September Science Update
Welcome to the public science update about the ongoings at WTF headquarters. There have been some changes to our observing program with Las Cumbres. We also have been closely working with the LCO team, to improve the robustness and statistical strength of the data pipeline. Lastly, we are in the process of designing a complementary observing program of WTF using Louisiana State University's Highland Road Park Observatory.
Observing Program Updates
On the command end of things, we have made some changes to the telescope scheduler to better improve the data and information we can pull from it. First off, we have requested the observing cadence to visit the target every hour (previously it was every 2-3 hours). When a request like this is submitted, the scheduler then plans ahead the night's observation sequence to optimize all submitted tasks. What this means is that when our star is high enough in the sky to observe, the telescope will point and make the observations. The telescope will then go off and do other things for an hour before returning to our target for the next round of observations. And so on. Whats nice about the LCOGT is that if your scheduled observations dont happen for some reason (weather, technical difficulties, etc), you are not charged for the time lost. Believe it or not, this is not common practice for an observatory - normally if it rains on your observing run, too bad!
The next schedule request change we made was to add another filter in our observing sequence. Observing through a filter allows you to capture a certain color (or wavelength range) of light. This is important because when we catch the star dipping again, we can get a lot more information out of what is blocking the light by using more filters to cover a larger wavelength range. Our previous monitoring was in the "B" and "r” filter — now we have added "i".
You can get an idea of the wavelength range of each filter from this plot (excuse the plot title, I borrowed it from here), where the area under each curve is the *only* light the filter will let pass to the camera:
The LCOGT Data Pipeline and Statistics
In order to properly flag spurious behaviors in comparison stars, modifications have been made to the data pipeline. These changes can largely be summarized as moving away from a weighted average to a mean absolute deviation. Further changes are being made to update adaptability and readability of the reduction pipeline.
In the current version of the software (which is freely available for your review on GitHub), statistical outliers are detected when they surpass the weighted average by three sigma. This sigma is computed from the uncertainties in each individual measurement. As some of you are probably currently yelling at the screen, this technique is prone to accepting outliers as within bounds and can produce weaker results with larger uncertainties. As you can see below, this change has produced some very nice looking results!
So, there's quite a lot to take in here. First, we have datasets from two different locations in the LCO network - Tenerife and Kahului. These are both taken in a blue filter as shown in the first picture. Each data point is marked as a colored triangle, with Tenerife's data generally being on bottom in case colors don't work well for you. The error bars are derived from the camera and sky conditions as well as the brightness of the object it self. You will notice that some of the points are marked with an X. This denotes that there was some anomaly in the behavior of the comparison stars. We need comparison stars to see how our star changed relative to the change all stars experienced (due to atmospheric effects).
At the moment the star is calm. The scatter is due to normal atmospheric variations or calibration. If we were to experience another dimming event, these points would literally fly off of the graph at this scale. For now, we are going to continue to brush up our pipeline and keep producing results like this.
As we are reaching the end of this update, we thought we would close with some exciting news. Tyler is currently testing the university's 20" Highland Road Park observatory for inclusion into the WTF monitoring scheme. This telescope compensates for its smaller size and poorer site with our ability to use it to our heart's content.
With HRPO different types of monitoring are also possible. For example, HRPO can be used for high cadence photometry. We can obtain sufficient signal to noise in a just a few minutes allowing monitoring between the LCO frames. This program is being tested and developed as part of an observational course, but if successful will be adapted into a long term observing plan.
Thank you again for all of your support and interest in our work!
Tyler and Tabetha