WISE/NEOWISE Coadder
Overview and Instructions
Overview (version 2.3 - Nov 14, 2024):
The IRSA WISE/NEOWISE Coadder enables users to create custom coadds (mosaics) from single-exposure images. This is different from the "Atlas" mosaics, which are pre-built. With this tool, one can specify the desired center position (or object), the output size (up to 3.0 x 3.0 deg), rotation angle, and output pixel scale. The input images may be filtered by observation date/time, or other properties. Some uses for this tool might be studies of time-variable sources or moving objects.
Coadds may be generated for a single position or multiple positions listed in an uploaded table. For single positions, user selection of images to avoid any problem in a given image is available. Moving object coadd creation is also available.
An optional resolution enhancement using the Maximum Correlation Method ("HiRes") is available (see below), but only when generating area-weighted coadds, which is the default.
For single positions, the service can also be accessed by an HTTP program-line interface. There are descriptions of all of the input values on this page.
The WISE/NEOWISE Coadder uses the "Image Coaddition with Optional Resolution Enhancement (ICORE)" software that formed the basis for the construction of the WISE Atlas Images. The web interface allows adjustment of the input variables most likely to be needed by the majority of users.
Important Note on Interpolation: The WISE/NEOWISE Coadder uses simple area weighting for its interpolation by default. Also available is Point-Source Response function (PRF) weighting, the method which was used to generate WISE Atlas mosaics (see section 6 of the ICORE Software Description). Sources in mosaics generated using area weighting will be slightly sharper than those in Atlas images. The reason for using PRF-weighted interpolation is to create "matched filter" mosaics which are optimized for point-source detection. If PRF weighting is selected, note that no further "matched filtering" should be done, say by a point-source detection program.
Important Note on Running Time:With the final release of NEOWISE data, coadds made with the defaults (i.e. all available data) can take a long time. If using the defaults, it is best to submit through the web page and supply an e-mail address when the monitor page appears. The time estimates displayed are very rough and vary with server usage; for the defaults at low ecliptic latitudes, count on at least 20 minutes, though one can reduce this by limiting the time range. The optional resolution enhancement can take up to twice as long. The time increases with ecliptic latitude because WISE/NEOWISE coverage increases toward the ecliptic poles. Close to the poles, that could mean running times of weeks, depending on coadd size and coverage. So a cutoff for estimated times of > 5 days is in place. Coadds taking longer may be made using the downloadable version of ICORE, see the ICORE Software Description.
Important Note on WISE Band 3: By default, the Coadder will only include WISE band 3 (12 micron) data from the "3-band" phase of the mission when there are no valid "4-band" W3 data in the requested range. It will not combine data from the two periods and "4-band" data will take priority. This is to avoid problems in background correction and outlier rejection in some fields due to changes in noise level during the warm-up ( see the 3-band Summary ). You can include all W3 data when doing user selection of images.
If you use the WISE/NEOWISE Coadder in your research, please cite the following paper on ICORE:
Masci, F.J., and Fowler, J.W., 2009, in ASP Conf. Ser. 411, ADASS XVIII, ed. D.A. Bohlender, D. Durand, & P. Dowler, (San Francisco: ASP), 67.
Input:
Single Position:
Select the "Single Position" radio button on the input form, and enter a coordinate or astronomical object name. Your entry can be any valid coordinate, or an object name to be automatically resolved by NED or SIMBAD.
If you want values for parameters different from the defaults, enter them in the boxes. A description of the parameters and the other frame qualities which may be constrained is on the HTTP program-line interface page.
The default date limits are 07Jan2010 01:45:14 to 01Aug2024 06:59:48 (see table below). You can narrow your search to specific dates. The latest date for band W3 is 29Sep2010 18:39:37 and for band W4, 06Aug2010 10:22:03.
For reference, here are the date limits of the mission phases:
Mission Phase | Earliest Date | Latest Date |
4-band cryogenic, W1-W4 | 07Jan2010 01:45:14 | 06Aug2010 10:22:03 |
3-band cryogenic, W1-W3(1) | 06Aug2010 10:35:04 | 29Sep2010 18:39:37 |
NEOWISE Post-Cryo (2-band), W1, W2 | 29Sep2010 18:41:16 | 01Feb2011 11:03:02 |
NEOWISE-R (Reactivation), W1, W2 | 13Dec2013 18:39:35 | 01Aug2024 06:59:48 |
(1)By default, the Coadder will mosaic W3 images during this period only when there are no valid 4-band W3 data in the request, i.e. it will not combine data from the two mission phases and 4-band takes priority.
Note that the default value for the frame quality minimum is 5. To choose only the best frames, use 10. Choosing 0 would give all frames, including some of low quality, so is currently excluded.
For single objects, there is the option for user-selection of the frames to use to avoid any bad frames. For details, see user selection of images.
Multiple Positions:
Select the "Upload Table" radio button, and then enter the name of a table on your local file system which contains a list of coordinates or astronomical object names.
There is a web page with table upload help.Select the "Moving Object/Stacking" radio button. There are two options for input: a Solar System object identifier, or a table of WISE/NEOWISE frames containing the object. For details, see the Moving Object/Stacking help.
Select the "Moving Object/Stacking" radio button. The user may select images to stack, either of the same object or of a class of objects with suitable properties for combining. The code will shift images to a common position. For details, see the Moving Object/Stacking help.
Resolution Enhancement (HiRes)
Selecting this option when using the default area-weighted interpolation allows an iterative image-reconstruction solution using the Maximum Correlation Method. The algorithm tries to achieve the maximum correlation between adjacent pixels, i.e. the smoothest image, consistent with the data. It is the same HiRes technique developed for use with IRAS data (Aumann et al. 1990). Complete details are in Section 8 of the ICORE Software Description.
Results:
Hit the Submit button once you're done specifying the request. A status page is returned with an approximate estimate (in minutes). If you prefer not to wait, you can fill in the e-mail box to be notified on completion.
The final results page contains a summary of the mosaics that were generated. Up to 50 objects are displayed in a Results Table:
- A Quicklook JPEG thumbnail of the mosaic (don't click on this, go to Save/View).
- Index from original table.
- RA and Dec of the center of the coadd.
- WISE band number.
- Type of Data (data, uncertainty, coverage, standard deviation if area-weighting is done, cfv uncertainty for HiRes, see ICORE Software Description).
- Root filename of the FITS image mosaic.
- Links to download the FITS file to disk, or send it to IRSA's interactive FITS viewer.
For area-weighting the uncertainty image should be a valid estimate of uncertainties, but the standard deviation mosaic is useful for seeing the spread in data at that pixel position.
Note: If no cutouts are presented on the results page, it may be because there was no coverage for those input parameters.
There is also a table containing lists of frames used in the coadd.
Instructions are given regarding bulk download of all mosaics, using "wget". There is also the option to download everything within one "tar" file.
Notes on Units: The units of the intensity and uncertainty mosaics are WISE data numbers (DN) or values, as noted in the header keyword ''BUNIT''. (Note: the coverage has the same BUNIT, but is actually just a number.) Aperture photometry on the intensity mosaics is possible using the value of the header keyword ''MAGZP''. Conversion to WISE magnitude of a background-subtracted sum of pixel values, S, in a sufficiently large aperture, or with appropriate aperture correction, is:
magnitude = MAGZP − 2.5log10(S)
See section 13 of the ICORE Software Description for full details.