As we continue commissioning the PUNCH Observatories, we're knocking off action items day by day and also bringing in the first scientific data sets.
During this time we're all impatient to see the awesome results that are coming. Although it can feel like it is taking forever, we're actually on track to deliver data on schedule. It's worth discussing exactly why the PUNCH commissioning schedule is three months long.
One aspect of PUNCH commissioning is that the spacecraft have to drift apart in their orbits to get into the right configuration.
Another is that we have many items to measure and adjust, to "sight in" the observing sequence for each Observatory and to calibrate the cameras themselves. We need to adjust the pointing precisely to where we want it, and optimize exposure times, and make sure that we understand exactly the position of each polarizing filter wheel. All of that is multiplied times 4 Observatories.
To add to that complexity, we can talk to each spacecraft for just 7-9 minutes at a time as they fly over ground stations. That means everything we do has to be pre-planned days in advance, and shuffled together across the fleet, as we learn how to fly four separate spacecraft and operate four separate instruments, all at the same time.
But even once we get our calibration observations collected, it will take weeks before we'll have our science-grade, background-subtracted images of the solar wind. That's because those images require extraordinary measures, and we have to shake out and tune the ground processing systems – using real data from the spacecraft themselves – to produce our images.
PUNCH data will be perhaps the most photometrically precise images ever collected from a Small Explorer mission. To create them we needed to set up an extraordinary data reduction pipeline. Each step of that pipeline has been tested on simulated data, but now we need to test, and refine, the parameters using the actual data.
Most heliophysics missions download "Level 0" data direct from the spacecraft, and develop those into a "Level 1" photometric data product that goes straight out to the science team. Level 1 data are typically de-streaked if necessary, corrected for linear flat-field effects, set into scientific units, and ... that's about it.
PUNCH Level 1 data are linearized using a novel Bayesian method, despiked, destreaked, subjected to quartic (fourth-order polynomial) flat field correction, stray-light corrected, aligned to better than 0.03 pixel RMS using the background starfield, and corrected to a uniform point-spread function on the celestial sphere. Each of those steps requires careful calibration, testing, and vetting before the pipeline is complete and Level 1 data are ready to go out the door.
But most PUNCH Level 1 data are not useful for our final purpose. Images are combined into mosaics that cover the full field of view. For each four-minute image set, that requires remixing each triplet to three "virtual polarizer" images that are aligned to the final coordinate system, including correcting for something called the IMAX effect, photometrically resampling to a fixed coordinate system, merging all twelve synchronous images onto a single coordinate plane, and marking diffuse transient light sources (aurora) in each image.
If all goes well, we hope to be delivering preliminary Level 2 data by the end of May 2025.
Once Level 2 data are available, then (and only then) can we really begin the final work of peeling back the layers of background F corona and starfield using modern updated variants of the six steps we used on the prototype STEREO data and improved across several years. All of these steps have been refined into a pretty amazing software system that is documented and being written up in a scientific publication.
All of that is to say ... it takes time to make the kind of exquisitely sensitive polarimetric measurements that define PUNCH. The team are doing something extraordinary. Full commissioning requires a tremendous number of steps to be complete, both on the spacecraft and on the ground processing side of the mission.
That's why, even though we've released our first-light images, full image sequences will take a few more weeks to get here. It's planned, we're working on it, and we're doing something, well, extraordinary.
In the mean time, we're planning to do something else unusual: as soon as possible (and hopefully before the end of April) will will begin publishing every image created by a PUNCH camera, to shake out the distribution pipeline. We feel that full transparency is important, and to ensure full reproducibility of PUNCH results we want you (and everyone else) to have access to the commissioning and calibration images also.
So as preliminary data products roll out over the next few weeks, enjoy each new layer of improvement -- but also please be aware that we're working hard on the amazing things still to come!
