predicted source plane morphology being huge

[Xin Wang]

Hi all,

I’ve been modeling RXJ1347 using Lenstool and one of the multiply lensed systems I used is a quintuply imaged system No.9. In previous trial runs, I found out that one of the cluster scale halos should lie somewhere close to the two central BCGs so that the image and source plane rms values for system 9 are quite good. However, as I’m about to finalize the best-fit model with larger number of sampling iterations, I’m always returned with a solution that the cluster-scale DM halo sort of align with image 9.3’s location such that the predicted source plane image of 9.3 is huge (as shown by the green ellipse in the attached plot). In this solution, the chi2 for system 9 is very small, but the image/source plane rms is not as good as in the previous trial runs.

Could I please ask if there exists a way that I can get around this without dropping image 9.3 as a constraint? Thanks!

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

[Johan Richard]

Hi Xin,

This would depend a lot on the actual image shape you use as an input. It seems the model predicts image 9.3 to be undermagnified, and this

does not match your observation. Central images are very sensitive to the inner slope of your mass model you could try to adjust different potentials

such as the gNFW profile (allowing a varying inner slope alpha for the density profile).

Note that you have an uncertainty on the location (and shape) of the soruce from this system. One more robust measurement would be to use the

source barycenter instead, you can find it with the command "bayesImage -b input.par best", input.par being your best model. This will create the

source barycenter position and predicted images of this barycenter, which are the actual images used by lenstool to compute the chi2 / rms. I expect

the shape of this source barycenter will not have the problem.

Johan

--
--
_____________________________________________________________________
LENSTOOL: http://projets.lam.fr/projects/lenstool/wiki
To post to this group, send email to lens...@googlegroups.com
To unsubscribe from this group, send email to lenstool-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Xin Wang]

Hi Johan,

Thanks so much for your reply! I tried with fitting cluster halos with gNFW profile, which indeed helped to avoid the occurrence of this huge source plane ellipse of central images. The value of alpha is constrained to be ~1.5. Although should I adopt the gNFW profile for all cluster member galaxies as well?

I also tried with “bayesImage -b best.par best” but it just created a folder named “images” containing 1e4 files with just one line “#REFERENCE 3 206.8775560 -11.7526250” and nothing else.

Last but not least, do you know if Lenstool can output the magnification estimates for all input arcs on the fly, i.e., in the process of model optimization? I think that’s a very important feature, because that might be a key feature to look at besides the chi2 values (which can be affected quite a lot by the input positional uncertainties), isn’t it?

Thank you very much for your patient guidance! I really appreciate it!

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

On Mar 23, 2017, at 12:47, Johan Richard <nova...@gmail.com> wrote:

Hi Xin,

This would depend a lot on the actual image shape you use as an input. It seems the model predicts image 9.3 to be undermagnified, and this

does not match your observation. Central images are very sensitive to the inner slope of your mass model you could try to adjust different potentials

such as the gNFW profile (allowing a varying inner slope alpha for the density profile).

Note that you have an uncertainty on the location (and shape) of the soruce from this system. One more robust measurement would be to use the

source barycenter instead, you can find it with the command "bayesImage -b input.par best", input.par being your best model. This will create the

source barycenter position and predicted images of this barycenter, which are the actual images used by lenstool to compute the chi2 / rms. I expect

the shape of this source barycenter will not have the problem.

Johan

On Thu, Mar 23, 2017 at 8:18 PM, Xin Wang <albert...@gmail.com> wrote:

Hi all,

I’ve been modeling RXJ1347 using Lenstool and one of the multiply lensed systems I used is a quintuply imaged system No.9. In previous trial runs, I found out that one of the cluster scale halos should lie somewhere close to the two central BCGs so that the image and source plane rms values for system 9 are quite good. However, as I’m about to finalize the best-fit model with larger number of sampling iterations, I’m always returned with a solution that the cluster-scale DM halo sort of align with image 9.3’s location such that the predicted source plane image of 9.3 is huge (as shown by the green ellipse in the attached plot). In this solution, the chi2 for system 9 is very small, but the image/source plane rms is not as good as in the previous trial runs.

Could I please ask if there exists a way that I can get around this without dropping image 9.3 as a constraint? Thanks!

<p2425.png>

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

--
--
_____________________________________________________________________
LENSTOOL: http://projets.lam.fr/projects/lenstool/wiki
To post to this group, send email to lens...@googlegroups.com
To unsubscribe from this group, send email to lenstool-unsubscribe@googlegroups.com
For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
--
_____________________________________________________________________
LENSTOOL: http://projets.lam.fr/projects/lenstool/wiki
To post to this group, send email to lens...@googlegroups.com

To unsubscribe from this group, send email to lenstool-u...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.

To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+u...@googlegroups.com.

[Johan Richard]

Hi,

Sorry I wrote a mistake in the command "bayesImage -b input.par best", input.par should be your *input* model. The bayes.dat should be

in the same directory.

Regarding your other command there is a tool to output the amplification of the images "a posteriori" based on all models

in the bayes.dat. It is called "bayesAmp" in the "utils" directory and again should be used with the input model, the bayes.dat being in the

same directory.

Best,

Johan

[Xin Wang]

Hi Johan,

Thanks for your reply! I tried with $ bayesAmp input.par and got a file showing the *average* amplification for each strongly lensed system (exactly as said in the header of ~/utils/bayesImage.c file). However, it might be more informative to show the amplification of each individual image within a system. Do you think there exists such utility in Lenstool? If not, I’d be happy to hack the code myself. ;)

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

To unsubscribe from this group, send email to lenstool-u...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.

To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+u...@googlegroups.com.

[Johan Richard]

Hi Xin

There are multiple ways to do this directly: either through bayesMap to produce the magnification map around a given image for each model,

or more quickly using bayesImage: from the ratio of the sizes in the image.all and source.dat you have directly access to the amplification.

Johan

[Xin Wang]

Hi Johan,

About the second way, does it make sense to just look at the values of the 8th column in image.all files? Based on the manual, the 8th column corresponds to the quantity “mag”. Is it simply -2.5log(abs(Amp)) at the image's redshift?

Thank you very much for all the help!

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

To unsubscribe from this group, send email to lenstool-u...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.

To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+u...@googlegroups.com.

[Johan Richard]

Hi Xin,

The magnitude (mag) in the 8th column will always be the one from the input catalog so you will only get access to the *relative magnification* between multiple images of the same source. Apart from the size ratios I mention you can also look at the AMP column of the dist.dat files to get the amplification: note that this is a recent addition of bayesImage and you will need to update to the latest svn version. Also it has only been tested on a few cases so let me know if there is any issue.

Best,

Johan

[Xin Wang]

Hi Johan,

Thanks for your reply! On comparing the sizes in source.dat and image.all to infer magnification, the values in 4th and 5th columns represent semi-major and -minor axes. Is there a straightforward way to estimate magnification based on the ratios of these axes, e.g. mu ~ a_img/a_src * b_img/b_src ?

Also could I please ask if there exists a way to compute dist.dat for all the lines in bayes.dat, as what bayesImage does to create image.all? Thanks again!

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

To unsubscribe from this group, send email to lenstool-u...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.

To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+u...@googlegroups.com.

[Johan Richard]

Hi Xin,

Yes, that is what I was refering about, the image/source area ratio is a proxy of the magnification.

But as I mentioned in a previous message bayesImage now produces dist.dat for all the lines in bayes.dat. You just need to update it to the last version. As it is still under testing, let me know if you find anything suspicious.

Best

Johan

[Xin Wang]

oh, I see. I was using version 6.8 but probably not the latest one. Thank you so much for all the fast replies! I really appreciate your help!

Best regards,

Xin Wang
--------------------------------------
Department of Physics and Astronomy
University of California, Los Angeles
CA 90095-1547, USA

To unsubscribe from this group, send email to lenstool-u...@googlegroups.com

For more options, visit this group at http://groups.google.com/group/lenstool?
---
You received this message because you are subscribed to the Google Groups "lenstool" group.

To unsubscribe from this group and stop receiving emails from it, send an email to lenstool+u...@googlegroups.com.