Rework list of compsets to align with the compsets needed for the ctsm6.0.0 release

We need to rework the list of compsets to align with the compsets that are needed for the ctsm6.0.0 release:

https://docs.google.com/spreadsheets/d/1nJl4lbbwBq60quyaERJ6OvVJxF4kDu4xX7zZenZgyB8/edit?gid=1789776230#gid=1789776230

In large degree we should leave the Clm45 and Clm50 compsets in place, but we should look for some to remove.

Currently there are 15 Clm45 compsets:

``` xml
I1PtClm45SpRs
I2000Clm45BgcCropQianRs
I1850Clm45BgcCrop
I1850Clm45BgcCru
IHistClm45BgcCrop
IHistClm45BgcCropQianRs
I2000Clm45Sp
I2000Clm45BgcCrop
I2000Clm45Fates
I2000Clm45FatesRs
I1850Clm45Bgc
IHistClm45Bgc
IHistClm45BgcCru
IHistClm45Sp
I2000Clm45VicCru
```

We should remove the Vic one for sure. And remove the Cru ones. We should probably support only the Rs or plain ones. And I think there is no reason to support both the Qian and GSWP3 versions.

The list of Clm50 compsets is longer having 54 of them. Again I think it makes sense to remove Cru. And we should remove the Rtm compsets and let the Clm45 compsets be the ones that work with Rtm. Other's should be examined for removal as well.

For Clm60 there's 37 compsets, and possibly some of the ones in the Clm50 list should be just for the latest Clm60 compset list. Do we need both Cru and Qian forcing for Clm60 or just use the older physics for these? Qian is used to save download and runtime for single point sites, so that might still be useful for Clm60.

Linking #3044 which is similar to this one.