Webb Office Hours Session 5:  April 11, 2024

Q&A's: 

Q1: Where are the Webb Office Hours procedures and guidelines?

A1: Webb Office HoursType your question into the WebEx chat. We will asynchronously copy questions from the chat to this main page and work through them as a group.  If you have images to share please give WebEx permission to share your screen (you may need to log out and log back in again to enable this feature.)


Q2.1: If I do not do anything to mitigate MSA leakage when running the pipeline, how does that impact results? If I'd like to remove it, how do I enable that in the pipeline? 

A2.1: In Figure 2 here: https://jwst-docs.stsci.edu/jwst-near-infrared-spectrograph/nirspec-operations/nirspec-ifu-operations/nirspec-msa-leakage-correction-for-ifu-observations, we show spectra of the 30 virtual slits. Leakage happens in the failed open microshutters, and this figure demonstrates the impact on the data and where the leakage can be seen. Figure 4 shows the clean, final IFU exposure, where the leakage exposure has been subtracted. You won't see the leakage signatures in the final image. The leakcal subtraction in the pipeline is described here: https://jwst-pipeline.readthedocs.io/en/latest/jwst/imprint/description.html. In the calwebb_spec2 pipeline, there is a step to enable imprint_subtract (True or False) and that turns on the leakage subtraction.

Q2.2: Is the correct strategy for mitigating leakage to have a leakage exposure at each dither position as the science dithers, and subtract them from the science exposure for each position? If I only have the leakage exposure for the first dither, is there any way to correct the leakage?

A2.2: The NIRSpec expert checked the program and it does not appear to be an issue that they only have one leakcal exposure for one dither (and not for all dither positions), because, as they explain in the "MSA leakage in IFU data and methods for correction" section of the JDox page, it only results in decreased accuracy of the leakage subtraction; however, since they're dithering, it shouldn't be a significant issue. 

Q2.3: We have two targets that don't have the leakage exposure. How do I reduce the impact from that? It's not a very crowded field but there are other targets nearby. 

A2.3: It shouldn't really be an issue since you are dithering. Dithering is particularly useful for filtering out leakage artifacts. However, if you see leakage signatures, you can do a more careful data reduction. Since you have several dither positions (9), if you can see contamination from leakage, create a cube with just one dither and then create a cube with all combined dither positions, and then compare the two. That will help you understand how significant the issue is.

Q2.4: I tried to combine a final cube using Stage 3 of the pipeline, but there were too many outliers. For my current cube, I'm using the Stage 2 outputs and then sigma-clipping, and then I reproject onto the final WCS. 

A2.4: Outlier detection is tricky for IFU since it's significantly spatially undersampled. Direct sigma-clipping can be risky. The main thing the pipeline outlier detection does for IFU in Stage 3 is that it looks for bad pixels rather than cosmic rays. There are two main sources of outliers: cosmic rays, and bad pixels that weren't flagged in the mask. The current outlier detection step is mostly targeted at bad pixels, with one main exception: there is a known issue where some cosmic rays aren't properly flagged in grouped data (particularly the very strongest and brightest ones), which may be an issue for this program since it uses a grouped readout. That will make its way through to the final data products in an unpleasant way. A fix is currently in progress. If you file a help desk ticket, pipeline experts can explain how to implement the fix manually until it becomes operational. 

Q2.5: In the final image, is the leakage contamination continuum-like? I am showing the result from Stage 1 pipeline and the data cube I have right now. For the continuum emission, how do I figure out what comes from the source and what comes from the leakage? 

A2.5: Only the trace with the leakage will be effected, but it doesn't only effect the continuum. Leakage has a horizontal trace, as well, so there is nothing that prevents it from going through emission line regions. There will be a profile on every single plane, but it also depends on how extended it is on the detector. To help separate the continuum emission from the source, please file a pipeline help desk ticket and we can look at it more closely. 

Q2.6: How accurate is the NIRSpec IFU astrometry? For all my sources, I often have to manually shift them by 0.2-0.3 arcsec. Is there any uncertainty associated with a rotation?

A2.6: This is a common issue. The absolutely accuracy is typically good to 0.2-0.3 arcsec, mainly due to the difficulties of tying such a small field of view to a global world coordinate frame. Depending on which guidestar was selected, it can be off by some amount, so this is not surprising. From everything we've seen so far, it can be effectively described as a shift, and not a rotation.

Q2.7: How do I use the pointing verification image? The WCS for the verification image is wrong in the MAST data.  

A2.7: Please send in a help desk ticket so we can investigate this issue.


Q3.1: Where should I start with the JWST reduction pipelines? Is there a step-by-step guide? I have a program for Cycle 3 that involves NIRCam imaging and NIRSpec MSA. Where should I get started? 

A3.1: We have documentation pages in JDox that give you an overview of the pipeline, along with some examples. All relevant pipeline pages are in the "Data" section of JDox. We have a "Getting Started" page here: https://jwst-docs.stsci.edu/getting-started-with-jwst-data. An overview of the pipeline and links to pipeline notebooks can be found in this section: https://jwst-docs.stsci.edu/jwst-science-calibration-pipeline. These all use the jwst package on Github. Installation instructions are here: https://jwst-docs.stsci.edu/jwst-science-calibration-pipeline#JWSTScienceCalibrationPipeline-installInstallation.

Q3.2: How often do I need specialized processing rather than the default processing for data in MAST?

A3.2: It depends on your science case and mode. For instance, if you look at the NIRCam Calibration Status page, it can help you decide if you should reprocess your data with more specialized or optimized parameters for your science case. Additionally, you can check the NIRCam Imaging Known Issues for more information about pipeline-specific issues.

Q3.3: What level should I start with for reprocessing?

A3.3: It may be best to start with Stage 3 products, and if you are seeing that there are issues (e.g., with image alignment), then you can revisit the Known Issues pages for more guidance on how to fix the specific issues that you might be noticing in the data.

Q3.4: If astrometry is the main objective, what should I check? Should I compare with Gaia to reconcile small deviations? 

A3.4: The NIRCam Imaging Known Issues page has a lot of information and tips for improving image alignment. Sometimes the spacecraft WCS aren't quite perfect for all the data. These pages have guidance for how to modify the pipeline offline to bring an exposure back into alignment. Some of these methods already use Gaia.


Q4: I am working on MIRI/MRS SZ Cha data from PID 1676. I am starting from the _uncal files. When I use the JWST-specific MAST search form, only some of the _uncal files are shown for download (24 files). But when I use the older, generic MAST form, all 48 are present and available to download. Why is this?

A4: In the original version of MAST, it includes uncal files for the background, as well, since they're used to construct the final data product. That should explain the differences you are seeing. 


Q5.1: We have observations of unresolved classical T Tauri stars with dedicated background exposures. All used the 4-point dither optimized for point sources. What is the best procedure for background subtraction, with the context that we are interested in absolute continuum variability? We have tried both 1) subtracting the background image from the science image (but this decreases the SNR), and 2) performing the master background subtraction in the Spec3 stage that takes the 1D background spectrum and uses a model to turn it into the 2D image (doesn’t decrease SNR). However, this causes a weird feature at the edges of the orders making it hard to scale them to one another. Before we could scale them more easily. Is this a byproduct of the different background subtraction method or is it some other change in the JWST 1.14.0 version? I'm using a modified version of the MIRI MRS Data Reduction Notebook (modified specifically for T Tauri stars). Using 1D master background subtraction (not pixel-based). 

A5.1: There may be an issue with how the master background subtraction is being applied to the data. It doesn't appear this would be an aperture correction issue, due to the nature of the feature they're seeing. Please submit a help desk ticket and MIRI experts can investigate more.  

Q5.2: Can you give a quick overview of the aperture correction? 

A5.2: For MRS cubes, the aperture correction is a conical circular aperture correction that depends on the source type. For point sources, it will do a circular aperture extraction whose radius grows with wavelength (corresponding to increasing PSF radius). Then, it subtracts an annular background. The parameters for these steps are in the extract_1d reference files. It applies a scaling factor to deal with what you would have lost based on some PSF model (the default aperture). Those values can be found in the aperture correction reference file.