I don't understand what exactly you want to do, because The title differs from the image in context.
Anyway , (as your title )i see the following tools maybe help you:
If there is indeed a trend, did you try to run a first order (linear) trend surface on the data and subtract the resultant from the input dem?
http://desktop.arcgis.com/en/arcmap/latest/tools/spatial-analyst-toolbox/trend.htm
http://desktop.arcgis.com/en/arcmap/latest/tools/spatial-analyst-toolbox/how-trend-works.htm
Not sure if the tool exists in the same locale (or if at all) in PRO.
Worth a first run... otherwise you could generate the surface I suppose from first principles (ie using numpy meshgrid and generate the z values from a 2D equation, then convert back to raster for subtraction
I have messed a bit with artificial terrain generation
https://community.esri.com/blogs/dan_patterson/2017/12/28/terrain-creation-diamond-square-algorithm
But I haven't played with input parameters to generate exact slopes and aspect terrain... but this might give you some idea. See the link.
# ---- A 'slope' with a NW aspect ----
array...
[[ 0.250 0.351 0.420 0.488 0.542 0.604 0.656 0.704 0.750]
[ 0.342 0.383 0.443 0.502 0.559 0.613 0.663 0.706 0.732]
[ 0.392 0.427 0.469 0.525 0.578 0.630 0.677 0.714 0.740]
[ 0.438 0.463 0.503 0.551 0.601 0.650 0.694 0.732 0.754]
[ 0.458 0.493 0.531 0.577 0.625 0.673 0.719 0.757 0.792]
[ 0.496 0.518 0.556 0.600 0.649 0.699 0.747 0.787 0.812]
[ 0.510 0.536 0.573 0.620 0.672 0.725 0.781 0.823 0.858]
[ 0.518 0.544 0.587 0.637 0.691 0.748 0.807 0.867 0.908]
[ 0.500 0.546 0.594 0.646 0.708 0.762 0.830 0.899 1.000]]
Hope you get the idea, but you would have to mess with the corner values for your raster extent and put in values to reflect the slope's orientation and not only its slope
There are other algorithms out there
Thank you for your help!