Published, Not Perished!

I just published my first paper as lead-author! Today I thought I'd give you an idea of what the process was like, and what the paper is actually about.

Choosing a Journal
There are many physics journals out there, each covering different divisions of the field. Our first submission was to Classical and Quantum Gravity, which specializes in general relativity. They rejected us outright (more on that in a moment), and so we had to find another home for the paper. Our next choice was Physical Review D, a subset of the 125 year old Physical Review journal. D covers particles, fields, gravitation, and cosmology. There's an associated journal called Physical Review Letters for short, breakthrough papers – This is where the first detection paper was published. All these journals are under the management of the American Physical Society. When I went to one of their meetings last year, I performed some shameless pandering, which appears to have paid off:

Review Process
When a paper is submitted to a scientific journal, it gets distributed to one or more referees. These are fellow scientists who are in related fields, hence "peer review". The referee's job is to read the paper, give suggestions for improvement to the authors, and recommend to the journal whether it should be considered for publication.

As I mentioned above, our first submission was with Classical and Quantum Gravity, which send it on to two referees. After several weeks, I got an anxiety-inducing email saying that the two reviewers had not agreed in their decision on the paper, so it had been sent on to a third. The trouble was, we didn't know what the disagreement was over: It may have been that one thought it was ready for immediate publication, while the other had modifications to be made first. As it turned out, one wanted to reject it outright, while the other wanted significant changes. The third also dismissed it, and the paper was rejected.

Because the referees write reports instead of simply giving a yes/no/maybe answer, we were able to see the problem: They didn't see any innovation in what we had done, and thought it was more suited to a technical note. That led us to rewrite the paper to put greater emphasis on the concepts we were demonstrating, rather than the technical implementation.

After rewriting, we submitted to Physical Review D. The referees there still thought it needed rewriting, but were willing to consider it if we satisfied their requests. After several back-and-forths, the paper was accepted!

Copy Editing
At this point, the paper was passed off to the copy editing department, which did final checks and formatting. Every reference we used needed to be connected to a known object, so there were several they requested more details for, so they could be properly matched.

Preprint
Typically, academic journals require paid subscriptions to access papers. Most universities will have an account they share with their students and faculty, but for anyone else the paper may be impossible to access. That's where preprints come in – These are earlier forms of papers that may or may not appear in journals that are offered for free online. In physics, we have the arXiv (pronounced "archive" – the X is the Greek chi). After our earlier submissions, we posted a version of the paper there, but since PRD allows the accepted copy to be shared as well, that will go online Monday.

Content
But what's this paper actually about? I talked about aspects of it before while writing my thesis. The idea is this: When we make an observation of a gravitational wave, we want to know when that was emitted from its source, so we can line up our observations and figure out what the source is doing. That's a little harder than it sounds though, since we're making our observations here on Earth, which moves through space. That movement is more complicated than you might think.

The paper is about approximating the timing calculation so that we can compute it more efficiently. The maximum allowable error we chose was 42 microseconds, which translates into an error in position of about 12.5 km, or 8 miles! To get the Earth's position that accurately, you can't just use an ellipse – There are all sorts of little wobbles from the Moon, Jupiter, and Saturn.

To account for those wobbles, we imagined looking at a small patch of sky, like the ones we use in our searches, and getting the exact emission time at the center of the patch. Then we asked, "What's the difference in emission time between any point in the patch, and the center?" By using the difference in time, we were able to get an efficient model that stayed within our error limits.

It was an interesting experience to take a paper through the review and publication process, and I can't thank enough my co-authors Keith Riles and Vladimir Dergachev for their help reaching this milestone. I look forward to many more successful publications!