integrated flux changes with input.region (z_min,z_max)

[abhisek176]

Hi,

I am facing the following issue using Sofia: The integrated flux from the source is getting reduced if I put the spectral region in the param file (i.e., z_min and z_max). I am fixing the remaining parameters the same.

[abhisek176]

Just adding a note: f_sum changes when giving spatial ranges as well. For e.g. going from 400 pixels to 300 pixels increase the f_sum by ~2%.

[SoFiA-Admin]

Thank you for reporting the flux issue. My understanding is that you are changing the region size (parameter input.region) while leaving all of the other settings the same. Small variations in the integrated flux (and other source parameters) are to be expected in this case, as the source finder will apply a relative flux threshold based on the noise level. Changing the region size will effectively change the noise sample and inevitably lead to a slightly different noise measurement and hence flux threshold to be applied. This, in turn, will result in a slightly different source mask and hence small stochastic variations in all of the source parameters measured across that mask.

However, all of these flux measurements should be consistent with each other within the statistical uncertainties implied by the noise (plus any systematic errors). The fact that you are reporting changes at the percent level would suggest that this is indeed the case, and all flux measurements should therefore be correct within the limits set by their statistical uncertainties.

Another point to keep in mind is that changing the region size will also change the boundary conditions experience by the spatial and spectral smoothing filters. This too could have an impact on the final source mask if the sources are too close to the edge of the region.

[paoloserra]

It would be good if you could check whether the shape of the detection mask changes when you change the spatial/spectra ranges. That would confirm the explanation given by @SoFiA-Admin .