Troubleshooting AutoProf

Here we list some common troubles that users can run into when getting started or when beginning to access some of AutoProf’s more advanced features. By it’s nature this is an evolving page, so if you run into something that you think should be addressed on this page please reach out to Connor Stone!

The error bars on the SB profile diagnostic plot are too large

The diagnostic SB plot is intended for evaluating the quality of a fit. With many SB profiles, the statistical errors are too small to see in a plot and so they are scaled up arbitrarily to make them visible. The error-bar scaling factor can be seen in the legend as (err.xx) where xx is the scaling factor. When scaling is applied it is done such that the largest error bar on the plot is now 1 mag arcsec^-2.

I want to hold the PA and ellipticity fixed at a global value instead of fitting each isophote

This can be done with the “isophotefit fixed” pipeline step replacing the standard “isophotefit” step. For information on adjusting the AutoProf pipeline see AutoProf Pipeline Control.

The isophotes go far beyond the extent of the galaxy in the image

This can happen for a few reasons. First, and most straightforward, if the ap_extractfull parameter is set to true then AutoProf will typically evaluate out to the edge of the image regardless of if it has any signal. Second, AutoProf will by default extract isophotes out to three times the largest fitted radius. At the low S/N outskirts of a galaxy it is challenging to robustly fit isophotes, however with fixed PA and ellipticity it is possible to detect signal out quite a bit further. If you wish to control this behaviour you can use the ap_fit_limit, ap_sampleendR, and ap_truncate_evaluation parameters to control various aspects of the extrapolation (see: AutoProf Parameters). Third, it may be a result of the sources in the image. If the image contains galactic ciri or diffracted light from a very bright star it is possible that AutoProf will extend very far (beyond the extent of the target galaxy) before reaching the background noise level. In this case one can often see in the SB profile where the dominant light source switched from the target galaxy to the diffuse source as the SB profile will flatten at a non-zero signal. Unless you have some sophisticated way to remove the spurious signal, users may wish to heavily truncate the SB profile or reject the galaxy form their sample as accurate extended photometry will be nearly impossible to acquire.

The fitted isophotes are too smooth, they are missing features

AutoProf is intended for use on large surveys with thousands to millions of objects. By default it applies a certain amount of profile smoothing to achieve reliable automated results. In some cases this smoothing can overdo it and will cause AutoProf to skip over a real feature (it will still fit the global structure though). This smoothing is controlled by ap_regularize_scale which by default is 1. Decreasing the value will reduce the smoothing strength, at 0 there will be no smoothing applied. If the isophotes continue to come out too smooth then the culprit is likely the ap_isofit_robustclip parameter, which clips high flux values while fitting isophotes. It defaults to 0.15 meaning that the 15% highest flux values are clipped down to lower values. This parameter can also be set to 0 to turn off the effect. It is not advised to turn off these features when batch processing large numbers of galaxies since it will make the fit susceptible to spurious image features.