|
Connection: https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Yohan, Jeff, Amy, Nathalie, Sasha, Gemma, Allison, Roberto, Patricio, Nichol, Adam, Thomas, Leo, Alex, HongLi, Fabrice, Sylvain, Timea, Tapas, Quang, Ant Not attending: Fabien For memory: Minutes from our last telecon Members and their main interests Agenda:
W43-MM2MM3_1mm_cohe_sm_sk06.fits W43-MM2MM3_3mm_cohe_sm_sk05.fits Measurements: original images, PB-corrected W43-MM2MM3_1mm_bsens.fits W43-MM2MM3_1mm_cleanest.fits W43-MM2MM3_3mm_bsens.fits W43-MM2MM3_3mm_cleanest.fits → Final extraction to be done with getsf and GExt2D. Question to Sasha: is it possible to spare time and not measure cores’ parameters on the W43-MM2MM3_1mm_cohe_sm_sk06.fits and W43-MM2MM3_3mm_cohe_sm_sk06.fits maps? Answer: Sure!
Here is the GitHub link : https://github.com/Yohan-Pouteau/FlufiCarp Please only circulate this link between consortium members 
Questions to Alex: → why is the PPMAP coverage smaller than the 1mm ALMA coverage? Some cores fall outside the Tdust map… Answer: image cropped to avoid edge effect, filtering will solve this issue. → why such squared regions? Answer: edge effects as well, not present anymore in the new image → A couple of cores lie over the cold (10K) artefact created when fitting a modified blackbody toward pixels whose the 3mm emission is largely contaminated by free-free emission. Is there a method to ignore the 3mm image when fitting the SED of these pixels? Answer: Alex could substitute a model wo 3mm to the black rounded artefact on the image Adam: To mitigate/remove the free-free emission, we could in the future use the H41alpha image or VLA emission. Question to Fabrice/Nathalie: → Do you confirm the “hot core” nature of the red sources?
Jeff’s presentation from last week. Comments?  * Max recov. scale = ~0.6 x \lambda/Lmin See equation nº6.1 (p53) in https://www.iram.fr/IRAMFR/ARC/documents/cycle1/alma-technical-handbook.pdf
optical thickness to be treated (70 micron, 1mm?)
[This information cannot be used for the 1st mass estimates we’ll make but we can expect publishing more corrected core temperature and thus mass measurements in the future.]
0 Comments
Connection: https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Timea, Nathalie, Amy, Sasha, Adam, Jeff, Mélisse, Roberto, Fabrice, Allison, Thomas, HongLi, Yohan Not attending: Alex, Fabien, Sylvain For memory: Minutes from our last telecon Members and their main interests Agenda:
 * Max recov. scale = ~0.6 x \lambda/Lmin See equation nº6.1 (p53) in https://www.iram.fr/IRAMFR/ARC/documents/cycle1/alma-technical-handbook.pdf
Jeff will circulate his presentation so that everybody can judge if this methodology can be validated and used for our whole dataset. He will apply it to BOLOCAM images.
Connection: https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Sasha, Yohan, Timea, Alex, Sylvain, Nichol, Adam, Leo, Jeff, Tapas, Hongli, Quang, Nathalie, Fabrice, Mélisse, Not attending: Fabien, Roberto For memory: Minutes from our last telecon Members and their main interests Agenda:
by Sasha: spurious sources and reliability of flux measurements On sky250: No major trend/bias but sizes of small sources can have overestimated and those of large sources could be underestimated. On SimAlma240: 20% overestimation of the sizes and below S/N=10, measurements are more spread using the MnGSeg/denoised image. The measurement step done on the original image removes most of the spurious sources (because artefacts are badly measured) and allows a better flux estimate for most sources. → detection on the MnGSeg-BSENS image, measurements on the original BSENS image Ideas for improvement and cautions:
by Jeff: difference maps and wavelet reconstruction residuals
→ We can be confident in extractions done on MnGSeg images.
by Yohan: difference maps
Adjustment of the noise level TBD: Apply filtering scales to all observed images
 * Max recov. scale = ~0.6 x \lambda/Lmin See equation nº6.1 (p53) in https://www.iram.fr/IRAMFR/ARC/documents/cycle1/alma-technical-handbook.pdf
BSENS and CLEANEST are consistent, flat CMF → The temperatures of the 3? hot cores have to be estimated through line emission... Connection: https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Alex, Sasha, Thomas, Adam, Nichol, Leo, HongLi Not attending: Yohan, Jeff, Timea For memory: Minutes from our last telecon Members and their main interests Agenda:
→ 65% more sources (85→139) → common sources (85) should be better measured (from the goodness factor of getsf), fluxes only vary by 10% [TBC]. → up to 6 [TBC with Nathalie] hot cores, for which the CLEANEST flux should be used
→ 25% more sources (139→173) → common sources (139) should be better measured. → fluxes measured on MnGSeg vary by 20% for common sources to 40% for new sources [TBC] Question: shall we trust more fluxes measured on the MnGSeg image? We were thinking to 1. detect sources on MnGseg/BSENS/NoBeamCorrected and 2. Measure on BSENS/BeamCorrected On the synthetic images provided by Fabien (projected numerical simulations by Ntormousi+2019 processed by the ALMA simulator and MnGseg) → 20%-40% more sources (87→ 71 50sec, 133→190 240s), depending on the integration time/sensitivity of the input image. → common sources should be better measured. → 3% (5/190 and 3/87) spurious sources only (versus <1% without MnGseg) → fluxes vary by 5-15% for common sources To Do: Check on the truth table of synthetic sources if fluxes measured on the MnGSeg image are more trustable...
 * Max recov. scale = ~0.6 x \lambda/Lmin See equation nº6.1 (p53) in https://www.iram.fr/IRAMFR/ARC/documents/cycle1/alma-technical-handbook.pdf
--------------------------------------------------------------------------------------
Connection: https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Adam, fab, Gemma, Nathalie, Alex, Tapas, Thomas, Sylvain, Yohan, Jeff, Timea, Roberto Not attending: Sasha For memory: Minutes from last week’s telecon Members and their main interests Agenda:
 * Max recov. scale = ~0.6 x \lambda/Lmin See equation nº6.1 (p53) in https://www.iram.fr/IRAMFR/ARC/documents/cycle1/alma-technical-handbook.pdf
Thomas and Roberto should start investigating how to properly estimate the free-free emission at 1.3mm and 3mm of extracted cores. Who could help them?
--------------------------------------------------------------------------------------
Connection:
https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Yohan, fab, Alex, Ben, Tapas, Nathalie, Patricio, Adam, Thomas, Hongli, Sasha Not attending: Jeff, Gemma, Roberto For memory: Minutes from last week’s telecon Members and their main interests Agenda:
Mass-weighted temperature: from 20K (most of the image) to 35K (compact HII region), should be checked if there is a continuous rather than bimodal temperature distribution.
PACS 70, 160 micron not zero-point corrected yet (use Planck and HIRES) but wouldn't it be easier to filter all images to 5’ beforehand? We are losing the absolute Nh2 values of filaments but it might not be so important after all...
For ATLASGAL: could Timea provide the ATLASGAL images with Planck offsets?
Here are the propositions from Fabien, Yohan and Fred, already discussed with Sasha and Sylvain. Feel free to comment and suggest improvements.
with a detection step slightly different for evolved protoclusters with non pb-corrected for detection and pb-corrected for measurements.
Fred: It will be an issue. As for now detailed CMFs will only be built for young (no HII region) and intermediate (localized compact HII region, where extracted sources are removed afterward). More detailed work involving H41a images will need to be done for evolved regions.
with a detection step different for evolved protoclusters with cleanest used only for measurements with non pb-corrected for detection and pb-corrected for measurements with 12M+7M for tests possibly with MnGSeg denoised maps.
and, except for evolved protoclusters, 12M_B3_BSENS_NOPBCOR
12M_B3_BSENS and 12M_B3_CLEANEST
12M_B6_BSENS_denoised and, except for evolved protoclusters, 12M_B3_BSENS_denoised measurements3: 12M_B6_BSENS_PBCORR and 12M_B3_BSENS_PBCORR
12M+7M_B6_BSENS and, except for evolved protoclusters, 12M+7M_B3_BSENS measurements2: 12M+7M_B6_BSENS_PBCORR and 12M+7M_B3_BSENS_PBCORR
I (Yohan) then plan to develop a small code that will write this small header with universal column names to avoid human intervention. if you plan to use this user friendly procedure (python functions very detailed and simple to use), here are the different steps you will have to do: - obtain core catalogs by running getsf/GExt2D - run my small code to edit a header - run my automatic reading procedure to get your data
comparing the measured values with the truth table. --------------------------------------------------------------------------------------
Connection:
https://zoom.us/j/92270792020?pwd=b1dndW96TkI4ZlNjcWIvbE5MS1lrQT09 Attendees: Fred, Alex, Roberto, Sylvain, Yohan, Sasha, fab, Nathalie, Lee, Thomas, Tapas, Ben, Adam, Not attending: Gemma For memory: Minutes from last week’s telecon Members and their main interests Agenda:
Young: G327 (84 cores), G338 (63 cores), W43-MM2 and MM3 (63 cores), Intermediate: G12 (82 cores), W51E (159 cores) Old: G10 (<100 cores), G333 (<183 cores), W51-IRS2 (<98 cores) The topic of each paper is up to the responsible team since CMF alone might not be interesting enough. Some of us proposed to present MST and mass segregation analyzes, comparisons between ALMA-IMF regions. If you need other datasets (especially lines), please ask the DR team but these papers should rather be “continuum-focused” papers. Here are the paper projects that will develop in the coming 6 months:
Alex: Introducing 7M data could help but artefacts will most probably appear. And computing time will increase with the number of pixels...
Fred: What does comparable bands mean? Inaccuracy should be estimated. Adam: another benchmark for fourier-space combination = feathering: https://github.com/radio-astro-tools/uvcombine/blob/master/uvcombine/uvcombine.py (it is exactly the same thing Alex described; it includes tools for converting units and regridding) Lee: Test of clean on the MM2 field, with and without self-calibration. Strange behavior suggesting that small scales are missing… TBC later!
For ATLASGAL: Timea could provide the ATLASGAL with Planck offset but current data could be more sensitive, TBC For SABOCA, using Herschel + Planck is OK ALMA 1mm and 3mm will most probably remain filtered (MUSTANG? BOLOCAM? TolTec? NIKA2?)
Sasha and Sylvain agree that we should use the Non-pbcorr image for detection to avoid the effect of variable noise and work on a flattened image. We then could use the pbcorr image for measurement or, even simpler, multiply the measured fluxes in the Non-pbcorr image by the PrimaryBeam image itself. The MnGSeg segmentation process produces a noise image (Gaussian part of the structure at the small scales) could give a good estimate of the noise in the map. Sylvain and Sasha could contact Jeff for that if it helps the extraction method. Here are preliminary results of tests. We need volunteers to make others:
At minimum, these images will be use as masks for intermediate regions.
comparing the measured values with the truth table. --------------------------------------------------------------------------------------
Papers dedicated to individual (or a couple of) regions:
Connection:
https://ufl.zoom.us/j/96368629056?pwd=c1JmbWN6Rll0ZkNjYzJZamtXZFdmUT09 Attendees: Fred, Fabien, Alex, Yohan, Sasha, Adam, Gemma, Sylvain, Ben, Nathalie, Manuel, Hongli, Amy, Tapas, Mélanie, Antoine, Estrella, Patricio, Lee, Roberto Not attending: Thomas For memory: Minutes from last week’s telecon Members and their main interests Agenda:
Adam: Tests for W51 could be done using ammonia temperatures… Fred: The goal of Fabien’s paper is to show the integrated CMF of all ALMA-IMF regions and possibly the 3 integrated CMFs for the young, intermediate, evolved regions imaged by ALMA-IMF. It is mandatory to do a more detailed and more correct work on CMF for the regions that have enough cores. The topic of each paper is up to the responsible team since CMF alone might not be very interesting. Some of us proposed to present MST and mass segregation analyses, comparison between ALMA-IMF regions. If you need other datasets, please ask the DR team but these papers should rather be “continuum-focused” papers.
Walker is interested to work on the CMF of G327.
Fred: I’m not sure this will help since it is more an edge effect than a noise effect. Sasha could potentially improve on that.
Decided to use fixed 1.8 beta for first run (variable in future?) Decided to use just 12m for first run (maybe 12m AND 7m bands separate for future) Need to check equinox of ALMA regions (was at 0.1E-4) Before running W43 MM2/3 for final model: * Need to decided following: - Do we degrade 12m to 2"? Fred: Yes to 2.5’’ - Do we use combined 12m and 7m? Fred: In a second run yes. - Do we use just 7m instead? Fred: To be tested... Need beam profiles for all observations (currently missing B3 beam) To run other regions: Need observations from: Herschel (all 5 bands) SABOCA/LABOCA (or ArTeMiS instead of SABOCA) ALMA 12 B6 and B3 (1.3mm and 3mm)* If people are changing flux units and projection before sending to me, then please set everything to MJy/sr and the same projection (preferably RA---TAN, DEC--TAN, easiest to use Herschel images as a reference). --------------------------------------------------------------------------------------
Here are prelimnary results of tests. We need volunteers to make others:
At minimum, these images will be use as masks for intermediate regions.
comparing the measured values with the truth table. --------------------------------------------------------------------------------------
Papers dedicated to individual (or a couple of) regions:
Connection:
https://ufl.zoom.us/j/99224867876?pwd=Y1ZWMWE0QW44aGN4bDI4QmdlZTl4Zz09 Attendees: Fred, Yohan, Sasha, Alex, Adam, Gemma, Fab, Ben, Hongli, Tapas, Patricio, Roberto, Sylvain Not attending: Antoine Agenda:
Note from Sasha: It is absolutely much better to make simultaneous (multi-wavelength) extraction catalogs. Associating two independent catalogs for images with substantially different resolutions is a bad idea. Imagine this: a hires image shows 10 cores and a lores image only has one blob in the same place. Arbitrary association of the blob with just one source from the hires image is the worst idea. It is necessary to deblend the blob into 10 cores using the positions known from the hires image. Otherwise, the measurements will be completely wrong by a very large factor. Matching independent catalogs works only in the simplest (not realistic) case when both images resolve all cores. Comment from Fab: the 1mm and 3mm images have the same angular resolution.
Cleanest for measurement only Roberto said that the 3mm band should not be used for driving the detection for evolved regions... Questions:
comparing the measured values with the truth table.
idea. Much better to use the measurements from a code that gives better accuracy against the truth table of a benchmark and use the other measurement from the second code to assess the total uncertainties of the measured quantities.
https://docs.google.com/document/d/1dxZzeDgzZRRV9oFsjNnNStI79IdXHrp0DnB7X5gvcaU/edit?usp=sharing However, I don’t think it will get to the accuracy level of being able to run core identification algorithms in images with the free-free subtracted pixel-by-pixel. Rather, it can serve as a map to see where there is significant free-free contamination and estimate the free-free flux in selected areas. Fabien proposes to use these images as masks (for intermediate regions). Roberto and Adam say that VLA maps (W51, W43, G351) work better because the emission is optically thin and S/N is larger. We must investigate if extracting free-free flux at the position of cores is more efficient than measuring 1mm/3mm flux ratios to remove the free-free emission from the thermal dust emission of cores. Questions:
Adam and Nathalie said that we need a deep clean on each channel and it is difficult and will take long. Lee: Pure-continuum image is a dream toward hot cores like those in W43-MM2. Question: What else?
Necessary data are:
Papers dedicated to individual (or a couple of) regions:
|
RSS Feed