Lessons learned from 1/31

At this point I've seen the data from all of the telescopes for the 1/31 event and I cannot say for certain whether we detected anything or not (which is unfortunate, because that means we can't reduce the timing error bars for future events yet).

The main difficulty is that most telescopes (in particular, the 3 largest telescopes) had only one night of observation covering about 6 hours at most, so there is not complete overlap of the 3.91 hour light curve. What overlap there is comes from high airmass or deep twilight observations.

The only solution to this problem is to observe for multiple nights, to get a fiducial lightcurve to which to compare. This was done at MRO and the Faukes North telescopes and results are still ambiguous at best.

Much better will be when Haumea is up for ~10 hours in a row in the next month and we can observe 2+ rotations to have a direct comparison. With luck this will make a big difference.

In the meantime, we are trying to construct multiple-rotation light curves by stringing together data from different observatories. This stringing only works if we all use the same filter. In the past I had no recommendation for filter, but after these events I think I will suggest R filters for everyone to all ease of comparison.

Over the next few months we will collect enough full light curves with and without events that I believe we will be able to go back and find the event in the 1/31 data. We might even be able to find it when people do more careful reductions.

In the meantime it is now time to start focusing on the 2/18 event. It will be visible over most of Asia and we know of several ~2m telescopes that are attempting observations.

Analysis continues

By now I've seen at least preliminary data from most observing teams and I tentatively think that we detected Namaka going into eclipse. The difficulty is that the eclipse actually takes some time -- ~15 minutes -- during which Haumea is changing rapidly.

As an example of the sort of data to be analyzed, I show below the first reduction data from the Palomar 200-inch telescope (in blue) and the University of Hawaii 88-inch telescope (in black) scaled to match each other as best I could. Times are UT the night of Jan 31, and intensities are relative to field stars.


In this plot you see about 1 3/4 rotations of Haumea. Haumea is elliptical and rotating end over end with -- we think -- one of the faces slightly darker than the other, thus the slight differences between the peaks and troughs 2 hours apart.

The expected signal of the Namaka eclipse is a ~1% dimming of the signal for ~1 hour!

Nothing is obvious by eye, so the analysis requires careful examinations of the before and after light curves. With almost 2 complete rotations we can do that comparison here. Even better, though, at least 2 observatories did observations on separate nights to get good comparison data. These give the best shot of showing something.

I think I know where the event is in the light curve above, as I also compared everything to a well measured light curve from Lacerda (though it is clear that there have been changes from when he observed). But I am eager to hear from the multi-night observations for a better indication.

Observing strategy

The main goal for these observations is to obtain accurate relative photometry with sufficient cadence to resolve the events (ingress and egress each last ~15 minutes, but there might be interesting structure within) and sufficient accuracy to detect the event (Namaka's disappearance during an eclipse will make the overall brightness of Haumea decrease by only about 1.3%).

The optimal exposure time for your observations depends on your telescope and on your camera. The main things to worry about are:
-readout time for your camera
-time to saturation
-time until you are sky background limited

I would shoot for having exposure times be at least 2-3 times the readout time of your camera, as long as you are not coming close to saturating Haumea. Even with a very fast readout camera, exposure times under a minute or two are probably not worthwhile. The best cadence, if it fits your camera characteristics, is probably around 2 minutes per exposure. At Palomar, we will be doing 2 minute exposures with 45 second readouts in between.

Previously I said this:
In general, the optimal filter is the one that gives you the highest signal-to-noise. Haumea is almost precisely solar colored, so observing in something like a V filter is a good bet, though it is possible that you might want to go totally unfiltered (at Palomar we will be using a special filter that cuts out some of the worst of the sky lines from the bright Palomar sky).

However, after the 1/31 event I now believe that it is best if everyone observes in as common a filter as possible. Because some of the events will have heavy moonlight, R seems the best choice. At Palomar we used a slight variant of R that cuts out the strong sodium lines from San Diego, but as long as everyone is R-ish I think we will be able to compare all of the data well.

At some point in the night you will want to insert a single image in a different filter. This single image will be used to get the colors of all of the stars relative to Haumea. The relative photometry can then be done using only stars that are close to Haumea's solar color. This selection will be important given the very wide range of airmasses Haumea is likely to go through (at Palomar we will start at an airmass of 4!).

Before Haumea rises (or on an earlier night, at twilight, or with dome flats):

• take images of the night sky/twilight sky or just dome flats
• look for a particularly well behaved region on the chip where the flat field is as flat as possible, there are no bad pixels or columns, nothing funny is going on for a moderately good sized region.
• Note that location for later; that is where you will put Haumea
  

The observing strategy is quite simple:

• point to Haumea as early as your telescope possibly can.
• place Haumea in the very nice region of the chip that you identified earlier.
• track at Haumea's rate; do not dither! Dithering is great if you want to average out CCD variations over time, but we instead want to keep them constant with time.
  
• observe all night until you can't anymore
• pause for nothing.
  
• if you fear focus changes, focus on an image-by-image basis rather than stopping to focus.
• Be sure to insert one or two images in a second filter.
  

We believe that the best strategy will be to track on the rate of Haumea but perform no dithering of the telescope. If your telescope cannot track at a moving target rate, track siderally, but, again, perform no dithering.

If you believe your night to be photometric, good solar colored standard stars at the begining and end would be helpful, but not essential.