A little video to introduce the team (ft. Maddie Silcock, Robin Thorkildsen, Ivana Grozdanova, Eishwar Pawar and Supervisor David Sobral).

Welcome to the JEDI KENOBI Blog Page.

A Long Time Ago, In A Galaxy Far. Far Away (z~2-6) …

There exists at its centre an SMBH that may be emitting a high concentration of electromagnetic flux in the radio band spectrum, we call these radio sources ‘jets’ or ‘quasars’. The purpose of our NLUAstro Project ; Jet Emission Data Interpretation for Known Enviroment Nuclei Orbiting Blackhole Infrastructure (or more simply JEDI KENOBI), will be to find some of these jets from the COSMOS 2015 data catalogue of over 1 million sources and study their effect on different morphological characteristics within the galaxy.

Weeks 1: The Phantom Jets

“Hey! I’m Eishwar, and I am going to be completing this weeks blog post upto and including week 2. A little bit about my role, I am the Group Coordinator/Project Manager/Code Lead, I will be overseeing progress on the group project and taking the lead when it comes to any coding/scripting we need to do.

During the preparatory week, we first attempted to brainstorm a few ideas about some of the topics that we would be studying, as well as some potential group names. Some of the potential subtopics we found to be interesting were how jet galaxies affect the spectra and morphological characteristics of a galaxy; looking how this also changes with jet intensity; and also looking at the instantaneous growth (how much the Blackholes are currently growing by. We also wanted to have a look at how accretion disk size changes with intensity of jets. This was soon dropped as a potential subtopic we now know that it becomes almost impossible to study accretion disk characteristics of AGN’s. This is because an AGN is extremely bright and outshines any radiation coming from the accretion disk. One last topic we wanted to explore was whether the intensities of x-rays, radio waves and UV waves of AGN’S have any variance with redshift of the AGN/galaxy. However as we had too many subtopics to choose from, we decided that it would be better to include redshift considerations into our data analysis when analysing luminosities and spectra, and that we would not be looking directly at any redshift correlations in depth.

Week 1 consisted mainly of finalising group project details such as aims and goals, group roles and a group name. The final subtopics that we chose were to look at the morphological characteristics (spectra, sfr, mass, age) of rare jet galaxies to common non jet galaxies, as well as to further analyse the correlations between jet intensity of the jet galaxies and these morphological characteristics. As for group roles, it was decided that Eishwar (myself) was to be the Project Coordinator and the code lead for any simulation/data work we needed to accomplish; Maddie was to be the team administrator and the literature lead for research and report writing; Robin was to be the analysis lead for data interpretation and explaining findings; and Ivana was to be the theory lead, in order to link background literature to our findings. However, whilst we have leads for each section, this does not mean they will be working alone. Generally we will help each other with each section as needed, but each lead gets the final call as to how we will be proceeding.  Of the many popular (star wars themed) group project names, we ultimately had to choose one and Jet Emission Data Interpretations of Known Environment Nuclei Orbiting Blackhole Infrastructure –JEDI KENOBI – was chosen.”

Week 2 – Attack of the papers

During week 2, our primary focus was to get a presentation together so that we could explain what we were researching to our peers and to outline our aims and goals for the project. We discovered through reading literature on previous NLUASTRO group project work from BAHAMAS and SHREDS, that the main way of detecting AGN is through the detection of Lya emissions. These groups used data from SC4K and VLA catalogues to identify these jet galaxies. Using this method, once we have identified which galaxies we will be comparing, we can look at radio and X-ray/spectra and colour data and begin work on determining morphological characteristics. In the next two weeks we hope to start looking at the relevant literature on measuring these characteristics such as SFR, Stellar mass and Metallicity, and perhaps begin working with SC4K data to identify some of these jet galaxies using Lya emissions.”

Week 3: Revenge of the AGN’s

“Hello! Welcome to our second blog post, where some details of our third week will be discussed. I am Maddie, and I am the Administrator and Literature lead for my group; like my colleagues I will also dabble in other project areas as the investigation progresses, such as theory and data analysis.  

This week, we were able to finalise some of the aims we need to achieve solid methods for. Consequently, here is a list of some of the week’s aims: 

  1. Identifying AGNs (Active Galactic Nucleis) 
  1. Identifying which AGNs have Jets and which don’t 
  1. A way of knowing the intensity/luminosity variance of Radio Jets 
  1. A way of noting which galaxies DON’T have Jets, but may still have certain levels of Radio, X-ray etc 
  1. A way of extrapolating Star Formation Rate 
  1. Working out the instantaneous growth of Blackholes (this may link to Blackhole Accretion Rate) 

You may be able to tell that these goals centre around obtaining ways of achieving certain aims, rather than doing the specific aims themselves just yet. This is to accommodate for the fact that for some of them we will need more time to achieve, implement code for and so on.  

With this being said, AGNs were still able to be identified, and a plot produced shows their distribution and population: 

Graph 1: SC4K Calhau 2020 Detected Sources

Using the SC4K catalogue, it is possible to see detections of substantial Radio emissions (in the 1.4GHz and 3GHz bands) and X-Ray emissions. Abundance of such emissions are attributive of an Active Galactic Nuclei, and by using such identifiers in the SC4K survey we were able to produce a distribution of sources, highlighting which ones are likely to be AGN’s (i.e. the ones with red squares, and/or blue and green triangles).  

Looking at the raw data and the above plot, we realised some counts are actually results of overlaps. That is, a given AGN can sometimes be observed “twice” due to it having detectable X-Ray and Radio emissions for example. This means we needed to remedy these count overlaps in order to get an accurate count for our total number of AGNs. With the above plot, we were looking at our AGN count to be approximately 394 (using SC4K and Chandra data), however this changed when a new cleaner plot was made; such a plot had the aim of reducing the overlap miscounts, and is shown here: 

Graph 2: SC4K Calhau 2020 Detected Sources: Overlaps factored, our total count comes a bit closer to around 321 AGNs.  


We have decided that for identifying if given AGNs have Jets, we may use the NASA/IPAC Infrared Science Archive COSMOS Cutouts programme (available here: COSMOS Cutouts ( ). This allows us to input relevant RA and DEC coordinates, and then see visuals corresponding to given coordinates with a range of telescopes. The potential of this is quite exciting; as an example, using the VLA (Very Large Array) visual, we are able to see the Radio picture of our sources. Given AGN Jets emit abundances of Radio emissions, we could expect to see two areas of high radio emission corresponding to the two Jets coming off the AGN. Looking at source “SC4K-IA464-75921” from the SC4K survey, a radio visual can be seen: 

As you can see, this AGN appears to have Jets- indicated by the two “blobs” there- these are the lobes of the Radio Jets coming off the AGN. These Jets could be spanning across 100s-1000s of light years, which is really exciting!  

One of our investigation aims of the project was to see if there was a relationship between Radio intensity of AGN Jets, and their host galaxy characteristics. For example, is there a relationship between the Jet intensity of the central AGN and the star formation rate of the host galaxy? This is quite an interesting question, but it’s viability is threatened slightly by the fact that obtaining the Jet intensities is a very difficult thing to do. A variety of literature and papers has been traversed to try and help our exploration of this. One thing that has resulted from it is the discovery of a scaling relationship between Radio Power and Jet Power: 

Pjet ≈ 5.8 × 1043(Pradio/1040) 0.70 erg s−1 

This was found from this paper: [Cavagnolo KW, McNamara BR, Nulsen PEJ, Carilli CL, Jones C, Bîrzan L. A relationship between agn jet power and radio power. Astrophys J. 2010;720(2):1066–72. Available from: C: ( ], and could potentially be chosen as an alternate observable to look at as opposed to intensity. However, our endeavours to find a way of obtaining intensity will continue, and we are making steps to try and learn more about the issue of measuring Radio Jet intensity. For example, at present, an inquiry email has been sent from myself to Astrophysicist Dr Megan Argo, University of Central Lancashire, asking about this issue and if the alternate plan of focusing more on power appears more viable. We hope to be able to comment more on this subtopic focus as time progresses.  

Amongst some of the goals aimed for this week, one that made progress was the investigation into ways to quantify instantaneous growth of Blackholes. With AGNs, the hot gas and matter surrounding the central Blackhole is responsible for a magnitude of X-Ray emissions. The information they provide is therefore very valuable since they come from an area closer to the central Blackhole itself, and will therefore be more sensitive to any changes in this area. It follows that observing “instantaneous” Blackhole growth, parallel to Blackhole Accretion Rate, may be done best by observing certain X-Ray emissions. This notion was supported by one of our plots from the SC4K data, looking at the correlation between X-ray luminosity and X-ray measured Blackhole Accretion Rate. An excerpt of this can be seen here: 

Plotting Xray Luminosity against BHAR, as one is calculated from the other, we should see a straight line fit

Here you can see the direct link between these two observables. Further to this, to test the waters so to speak, another plot was made to see the distribution of X-ray Blackhole Accretion Rate with Star Formation Rate. This plot was again quite an interesting one: 

Premliminary Plotting (Flux against SFR)

At present it is difficult to tell, but the above plot looks as though it may have a very slight positive correlation. If this is well-founded, it may be the case that a more active AGN entails a more active host galaxy. We will need to study this further to make more solid conclusions perhaps. Something of interest to note, are the outliers; for example, the one towards the top left of the previous plot. This data point has an extremely high X-ray flux, but a relatively low Star Formation Rate- a very curiosity-inducing point and something we could investigate more, time-permitting. 

This week has been quite an important one in our journey; we have traversed and explored data sets such as SC4K, LEGA-C and VLA whilst building the paths to our future findings. It has been an exciting time and we look forward to building on our current achieved methods next week!” 

Week 4: A New Hope, for less noise…

Hello there, and welcome back to our blog. My name is Robin, I’m the data and error analysis lead and will be in charge of this week’s blog. 

This week may seem a bit slow in terms of progress, but we have taken important steps towards our goals for this project. We mainly want to focus on AGNs that have jets and how that might change the properties of the galaxies. The first obvious step we need to take is to identify these AGNs from the SC4K data set. This was done last week and left us with 312 unique AGNs. The next step was to identify which ones of these are jet AGNs. 

Like mentioned in last week’s post, we used the COSMOS cutouts from the NASA / IPAC infrared science archive ( and used the VLA (Very Large Array) in order to identify jets. This website lets us insert the RA and DEC coordinates of our AGNs, and by choosing VLA, we get images of the galaxies in the 3GHz radio band. This is perfect as the jet will be visible in this band.  

The process of going through all 312 galaxies was a tedious task since it all had to be done manually. When looking for jets, we are looking for a galaxy with two “blobs” on each side of it which would be our jets. This is also easier said than done, since in the 3GHz range there will be a lot of noise from around the galaxies, and the resolution of the images does not make it any easier. Underneath shows a galaxy we have determined to be a jet, one which is nothing but noise, and one which is currently undetermined (as of writing this) 

As you might see, this task is not as clear cut as it may seem at first. Most of the AGNs do not have jets, which is as expected since jet galaxies are quite rare. Although it might be easy to let the excitement take over and label a galaxy as to have jets, it is important to be critical as a wrongly labeled galaxy might change our conclusions if we are not careful.

In the coming week, we would like to determine all the uncertain galaxies we still have. Since jets are a rare phenomena, we would also like to look at some more AGNs if possible with an extended data set. As of now, we only have about 4 Jet galaxies (the same ones as BAHAMAS). With such a small sample size it would be difficult to draw any sort of conclusion from this. With regards to these more uncertain galaxies, we will discuss and gain the opinions of David Sobral as to what distinguishes between a Jet and Non Jet galaxy in these extremely close cases.

No description available.

Weeks 5: The Catalogue Strike Back

Week  5  has  been  a  week  of  getting  all  of  our  resources  organised.  This  has  been  an interesting  task  since  we  are  working  with  two  different  catalogues,  each  of  which  holds different  information  (are  even  if  they  have  similar  columns,  the  values  might  have been calculated in  a completely different way).  The smaller catalogue we  are working with is  the SC4K where we have 310 AGN (out of the nearly  4000 galaxies in  the  catalogue) of which we  know  the  masses  thanks  to  (Santos  at  al.  2020).  Our  larger  catalogue is  the  VLA catalogue  which  has  significantly  more  sources.  In  order  to  get  information  similar  to  what we  have  for  the  SC4K  catalogue  we  had  to  match  a  table  of  9000+  sources  (taken  from Smolcic et al. 2017) to the larger COSMOS15 survey which had over 1.1 million sources. At the  end  our  matched  table  has  3559  AGNs.  This  means  that  we  have  a  total  of  3869  AGNs which  we  can  study  between  our  two  catalogues.  Unfortunately, neither  one  of  the  VLA tables contained BHARs, which meant that we had to calculate them ourselves. We used the flux  in  the  0.5-10keV  band  (one  of  the  many  wavelengths  the  objects  have  been  observed) and found the BHAR for 529 AGNs with a mean BHAR of    0.1907 solar mass per year (the rest unfortunately do not have provided the flux, but we might be able to get it ourselves for some of them). An  additional  task  for  this  week  was  going  through  some  of  the  objects  that  we  were uncertain about whether they had jets or not. And of course we had to prepare ourselves for the real challenge -looking for jets in the VLA catalogue.


Week 6: Return of the Blobs

During week 6 we focused primarily on detecting jets in our VLA catalogue. Using our matched table from last week, we had roughly 3500 AGNS from redshifts of roughly 0-6. Therefore, we decided to take the same approach as we did in the SC4K data set, as without an advanced AI, it becomes impossible to automate this task in order to find jet galaxies. We started from the top of the VLA catalogue (high redshifts) and worked our way downwards. This is because jets have been known to be found more frequently at higher redshifts. However, one issue we found with this approach was that these jets of galaxies at redshifts >3.5 become increasingly difficult to detect due to negative fluxes, artefacts and the quality of the images due to these vast distances.

We found jets of a galaxy at a high redshift ~ 4 which was initially exciting. However this was just our target galaxy being obscured by one with jets. You can see our target galaxy in the centre, denoted by the yellow circle and our obscuring galaxy to the south with a jet being emitted further south.

As this was such a large task, it took most of the week for members to go over the VLA images in their spare time. By the end of the week, In addition to some of the Bahamas jets we confirmed in our sample, we found two more within the Bahamas redshift range that we thought provided enough evidence to be determined as jet galaxies and another one outside their redshift range.

Three of Bahamas’s Jets were omitted from our jet collection as they either did not qualify as being an AGN as described in Smolcic et al 2017 or were not a Cosmos 2015 detected source. The reason for this is because Jets are produced from AGNs. Therefore, by excluding any data not classified as an AGN, we can save doing unnecessary work, such as going through the images, for these galaxies as the chances of a jet being present are unlikely. As it turned out, there was a detected jet that was not classified as an AGN in the BAHAMAS dataset. This may be due to the constraints as to what is classified as an AGN in the Smolcic catalogue. We are still deciding whether or not to include it in our data set or include it within a separate category. We ended up going down to a redshift of 1.7 as this keeps our sources to a reasonable number given the limited time we have to finish data collection. As we are taking such larger redshift ranges compared to BAHAMAS, we will likely need to take this into account as part of our systematic biases.

Week 7: Rogue One

Hi! This is Maddie again; I’ll be writing this week’s blog post (Week 7) with some updates on things we have been looking into.  

This week, we began trying to plot some of our chosen and calculated variables. We also started with a goal to reclassify all our “uncertain” sources, to specify if we thought they were Jet AGNs or Non-Jet AGNs based on the result of their visual inspection. To do this, we used the NASA/IPAC Infrared Science COSMOS Cutout resource (available here: COSMOS Cutouts ( ) as in previous weeks and reconsidered some of our uncertain sources in order to successfully reclassify them, and therefore include them properly in our later research. 

An interesting source came up, one that had been classified as a Jet AGN by previous groups but not by us! It was SC4K-IA427-26126: 

At first, the two blobs visible there, on the left-hand side, seem to be radio lobes of Jets. However, upon comparing this VLA image with the Hubble image (on the right-hand side), it seems that these blob-like areas of radio emission correspond to sources visible in the optical range too. This is attributable to the notion that these are actually two separate sources, as opposed to two radio lobes of a Jet. We therefore made note to keep this source under review as our investigation progressed, as its visual inspection entailed interesting possibilities for reclassification.  

The above discussion has particular significance in the context of our literature research, as our potential classification of “Non-Jet” for this source contrasts with the conclusions from BAHAMAS last year! We are potentially disagreeing with them, however plan to keep this under review as we will need to see this source in the context of our future plots to help further analyse whether our proposed decision is correct. With all this in mind and taking inspiration from the connotations of our group name, we have colloquially renamed this SC4K-IA427-26216 source: “Rogue One”. 

Changing gears slightly, another focus this week was the formation of plots for our sources. These were the first steps of our plotting journey, and so with the graphs produced this week, we attached to them a future aim to continuously work on them- refining, implementing legends, adjusting ranges, confirming appropriate data sets and so on.  Despite this being the early stages of our plotting, it was still very exciting to see some trends and new behaviours of our sources! Some of our preliminary plots are visible here: 

This graph includes Jet Power (using the equation found in: Cavagnolo KW, McNamara BR, Nulsen PEJ, Carilli CL, Jones C, Bîrzan L. A relationship between agn jet power and radio power. Astrophys J. 2010;720(2):1066–72. Available from: C: ( ) plotted with Redshift, for our catalogue sources (AGNs and Non-AGNs). Remembering this is a preliminary plot, this was done for even our Non-AGN and Non-Jet sources, as mentioned, just to get a first look on what the trends and behaviours might be acting like. 

Another plot produced this week, again concerns Jet Power: 

A note can be made here about the points being plotted. For example, in the above plots, the points visible are the VLA AGNs we identified. It may seem counterintuitive to plot Jet Power for known Non-Jet AGNs too, and you’d be right! At this stage, this occurrence had not yet been fully remedied. It was speculated that plotting Non-Jet AGNs may allow us to predict the existence of Jet-AGNs where the method of visual inspection had given an ambiguous result. Could sources in close proximity to known Jet-AGN areas be predicted to be Jet-AGNs? This was uncertain, but times like these were quite exciting and helped refuel our curiosity! 

With the finish line clearer in sight, we have also started our report writing in our Overleaf report file! We aim to continuously add to this as we go and have already made good progress with our introduction. We hope to be able to also provide updates of this in the future.  

This week had a significant component consisting of coding work; as shown here, we were beginning to enter a stage where we could more abundantly produce plots to make our interpretations from. I therefore hope you enjoy our next blog post too, where we expect to demonstrate more of our plotting progress, and any more side quests we pick up along the way! 

Week 8: The Data Awakens

Welcome back, this will be an update for week 8 written by none other than Robin.  

Week 8 has seen a lot of progress in terms of writing code in order to make some nicer looking plots than what we can get in topcat, and our goals and plans for the project are starting to come into view.  Thanks to Ivana and Maddie, we were able to get some quick first draft plots in order to present some of our work we have done so far. Some examples are the black hole accretion rate calculated in both radio and X-ray band in order to see how they may differ depending on what band the calculation is done. As seen in the figure below we can see most of the values are in the area of 0.1 solar masses per year in the radio, but scattered past one solar mass per year in the X-ray. This also happens the other way, especially for the two VLA jets. (note that not all the jets are on the plot since not all of them have the data needed). 

This week we also made the decision to take away the “jet AGN” label given to SC4K-IA427-26216. This galaxy has long been on our mind (we also mentioned it in the blog post for week 4) as it was also labeled a jet by BAHAMAS. But after some plotting, we kept seeing one lonely jet marked AGN together with all the other AGNs. This point can be seen in the plot above, but also in the plots underneath showing stellar mass and star formation rate against BHAR 

This little rascal needed some investigation, and after looking up which galaxy this was, we were amused to see that it was our good ol’ pall SC4K-IA427-26216. Taking another look at the hubble and VLA images using the cosmos cutouts we did before lead to the final piece of the puzzle. Although a bit difficult to see, when matching the location of the two lobes on the two images we can see that the two larger points were two different galaxies.  

It can be argued that the galaxy could be a jet, by looking at its slight elongation and diffuse edge in the top right of the galaxy. However the resolution is not high enough to know for sure, so instead of removing the galaxy from our jet table entirely we have placed it in “Jet limbo”. It will stay labeled as a jet in order to see if it still deviates from the upcoming plots. It is a part of the council, but we will not grant it the rank of confirmed jet. 

As a final not, Eishwar and myself (Robin) have done the most important work this week by keeping morale up in the only way we know how to during a pandemic: Memes 

P.S. Nothing personal, we really love you BAHAMAS. Your work have helped us immensely 

Week 9 – The Last Programmer

Week 9 is when we finally started seeing the light at the end of the tunnel. After weeks of going through hundreds of images of galaxies, searching through tables with tens of columns and spending days, trying to make a few lines of Python code do what we need it to, we are finally at the point where we can put everything together. Even though we had made graphs of almost everything we needed, they weren’t the final product we desired. This is why, every day from that week we woulda make many small changes to what we had. We would change the colour, size and shape of the points until we were satisfied with how the plots looked. We even tried to incorporate the redshift into graphs of other values using a colour bar and changing the outline of the datapoint, however, this only made the plots more complicated to read.

This week was also when we started to make a significant progress on the report. By Monday Robin had already done a brilliant job with the introduction, so the rest of us worked on the Methods and Results sections. By the end of the week everything was headed in the right direction.

Week 10: The Rise of our Paper