Galaxy Zoo: Clump Scout II

Galaxy Zoo: Clump Scout I

The initial results of the first Galaxy Zoo: Clump Scout project were based on your Zooniverse volunteers' annotations of images from the Sloan Digital Sky Survey (SDSS) Legacy Survey. Since then, our team has been putting your clump markings to work on several other projects.

Training a Deep Learning Model

Our first step was to use your annotations from Galaxy Zoo: Clump Scout to train a deep learning model that we could use to search 240,000 galaxies from SDSS survey for giant star-forming clumps. Our model found 41,445 clumps distributed among 34,246 of the galaxies we searched.

You can find a paper that describes the development and testing of this model on ArXiv at: https://arxiv.org/abs/2312.03503.

The clump formation mechanism and the environments of clumpy galaxies

Giant star-forming clumps can be divided into two broad categories. In-situ clumps are clumps that form within their host galaxies via the gravitational collapse of huge, cold, (mainly) hydrogen gas to form hundreds of millions of stars in a relatively short period of time. On the other hand, ex-situ clumps are thought to be the remnants of other galaxies that have merged with the clumpy galaxy at some point in the recent past.

Using our expanded clumpy galaxy catalogue, and the galaxies from our original test set that were found not to be clumpy, we measured the fraction of galaxies that appear clumpy in dense environments (where the average galaxy-to-galaxy separation is small) and in sparse environments (where the average galaxy-to-galaxy separation is large). This measurement allowed us to test whether clumps in our catalogue are primarily formed in-situ or whether ex-situ clumps comprise a large or even dominant fraction of their population.

How does measuring clumpiness as a function of environment tell us about the clump formation mechanism? Well, as you read earlier, the formation of in-situ clumps requires large reservoirs of cold gas to be present in a galaxy. The gas needs to be cold so that motion of its constituent molecules is relatively slow. In hot gasses the motion of the molecules is very fast and this stops the gas from collapsing under gravity to form stars.

In dense environments, interactions between galaxies, as well as energy produced within those galaxies, heats up the gas in the intergalactic medium. When that gas falls onto a galaxy it is too hot to collapse quickly to form stars and clumps. On the other hand, the intergalactic gas in relatively sparse environments is much cooler and when it falls onto galaxies in those sparse environments, it can collapse to form clumps. Putting all this together, we expect that in-situ clump formation is more likely to happen in sparse environments.

This would not be the case for ex-situ clumps. These form outside the galaxy and their presence do not depend on the availability of cold gas in or around the clumpy galaxy. Ex-situ clumps should be more common in denser environments where the galaxy merger events that produce them are more common.

If we found that clumps are much more prevalent in sparse environments, this would be good evidence that they were mainly formed in-situ. If we found that the fraction of clumpy galaxies was higher in dense regions, this would suggest a significant population of ex-situ clumps.

So what did our results reveal? When taken at face value, they show that the fraction of clumpy galaxies is higher in more sparse, lower density, environments. But how should this be interpreted? Does it mean that most clumpy galaxies in our catalogue are dominated by in-situ clumps or is there something more subtle happening?

In fact, the answer is probably "both". Here's why. For any galaxy, one can calculate the ratio of the rate at which that galaxy is are forming stars and the total mass of the stars it already contains (this ratio is called the "specific star formation rate" or sSFR for short). Now, it turns out that when we group our clumpy and non-clumpy text galaxies into subsets with similar sSFR values, the dependence of clumpiness on environment vanishes. A galaxy in a particular sSFR subset is just as likely to be clumpy in a dense environment as it is in a sparse environment.

The conclusion is that it seems to be the sSFR of a galaxy that directly determines how likely it is to be clumpy or not, rather than its environment. Galaxies with higher sSFRs are more likely to be clumpy. The reason that we initially found a higher fraction of clumpy galaxies in sparse environments is because the average sSFR of galaxies in sparse environments is higher. It's a perfect example of the adage that "correlation is not causation"!

So how does this affect the in-situ versus ex-situ question? Consider what we would expect to find if the population of ex-situ clumps was large or even dominant. In that case, we would expect to find a greater prevalence of clumpy galaxies in denser environments, regardless of the test galaxies' sSFRs because the prevalence of ex-situ clumps depends on the rate of galaxy mergers, which does not depend on the sSFRs of the merging galaxies. The fact that we do not observe any dependence of clumpiness on environment (and a preference for sparse environments when we do not control for sSFR properly) strongly suggests that the number of ex-situ clumps in or clump catalogue is much smaller than then number of in-situ clumps. Moreover, the processes that lead to clump formation are likely to be extreme examples of the processes that drive ordinary star formation so it makes sense that as the sSFR increases and the star formation in a galaxy becomes more intense, the prevalence of giant star-forming clumps should increase.

You can find a paper that describes this research in much more detail at: https://iopscience.iop.org/article/10.3847/1538-4357/ad7119.