SPHEREx Data Explorer: Spectrophotometry Tool

Contents of page/chapter:
+Introduction
+Background
+Fitting Modes
+Data Inputs
+Implementation Overview
+Outputs
+Initiating a Process <-- Jump here to learn more about starting a job
+Job Monitor
+Results
+Exploring Bitflags
+Saving Results
+After SPHEREx
+Different Behaviors: Loading from the Job Monitor, Pinning, and Loading from Disk
+Tips for Success

Introduction

The Tractor

The tool supports two models for the intrinsic source morphology (i.e., the source structure before convolution with instrumental effects or the SPHEREx PSF):

• Point Source
• Sérsic Galaxy Profile

If a source of interest is known to have a close companion, they can be fit simultaneously with Tractor by providing the coordinates of both sources. This approach can mitigate the impact of blending on the recovered flux of the source of interest.

Background

The Tractor

Fitting Modes

Point Source

The intrinsic source is treated as a delta function located at the user-specified position.

Sérsic Galaxy Profile (or Sérsic Profile )

The intrinsic source is modeled as an extended galaxy. Users must provide the following fixed parameters:

• Sérsic index (n): 0.5 ≤ n ≤ 6. While some galaxies may be well-characterized by more extreme values, the allowed range includes the values over which the tool has been tested.
• Minor-to-major axis ratio (b/a): 0 < b/a ≤ 1.
• Position angle (PA): -360 degrees ≤ PA ≤ 360 degrees, measured east of north.
• Effective radius (Re): 1 ≤ Re < 20 arcsec. Values below 1 arcsec are interpreted as a point source.

Data Inputs

The primary data come from the SPHEREx Quick Release Spectral Images (version 2), which are multi-extension FITS files with the following extensions:

• IMAGE: Per-pixel flux in MJy/sr.
• FLAGS: Pixel-level flags indicating issues such as saturation or strong non-linearity, as well as pixels containing flux from known sources so they can be masked when estimating the local background.
• VARIANCE: Per-pixel uncertainty estimates corresponding to the spectral image.
• PSF: The high-resolution, position-dependent point spread function for the corresponding detector region.
• ZODI: (Not used by the tool.)

These inputs, together with the user-specified intrinsic morphology parameters, allow the tool to accurately fit source fluxes and propagate relevant flags to the output catalog.

Implementation Overview

For each source, the tool:

• Extracts a cutout from the IMAGE extension of the SPHEREx Quick Release Spectral Image, centered on the source (per-pixel flux in MJy/sr).
• Computes and removes a local background from nearby pixels, masking pixels flagged in the FLAGS extension as containing flux from known sources.
• Retrieves the position-dependent PSF from the PSF extension for the cutout region.
• Uses Tractor to fit a generative model combining the source's intrinsic morphology and the local PSF, adjusting only the source flux amplitude. The fit uses per-pixel uncertainties from the VARIANCE extension.
• Converts fitted fluxes to μJy using the Pixel Solid Angle map.
• Propagates pixel-level flags from the FLAGS extension to the output catalog and applies additional photometry-specific flags for issues such as strong non-linearity or saturation.
• Computes a fit quality metric based on the residuals between the model and the spectral image within a 2.5-pixel radius aperture; values near 1 indicate good fits.

Outputs

Core Spectral Columns

These provide the fundamental science spectrum.

• wavelength (μm) -- Wavelength at which photometry was measured.
• flux (μJy) -- Source flux density.
• flux_err (μJy) -- Uncertainty on the flux measurement.
• lambda_width (μm)-- Effective bandpass width associated with the wavelength measurement.

Quality Assessment Columns

These help you evaluate the reliability of each flux measurement.

• flags -- This column contains an integer bitmask value representing issues that affected the photometry measurement. Each bit in the value corresponds to a specific condition, and multiple conditions are represented as the sum of their corresponding bit values. To determine which conditions apply, convert the flag value in the output to binary and examine the bits that are set. Detailed definitions of each bit are provided below and in the Explanatory Supplement, available from the SPHEREx Mission Page.
• fit_ql -- Photometry quality metric. For each source, a circular region of radius 2.5 SPHEREx pixels around its subpixel position is used to compute:
  fit_quality = < |p_k - m_k| / sigma_k >
  where p_k is the pixel value in the Level 3 image, m_k is the Tractor model value, and sigma_k is the per-pixel uncertainty from the variance map. Values near 1 indicate good fits; larger values suggest potential problems.
• flux_bkg -- Background flux estimate used in the fit.

Provenance Columns

These identify the exposures and detectors that contributed to each spectral point.

• lvf_id -- ID of the LVF exposure used for photometry.
• det_id -- Detector ID.
• deep_flg -- True if the measurement is from a Deep Field.
• mjd -- Modified Julian Date of the exposure.

Inspection Columns

These columns indicate where the requested position falls in each overlapping spectral image.

• x_image, y_image -- Pixel coordinates of the requested position within the spectral image used for the photometry measurement, derived from the image WCS.
• ra, dec -- Sky coordinates corresponding to the same position.

Notes on SPHEREx Photometry Flags

Bit-level flags are set in the images and are propagated through the spectrophotometric process to indicate if the spectrophotometric measurement may be affected by any pixel-level flags. If any pixel in the constituent images is flagged, then that flag is carried over into the resultant spectrophotometric measurement flag. The number of impacted pixels is currently not returned.

(*) for full definition and discussion, see the Explanatory Supplement on the SPHEREx Mission Page

See Exploring Bitflags below for how to explore bitflags using this tool.

Some bits are not listed here; those bits are used in products not publicly released.

The tool keeps track of whether the photometry included any "bad" pixels (defined by the list above) using the CONTAINS_BAD_PIXEL (bit 32) flag. This provides a convenient way of selecting down to only measurements unaffected by known problematic effects. In other words, if the bitflag is a very large number, 4294967296 or more, then that point is problematic.

The existence of any of these flags in the measurement should be considered a warning when using / interpreting the point, and do not necessarily mean that the measurement is unusable. However, there are two exceptions. The first is DICHROIC (bit 7). Currently the calibration in the dichroic region is not perfect, and users would be advised to be skeptical of these measurements until further refinement of the absolute calibration. The second exception is FIT_ERROR (bit 33), indicating a failure to converge.

Pixels with any of the following flags will be ignored when fitting with Tractor:

• SUR_ERROR
• NONFUNC
• MISSING_DATA
• HOT
• COLD
• NONLINEAR
• PERSIST

Initiating a Process

Just like in a Spectral Image Search, you have a HiPS image loaded that takes up most of the browser area. Overlaid on top of the HiPS image, there is a MOC indicating the sky coverage for the data currently available in the archive, and one for the deep fields coverage. As you zoom in, the shaded regions become more transparent, or you can change them in the layers pop-up.

Again, just like in a Spectral Image Search, you can enter a single target or upload a list of targets; please see that section for details of how to do that.

And, just like in a Spectral Image Search, these spectrophotometry jobs are managed in the Job Monitor. If you are submitting a lot of jobs, keeping track of which job is which in list in the Job Monitor can be difficult. Using the "Title" field near the bottom of the spectrophotometry window, you can change the title by which the search will be listed in the Job Monitor. You don't have to change it from the default to be able to submit jobs, however. (Once you change it, though, all subsequent jobs you make will still have that same title unless you change it each time.)

Point Sources

As discussed above, one of the two modes the tool supports is point sources. You specify the necessary parameters here. At minimum, you just provide the location(s), and let it take the rest of the default parameter values; the defaults work fairly well. ⚠ Tips and Troubleshooting Parameters: The tool doesn't recenter the position(s), so be sure you are providing a good location(s). Nearby sources: If there is a nearby, potentially confused source (recall SPHEREx's pixels are ~6 arcsec; see the Explanatory Supplement on the SPHEREx Mission Page), you need to provide the tool the locations of all of the sources you want to deconvolve. Upload a list of targets with all of the sources included. However, if there are sources within the 15 px box that the tool uses for background estimation, you do not need to specify those; the tool knows about bright sources all over the sky. (See the Explanatory Supplement on the SPHEREx Mission Page) Selecting all wavelengths or no wavelengths has the same effect -- it will attempt to provide all the data it can find. If you select just one (or a few) range(s) of wavelengths, it will do only what you request. Run time: The more frames the tool has to work through, the longer the process will take. If you ask for a source near the ecliptic poles, where there might be at least one image per orbit, it will take a long time for the tool to work through all the images to get a spectrum. If you ask for a source near the ecliptic plane, where there are far fewer images, the process will complete relatively quickly. If you submit a list of targets all over the sky, it will take a long time for the process to run. If you submit a list of targets close together on the sky, the tool is clever enough to only pull the images it needs once, and obtains fluxes for all the sources at the same time. A back-of-the-envelope calculation for run time is as follows -- do an image search to figure out how many images cover your target. Then, to obtain the estimated run time in seconds, calculate 1.67*number images + 50; for run time in hours, calculate 0.000463*number images + 0.013. Throttles: To allow computing resources to be shared fairly among all users, there are built-in throttles. You are currently allowed a maximum of two spectrophotometry jobs running simultaneously. Additional jobs will be queued and run when resources are available. Lists of targets are currently limited to 20 sources. Additional tips for success are below.

Sérsic Galaxy Profile

As discussed above, one of the two modes the tool supports is Sérsic Galaxy Profile. You MUST specify all the necessary parameters here. You can submit a list of targets here too; if you upload a catalog, there should be columns for all of the necessary Sérsic parameters, and you will be asked to indicate which column from your input catalog should be used for these values. ⚠ Tips and Troubleshooting Parameters: The tool only fits the flux; it takes from your input the position, the Sérsic index, the major/minor axis ratio, position angle, and effective radius. Be sure you are providing good values for these parameters. Selecting all wavelengths or no wavelengths has the same effect -- it will attempt to provide all the data it can find. If you select just one (or a few) range(s) of wavelengths, it will do only what you request. Run time: The more frames the tool has to work through, the longer the process will take. If you ask for a source near the ecliptic poles, where there might be at least one image per orbit, it will take a long time for the tool to work through all the images to get a spectrum. If you ask for a source near the ecliptic plane, where there are fewer frames, the process will complete relatively quickly. If you submit a list of targets all over the sky, it will take a long time for the process to run. If you submit a list of targets close together on the sky, the tool is clever enough to only pull the images it needs once, and obtains fluxes for all the sources at the same time. Throttles: To allow computing resources to be shared fairly among all users, there are built-in throttles. You are currently allowed a maximum of two spectrophotometry jobs running simultaneously. Additional jobs will be queued and run when resources are available. Lists of targets are currently limited to 20 sources. Additional tips for success are below.

Multiple targets

Once you select "Multi-object", if you haven't yet uploaded a table, you get the upload pop-up window, from which you can upload a file. That upload section contains more information about file formats, etc.

If you have columns "ra" and "dec", the upload tool may be able to guess that those are the position columns. If it can't guess the position columns, or it guesses incorrectly, then you can select from the uploaded columns by selecting the magnifying glass and choosing the correct column from the available columns it provides. You can also choose to select a target label from your uploaded table; this is optional. The "optional parameters" for multiple targets are the same as for single targets, and apply equally to all of the uploaded targets. ⚠ Tips and Troubleshooting If your list has more than 20 sources, as the one in the screenshot here does, then the tool will only take the first 20 sources. Break your list into sets of 20 to submit them. If you are submitting a list for Sérsic Galaxy Profile fits, you should provide additional columns for the Sérsic parameters; just like for the ra and dec columns, you can teach the tool which columns to use for these parameters. Related to the prior item: You can't upload a list of targets and pick the same Sérsic parameters from the target submission screen for all the targets at once. Your uploaded catalog has to have values for each of the Sérsic parameters.

Job Monitor

Click on "Send to Background", and the job monitor will take over management of the job; you can then continue to work in the tool while you wait. See the Job Monitor section of the Downloads chapter for more information.

Results

Here are the results of a spectrophotometry run on a single target:

On the left is a list of all the requested targets (here, only one), and on the right, the "Data" tab is in the foreground with the extracted spectrum plotted. Across the bottom of the screen are cutouts from the constituent images, centered on the target.

The plotted spectrum has some properties like that of generic plots in this tool, but because the tool realizes that it is a spectrum, it has some properties that are specific to spectra. Moreover, if you click on the table/spectrum toggle in the upper left of the right hand pane, you can view the spectrum as a table. Then you can interact with the spectrum as you can with any other table in this tool.

All the columns that the tool produces are defined above.

The images at the bottom and the plot (or table) are linked. If you click on a point in the plot, the cutouts are updated to reflect your selection. In general, the currently selected point is the center cutout in the display. From the row immediately above the images, you can set the number of images shown at the same time via radio buttons (), set the size of the cutouts () where the default is 1.2 arcmin, and control whether or not the cutouts are linked together by WCS () where they are linked together by default. The image toolbar works here just as for all other images in this tool.

This is what the results look like for a submitted list of targets -- very similar to that for a single target, but with multiple targets on the left.

Exploring Bitflags

Look at the flags column

You can also change the color of the points in the spectrum plot to correspond to the flags column -- this example shows how to change the plot point colors to correspond to the dates of the observation, but you can do the same sort of thing to make the colors correspond to the flags column. Then, as you move your mouse over each point in the plot, the pop-up will tell you the value of the correpsonding flags value. (The string "#this" is a strange-looking IVOA term that means "the main data product.")

Look for the dichroic flag

As described above, you can view the spectrum as a plot or a table. Change your view of the spectrum to a table and look for the column called "obs_publisher_did." This is the field that the tool uses to go find the images it shows at the bottom of the screen -- it is basically a link to the images. But the images have the band in the filename! Turn on the filters if they're not on already, and add a filter for the strings "D3" or "D4" in obs_publisher_did, e.g.,

like '%D3%' or like '%D4%'

Look at the images

As described above, the tool pulls the cutouts for each of the constituent points in the spectrum and shows them to you along the bottom of the window. As described in the Visualization chapter, you can control the layers that are shown on your images. At the bottom of the layers pop-up, you can enable the mask layer. Do it, and turn on all of them ("show all") but then turn off bit 21 ("Source"). You'll be left with color-coded indications of which pixels might be problematic:

In this example, the selected row has flags = 4297064449, suggesting that something might be questionable about this point. The corresponding image is the one outlined in brown at the bottom. Indeed, one of the points involved in the source appears to have been affected by a something flagged as a transient -- the red pixel.

In this fashion, you can then decide on a case-by-case basis what to investigate in more detail.

Decode the flags

Change your view of the spectrum to a table. Following the add column instructions, add a new column to the spectrum table. Call the new column something useful like BITN, where "N" is the number of the bit you are exploring. For the expression, enter:

MOD(FLOOR("flags" / POWER(2, N)), 2) = 1

The new column will be 1 if that bit is set, and 0 if it is not. You can create as many new columns as you need to explore which bits are set.

Investigate background fluctuations

Export it and explore with Python

The existence of any of these flagged pixels in the photometry measurement for a source can be tested using the "flags" field -- for example, to see if a source measurement included any pixels flagged dichroic, do:

dichroic = 1 << 7
phot_flags = phot_result["flags"]
idx_dichroic = np.where(phot_flags & dichroic)[0]

Saving Results

Saving from the left pane

(This is the method that is just like the Spectral Image Search results as well as most like what you find in other IRSA tools with this look and feel.) From the Results tab, in the left pane, just like with the Spectral Image Search or other IRSA tools, select the rows you want to download, and then select "Prepare Download." From there, it's the same as other downloads here. This is the best way to save the data. Note that the filenames of the downloaded files are of the form [ra]_[dec].xml but they are actually *.vot (VO tables), which is a subcategory of XML files. (Note also that if you use the diskette to save this table, you are only saving the table itself, not the data that spring forth from this table, e.g., the spectra.)

Saving the plot

(This works just like saving any other plot.) Click on the diskette icon in the upper right of the plot pane to save the plot.

Saving the table from the plot

(This is somewhat more obscure than the above options, but works just like saving any other table.) In the top left of the plot, use the toggle that allows you to see the spectrum as a table. Click on the diskette icon in the upper right of this new table pane to save this spectrum table.

Saving the results from the Job Monitor

(This is much more obscure than the above options, but works in a pinch if for some reason the results can't be loaded into the tool.) Go to the Job Monitor tab, and click on the info icon (). From that pop-up, click on 'Results' to expand that section, then click on the link or click on the clipboard icon to copy the link and paste it into a new browser. Then you can save the file that is loaded into your browser (probably your browser's "File" menu then "Save as"). If you save the file this way, it arrives in an xml vot format (and you can't control that format; this is sort of an emergency save).

⚠ Tips and Troubleshooting

• The Job Monitor is supposed to hold on to your jobs for 14 days, but in testing we have found some transient gremlins suggesting that (a) you should download the data as soon as you can; (b) logging in before submitting your jobs (and before attempting to download the results) seems to keep the gremlins at bay for longer.
• IRSA tends to have scheduled blocktimes every other Tuesday morning Pacific time. Those blocktimes may take down our servers and terminate running jobs and/or lose old jobs. We are working on ways to ameliorate this, but, again, it's probably wise to download your data as soon as possible.
• This spectrophotometry process takes a long time, with a lot of disk i/o. Lots of people are accessing IRSA's servers in general, some of whom in recent months have been malicious. We are working as fast as we can to block the malicious and shift more resources to SPHEREx; we thank you for your patience. You can also consider using SPHEREx data in the cloud . (There are Python Notebook tutorials .)

After the SPHEREx Data Explorer

Another tools that can read any of the files written by the SPHEREx Data Explorer is IRSA Viewer , which also has special capabilities for handling spectra. Currently, some additional capabilities are available in IRSA Viewer which are not present in this SPHEREx tool. However, IRSA Viewer doesn't have all of the SPHEREx-specific capabilities of this SPHEREx tool.

⚠ Tips and Troubleshooting

• If you're in a situation where you've saved your spectrophotometry results as an XML or VOT file, but you don't have any software ready to read that, you can use IRSA Viewer to convert the XML or VOT file into something else, like tbl or FITS format.

Different Behaviors: Loading from the Job Monitor, Pinning, and Loading from Disk

If you just let the tool run and don't explicitly put the spectrophotometry job in the background, the results will be automatically loaded into the tool, complete with cutout images.

If you put the spectrophotometry job in the background, once it finishes, you'll have to click on the little green plot icon in the job monitor to load the results into the tool, and the results will be loaded into the tool, complete with cutout images.

If you come back to the same browser session on the same computer anytime within the same few days, you can still click on the same little green plot icon in the job monitor to load the results into the tool, and the results will be automatically loaded into the tool, complete with cutout images.

When you load the results this way, the linkage to the images is preserved, and the tool understands how to get the images.

Potential confusion #1.

If you have requested a set of <20 sources, then you will get the list of sources on the left and the spectrum corresponding to each object on the right. Or, if you have multiple single spectrophotometry results loaded in different tabs on the left, as you change what you have in the foreground on the left, the spectrum on the right changes.

Each time you change what you've clicked on the left, it reloads the entire data product on the right. So, if you impose a filter on the plot or the data table, that filter is removed when you go back and reload the data product. In order for the filter to "stick", you need to first pin the spectrum (), and then filter the table associated with that pinned spectrum. Then the filter persists even if you click away from that spectrum and then come back to it. However, pinning breaks the linkage to the images -- you no longer have the direct link between the spectrum and the cutouts.

Potential confusion #2.

If you want to save the spectrum, in general, you need to use the "prepare download" button, not the diskette icon. The diskette icon usually saves just the table you're looking at. If you are looking at the data table corresponding to the spectrum, you're fine using the diskette, but if you're looking at a table that lists your source(s), it will save that table and it won't save your spectrum/a.

Potential confusion #3.

If you download your spectrum, it's most likely going to save as an XML file. If you turn right around and upload it back into this tool via the upload tab, it will behave as if you have pinned the spectrum -- that is, the linkage to the images is lost. However, it will understand how to interpret the spectrum and treat the errors, etc. properly.

For a spectrophotometry job loaded directly from the job monitor, the "data" tab has the spectrum and "active chart" has a boring ra/dec plot (with a single point). For spectrophotometry results loaded from an xml file, "active chart" has the spectrum. "Data" may attempt to grab one image, but will not know how to grab all of the images corresponding to all of the images that went into the spectrum.

If you upload the XML file into IRSA Viewer (the more generic IRSA tool with this look and feel), you have to warn it that the file contains a spectrum, and then "active chart" plots the spectrum. It doesn't have any idea how to retrieve the images, though it behaves as if it is trying to do so.

Tips for Success

SPHEREx Mission Page

Parameters:

• The tool only fits the flux; it takes all the other inputs from you.
• The tool doesn't recenter the position(s), so be sure you are providing a good location(s). Did your source move between its epoch and equinox (is it a high proper motion source)?
• For galaxies, you need to provide not only the position, but also the Sérsic index, the major/minor axis ratio, position angle, and effective radius. Again, it doesn't fit those, so you need to provide good values for these parameters.
• If you upload a catalog, there should be columns for all of the necessary Sérsic parameters, and you will be asked to indicate which column from your input catalog should be used for these values.
• For single sources, there is error trapping when you input parameters, but when you upload a list of targets, there isn't error trapping. If you ask it to calculate fits where the Sérsic parameters aren't physical, it will still attempt to do this.

The drill is not a spectrophotometry run.

• The Mosaic Tool is the newest SPHEREx tool, and it runs to completion much more quickly than the Spectrophotometry Tool. It generates a FITS cube, where each plane is a different wavelength. You can use the extraction tools to very quickly drill through the cube to generate a spectrum-like product.
• The drill extraction tool is NOT THE SAME as the spectrophotometry tool! The drill extraction tool provides a quick look through the cube, intended for exploratory use, and does not yield a research-ready source flux spectrum.
• The Spectrophotometry Tool takes a lot longer because it is doing a far more sophisticated calculation, and the product of a Spectrophotometry Tool run is, in fact, research-ready, providing that you gave it good input parameters (as described above).

Names:

• When you upload a list of targets, it's not supposed to require that you identify a column for names from you. However, if you did it one other time in the session, it will probably require that you do so in subsequent requests. (Reload the tool to reset this. You should still be able to see the job monitor if you do this.)
• Uploaded files with duplicate source names causes chaos (and uninformative errors). Given the prior bullet, I recommend restarting the tool, re-uploading your file, and not specifying a name. OR, loading your file into IRSA Viewer, removing ("hiding") the offending name column, saving it as a tbl file without the name column, and then uploading that modified file into the SPHEREx Data Explorer.

Nearby sources:

• If there is a nearby, potentially confused source (recall SPHEREx's pixels are ~6 arcsec; see the Explanatory Supplement on the SPHEREx Mission Page), you need to provide the tool the locations of all of the sources you want to deconvolve. Only the sources you provide, together with the SPHEREx PSF model, are used to measure and return Tractor forced photometry from the images. So, upload a list of targets with all of the sources included.
• However, if there are sources within the 15 px box that the tool uses for background estimation, you do not need to specify those; the tool knows about bright sources all over the sky. (See the Explanatory Supplement on the SPHEREx Mission Page.)

Run time:

• The more frames the tool has to work through, the longer the process will take. If you ask for a source near the ecliptic poles, where there might be at least one image per orbit, it will take a long time for the tool to work through all the images to get a spectrum. If you ask for a source near the ecliptic plane, where there are fewer images, the process will complete relatively quickly.
• If you submit a list of targets all over the sky, it will take a long time for the process to run. If you submit a list of targets close together on the sky, the tool is clever enough to only pull the images it needs once, and obtains fluxes for all the sources at the same time.
• A back-of-the-envelope calculation for point source run time is as follows -- do an image search to figure out how many images cover your target. Then, to obtain the estimated run time in seconds, calculate 1.67*number images + 50; for run time in hours, calculate 0.000463*number images + 0.013. For ~5700 images, it takes a little longer than 2.5 hours. Elliptical fits take longer.

Throttles:

To allow computing resources to be shared fairly among all users, there are built-in throttles.

• You are currently allowed a maximum of two spectrophotometry jobs running simultaneously. Additional jobs will be queued and run when resources are available.
• Lists of targets are currently limited to 20 sources. If you need to get spectrophotometry for more than 20 sources, break up your target list into groups of 20.

Understanding your results: Images:

• The cutouts provided as part of the spectrophotometry results embedded in this tool enable you to easily explore whether that outlier in the spectrum looks ok in its corresponding image, and explicitly check on its neighborhood in the image(s) by overlaying masks, etc. This tool makes it really easy to check the masks -- see the visualization chapter.

Understanding your results: Returned values:

• All the returned values, ranging from ext_flg and nyquist_comp to fit_ql and lambda_width and more, are defined either above or in the Explanatory Supplement on the SPHEREx Mission Page.
• See the bitflag definitions above.
• The background estimated under the source is returned with the photometry as "flux_bkg". The background can be problematic in some (rare) cases. Check for sharp discontinuities between adjacent wavelengths in the background estimates if you suspect an issue with your photometry result.

Evaluating Photometry Tool Outputs:

Here are several steps to evaluate the output of the spectrophotometry tool, especially if the results were not what you expected.

1. Check the location. The spectrophotometry tool does not fit the position; it uses the exact position as entered by the user or returned from the name resolver. Check that this position is correct by looking at the images that are returned as part of the tool. This is especially important for high proper motion sources. Is the position really centered on the source?
2. Check nearby sources of emission. Unless multiple sources are specified, the tool will fit a single source. If there are other sources of emission, either a point source or diffuse emission, this will impact the PSF fitting and the resulting flux. Check this by looking at the images that are returned as part of the tool.
3. Check the flags on pixels used in flux measurement. The tool output includes a per measurement combination of the flags from the pixels used. See photometric flags, above.
4. Check the fit metric. The tool output includes a per measurement fit metric. Values closer to 1 are better.
5. Check the background level. The background level is determined per image and can vary from wavelength to wavelength, particularly in crowded fields. If the background is a significant fraction of the measured flux and varies across wavelengths, this can create the appearance of spectral features that are not real.

Downloading the results:

• There are several ways to download the spectrum you have extracted, and you can save it in any of a number of formats.
• If you have submitted a list of targets, make sure that you are saving the actual spectrophotometry and not just the list of targets -- in general, use a 'download data' button, not a diskette.
• If you have saved your spectrum in a format that you can't immediately read, you can upload the file back in here, or IRSA Viewer can read in the file and then you can turn around and save it in another format that you can more easily read (tbl, FITS, etc.).
• The default filenames of the downloaded files are of the form [ra]_[dec].xml but they are actually *.vot (VO tables), which is a subcategory of XML. (All VOT files are also XML, but not all XML are VOT.)
• The Job Monitor is supposed to hold on to your jobs for 14 days, but in testing we have found some transient gremlins suggesting that (a) you should download the data as soon as you can; (b) logging in before submitting your jobs (and being sure you are logged in before attempting to download your data) seems to keep the gremlins at bay for longer.
• IRSA tends to have scheduled blocktimes every other Tuesday morning Pacific time. Those blocktimes may take down our servers and terminate running jobs and/or lose old jobs. We are working on ways to ameliorate this, but, again, it's probably wise to download your data as soon as possible.