It may not be completely obvious but flying a successful space mission takes a lot of work. Space is a hostile and difficult to access region. A spacecraft must be carefully designed to do the work desired and then just as carefully constructed and operated. Once you launch, that’s it. No do-overs, no going back to tweak this mirror or that heater panel or any number of things that could be less than perfect when first constructed.

Before you can even start building a spacecraft you have to be sure you understand the job ahead. In the case of the Sentinel Space Telescope, it’s all about searching for and finding near-Earth objects (NEOs). It’s no secret that Ball Aerospace put their best and brightest on the problem as they started their design for Sentinel and they tell us it will get the job done. At the B612 Foundation, we have the utmost confidence in their abilities and the design they came up with but, remember, space is unforgiving. It makes good sense for us to independently analyze the mission design and make sure nothing was missed.

One of the stereotypical views of scientists is that they are a pretty sedate crowd, standing around in their lab coats calmly discussing the fate of the universe. Usually. There is one concept in the realm of searching for NEOs that gets everyone riled up just as much as you’d see between bitter rival sports teams.  I’ve seen this myself more than once. Here you are, talking about your work (like Sentinel) and someone asks, “Well, what about your cadence?”  From there the gloves come off and it can get quite animated—often heated—in the debate that follows.

So what is “cadence,” anyway? This term refers to the pattern and timing of the images you take while searching for NEOs. No one has yet invented the system that can simultaneously look everywhere in the sky to find things that move. That would be really great and would eliminate the concern over cadence. What we have, and this is true for everyone looking for NEOs, is a camera that can look at just a small piece of the sky at one time. Adding to the problem, when you look at the sky it takes a finite amount of time to collect an image and you have to look at the sky multiple times at the same location in order to spot things that are moving. Cadence is a term that refers to the pattern, in location and in time, of your observations of the sky from which you have to discover NEOs.

Simple geometry dictates at least two images to detect a NEO and measure its position and velocity. In principle, you could do your survey with just pairs of images. In practice, almost everyone would agree that this is just too risky. Something could happen to one of the observations and then you have nothing. Most people agree that you need more than two measurements. After that, the arguments start. Thirty years ago most people were happy with three measurements: you can lose one and have the minimum needed. As the community has worked ever harder to find NEOs, we’ve tried to optimize our use of available resources to get more data. When computer software is used to look for NEOs in an automated search, it turns out that you’re better off with five images.

There are many other strategic decisions to make beyond deciding how many images to take.  How long do you wait between images to let the object move? How much of the sky do you survey? What is the pattern of images on the sky? Beyond these questions, there are detailed issues related to the camera and telescope design.

Decisions have been made on all of the questions in the design of the Sentinel Space Telescope. We will have a camera that is sensitive to infrared radiation (think of glowing red coals in a fireplace, 5- to 10-micron light). In one shot, the camera will cover an area that is 2×5 degrees on the sky. Our moon is 0.5 degrees across. Taking one image will require 3 minutes. We will survey the sky that is 80 degrees or more from the sun. We have a term—“observing cycle”—a pattern of observing that covers the available sky enough to see objects four times in the cycle. When you put all of this together, it takes 28 days for one observing cycle. This is the Sentinel cadence.

Here’s a simple example of the data we can get based on this cadence. Consider just a single object that will get picked up a couple of times in the survey. Let’s call the time we see the object first to be at time zero (t=0). In the first observing cycle, we see the object at 0, 1 hour, 2 days, and 2 days plus 1 hour. In the next observing cycle, the pattern is the same, just 28 days later, thus, at 28 days, 28 days plus 1 hour, 30 days, and 30 days plus one hour. This comprises a set of eight observations that span just over 30 days and can be used to estimate the orbit of the NEOs.

I have analyzed this pattern of observation to see just how good our orbit determinations will be.  After all, this is a key goal of the Sentinel project.  Yes, we want to detect NEOs but we don’t have what we need until we know their orbits because that is what tells us if the objects are dangerous or not. This pattern of eight observations I just described does a pretty good job of determining the orbit. However, an orbit based on eight observations is good enough to weed out most objects from further consideration while also letting us find them again years or decades later if we’re interested in them. A small number of objects will come out of this as objects of interest that we’ll want to study more.

Let’s hope that this all sounds reasonable but don’t forget that the observing cadence gets the NEO community pretty worked up. This is where SODA comes in. SODA stands for Sentinel Operations and Data Architecture and is a working group comprised of a top-notch collection of researchers active in searching for and studying NEOs. Their first task is to provide input on cadence and review my analyses of Sentinel performance. They have provided very useful suggestions on the analysis and provide a sounding board to discuss my modeling tools and the conclusions we draw from them. This process serves double duty ensuring that members of the scientific community are kept informed and have input into in the design process for Sentinel. In the end, we hope to build consensus on the merits of the Sentinel Mission and bolster everyone’s confidence in the project.

One example of my simulation tools is shown in this video.

Shown here is a top-down view of the inner solar system. The yellow circle in the center is the Sun.  The open blue circle shows the orbit of the Earth. The heavy green dot shows the location of Sentinel, remember it is orbiting the Sun inside the Earth’s orbit, looking out away from the Sun. The small dots show the position of a NEO that has been detected by Sentinel. Each time step in the video corresponds to one observing cycle.

Those NEOs that are seen in that time step are shown as white dots. If they aren’t white they were not seen in that time step but the color of the dot shows its past observing history. A red dot means the object was only seen once (that’s not great but better than zero). A yellow dot means the object was seen twice.  This is the minimum to be called a successful discovery by Sentinel. A green dot means it has been seen in three or more observing cycles. These will have very good orbits. In the upper right you see a completion gauge. If we find 100% of the objects three times or more this will become completely green. As the survey progresses you see the tally for that cycle and all that came before, keeping track of those seen one time (red), two times (yellow), and three or more times (green). Below the gauge you see the elapsed time of the survey in years.

This particular simulation was run with 20,000 individual objects all with a diameter of 140 meters. If you watch closely you might be able to follow a single object as it orbits the sun.  You’ll notice as time passes that there are lots of objects throughout the plot and most are green. But, keep in mind that they’re only seen when they pass within the white zone. Most of the time these objects can’t been seen by Sentinel because they are too far away. For this particular case, we get about 75% of the objects by the end of the survey.  What about the rest? Well, most of these objects take too long to orbit the Sun and don’t get close enough to be seen during the 6.5 year survey. If you run the survey for longer you’ll get even more.

With tools like this we can understand just how well Sentinel works and can investigate different observing strategy and system designs. The goal, of course, is to find as many as possible. As with everything, there are compromises to be made between cost and capability and these calculations help us make the right decisions. Once Sentinel is operating, we can also use these tools to evaluate how well the observatory is performing.

You can read more about Marc in this Q&A.