The OP is right about the problem... That said, I think the prescription for fixing it is wrong.
I think that surface detection should be first quantified by task force. Each task force, ships grouped together, should have at least the value of the highest surface detection ship in their group. It SHOULD NOT be the average. However, it also shouldn't be purely additive. I will explain why below.
Having multiple task forces in a sea zone SHOULD certainly be additive. Each task force can fully add their detection in that sea zone because they cover more area.
As per the example above, adding destroyers to a escort CV task force doesn't diminish the CVs surface detection. But it doesn't help much either. Why? The destroyers which are sticking close to the CV to protect it are not going to spot enemy ships much better than the CV they are near. More likely the height of the CV makes it better for spotting directly thorough LOS, and the aircraft, if they can fly missions, are going to spot anything way before the surface ships see anything on the horizon. If anything destroyers in a CV task force might add a little surface detection.
If you want destroyers to use their speed to move about the sea zone trying to find ships they should be in their own task force moving separately from the CV task force in the scenario above. Screening ships shouldn't be allowed to add their full surface detection and be able to screen capital ships in a task force at the same time. In a separate task force with their speed they should be capable of spotting the enemy and attempt shadowing any slower ships until another task force closes and engages in combat. Perhaps if you exceed the needed ratio of Screens/Capital ships the screening ships should add more to detection values?
Now I will admit that having multiple ships in a task force does increase the task forces spotting abilities a little bit... However, it isn't increasing it by its full value at all. Two ships of any type that are staying close together do not have double the surface detection as one. However in separate task forces spread out they do. Also, having 10 ships with radar in the same spot does not cover much more area than 1 ship with radar. However spread those 10 ships out all over a sea zone and you get some real coverage. Perhaps adding more ships to a task force can have a small additive effect?
The most confusing scenario is this. Lets say I have several types of ships together. A CV with its aircraft that have a very high detection during the day. Another ship has radar which is a different type of detection which also works at night unlike aircraft. And finally I have fast ships which have their LOS detection which greatly benefits from the speed of the ships.
Therefore if the system was to be perfected... IMHO, I think there should be three types of surface detection factored and then added into each task forces detection abilities.
One for air spotting provided by catapult aircraft and CVs (these should have some additive factor, more aircraft results in better spotting). One for radar from the ship in the TF with the best radar. One value for task force spotting by line of sight (LOS) incorporating both the speed and size of the task force. The slowest ship in the TF would define the base speed since the ships should be staying close together. Finally (Perhaps) per my previous example if you have extra screening ships above the needed screen/capital ratio those extras could boost the "speed" and size LOS detection value for the task force. This makes sense because the excess screens could spread out for search duty while still staying somewhat close to the main TF.
Then those values (air, radar, direct LOS) detection should be added together to give a final detection value for each task force.
Sub detection should be done in the same way but sonar would be added on top of the other 3 factors.
Finally, the game would need to be balanced again. However, this is the realistic solution I think everyone would love to see.
Unfortunately, I have no idea exactly how it works in game now? I can only guess from the information that has been provided by others on this forum. If it does average (which is how it appears to work) it is wrong. If it was additive within a task force it is also wrong. My prescription is how it did work in reality.