly.

Here is a sample of the shift calculation

using the same binary as the one previously posted. The shift isn't as accurate as for the M dwarf, but this might just be on account of the accuracy of the WDs in general.

Goal for next week: Calculate the system's velocity. Hopefully I will have it in time for group meetings.

It feels so good to be back in the lab!

back to work!

After a wonderful holiday, it's time to get back to the lab! Things to do today:

-read back over orbital velocity calculation steps

-calculate orbital velocity!

It's been a while since I've looked at code. Hopefully all will go well.