Because the RG size is set to 16 (which is the default for SAS,FC) and thats why you have the 16, 16, 8 situation right now.
In order to get to even RGs, you need to control where new disks go when adding them to the aggregate. If you do not influence this, WAFL will add them to the latest RG.
You can change the RG size anytime, this is nondisruptive. It is only in effect when adding disks to the aggregate. Keep in mind, that a RG size of 28 might not be be the best choice for what you want to achieve. The bigger the RaidGroup, the higher the risk if a triple disk failure within the same RG.