To answer [1] Why?: it's complicated.
There are more than 35 red blood cell groups (see https://www.science.org.au/curious/people-medicine/blood-typ... for a nice writeup). For each of those blood groups, there is more than one possible configuration of some protein or carbohydrate (something like more than one possible genetic sequence leading to more than one kind of molecule on the surface of the RBCs).
And, even with ABO, there can be infrequent variations that make things more complicated (see https://professionaleducation.blood.ca/en/transfusion/best-p... for more).
For the other blood groups, I think every case the groups were identified because a patient somewhere made an antibody, causing either a transfusion reaction (if not tested ahead of time) or, more likely, a positive (incompatible) reaction on in compatibility testing.
Each of those blood groups will have two or more versions. For example, there is the Kell blood group, which has two primary versions: you can express big K, little k, or both as Kell is a codominant system. Most people are homozygous little k (kk). When you are homozygous little k, you can make antibodies to big K (which is expressed on people who are Kk or KK). So, if you are homozygous little k, you are at risk of making antibodies on exposure to big K. Of course, not everyone is either big K or little k, that would be too simple. There are, I think, more than 20 variants of K out there...
More or less, each of those blood groups can cause antibody formation if you lack an antigen on the surface of those antibodies. In practice, the likelihood of an antigen causing antibody formation varies by antigen (the term there is "antigenic"). Also, the immunologic status of the recipient matters - obviously, someone who is immunosuppressed from an organ transplant may be less likely to form antibodies (that's the point of of the immunosuppression). Conversely, some people are more likely than most to form antibodies.
Also, it appears that the immune system is more likely to form antibodies in other blood groups if a person already has antibodies. In practice, there is a recommendation from NHLBI to match for C, E, and K in patients who are going to be transfused often to try to avoid forming new antibodies.
People 'naturally' form antibodies to A and to B antigens that they lack (I am group O and have both anti-A and anti-B antibodies). This is because they are exposed to those antigens in everyday life (the A and B antigen is found on some gut bacteria). And Rh-D is very immunogenic and clinically significant.
When we test blood prior to transfusion, we attempt to identify what antibodies the patient may have. If they have no unexpected antibodies, just honoring blood and Rh-D type is usually enough. If they have unexpected antibodies, we go through a complicated process to identify blood that lacks those antibodies to try to avoid a transfusion reaction.
(Now to modify that testing point, we know some antibodies are not clinically significant - that is, they do not cause significant transfusion reactions. So, we can, if we have to, give blood that is incompatible for clinically insignificant antibodies. We still try to provide blood negative for those antigens because we wouldn't find out this patient is the first to have a clinically significant version of the antibody. Also, the blood tests as incompatible with clinically insignificant antibodies, and we would prefer to give blood that tests as compatible).
Sorry this is rambling a bit - I have to take my daughter to school. Happy to expand/clarify/etc. when I can get back to it.
Source: I am a practicing physician board certified in "Blood Banking/Transfusion Medicine". My statements in this are still likely inaccurate, but hopefully in a useful way. I am not your lawyer, and I am not your physician.