WISE Tom Jarrett

Galaxy Deblending: KPAIRSMALL18

WISE view of galaxy pair: kpairsmall18.

KPAIRSMALL18: Deblend Iteration Sequence

The following sequence shows 4 total interations. This pair requires human intervention because of the (extreme) tidal-distorted disk of component "A". There is no clean way to model this galaxy, so a few tricks are needed.

Galaxy A Galaxy B

Iteration 1: The red (star formaing) galaxy is component "A". It is highly irregular due to tidal distortion from the neighbor interation. The outer disk has been turn off and skewed at right angles to the main disk. The initial guess at the shape is somewhat confused because of this. The trick is now to intervene and change the shape. The good news is that neighbor "B" is recognized as a 2MASS XSC galaxy, and elliptically-masked according to the XSC shape (as stored in the WISE "g" parameters). Note the "mask" pixels are replaced by the local background.
Iteration 1: The blue galaxy is component "B". The neighbor, "A", has been model-subtracted, but it's a very poor result. Human intervention at this point:

After iteration 1: change the initial shape of "A", forcing a solution that is more aligned with the total disk. Since this is not a symmtric distribution, we are forced to use a "self model" instead of an ellipsoid. Self-model is the last resort and only used when the galay is highly irregular.
After iteration 1: Constrain the shape and size of the solution for "B". This is necessary to avoid the train-wreck mess of "A".

Iteration 2: Now instead of a mask, the component "B" has been removed using an ellipsoidal model of its light distribution.
Iteration 2: Neighbor "A", has been removed using the "self" model of its light distribution.

Iteration 3:
Iteration 3:

Iteration 4:
Iteration 4:

Photometry after each iteration

Galaxy A	Galaxy B
Iteration 1: The red (star formaing) galaxy is component "A". It is highly irregular due to tidal distortion from the neighbor interation. The outer disk has been turn off and skewed at right angles to the main disk. The initial guess at the shape is somewhat confused because of this. The trick is now to intervene and change the shape. The good news is that neighbor "B" is recognized as a 2MASS XSC galaxy, and elliptically-masked according to the XSC shape (as stored in the WISE "g" parameters). Note the "mask" pixels are replaced by the local background.	Iteration 1: The blue galaxy is component "B". The neighbor, "A", has been model-subtracted, but it's a very poor result. Human intervention at this point:
After iteration 1: change the initial shape of "A", forcing a solution that is more aligned with the total disk. Since this is not a symmtric distribution, we are forced to use a "self model" instead of an ellipsoid. Self-model is the last resort and only used when the galay is highly irregular.	After iteration 1: Constrain the shape and size of the solution for "B". This is necessary to avoid the train-wreck mess of "A".


Iteration 2: Now instead of a mask, the component "B" has been removed using an ellipsoidal model of its light distribution.	Iteration 2: Neighbor "A", has been removed using the "self" model of its light distribution.
Iteration 3:	Iteration 3:
Iteration 4:	Iteration 4:

Photometry after each iteration