There are several points I need to cover because your questions seem to cover the same subject. You just ask the same question differently but the answer is the same.

Replicas

As I explained in your previous question (#6523), "read replicas" and "write replicas" are one and the same. No one uses those phrases in Cassandra -- they are just replicas.

I also mentioned in that post that you're asking about Cassandra concepts one tiny bit at a time. If you at the very least watch the videos in the DS201 Cassandra Foundations course, you will save yourself a lot of time. Help me help you -- do the course. :)

Auto-scaling

Automatically adding or decommissioning nodes is never a good idea with Cassandra because it is an expensive operation. Bootstrap and streaming operations take a long time depending on how much data each node owns. Once the bootstrap or decommission process has started, it can't be interrupted without consequences.

For the same reason, the cass-operator does not support on-demand scaling. John Sanda has covered this briefly in his response to your other question (#6528). This is effectively the same question so the answer is the same.

Scaling up a DC

It is possible to manually scale up the number of nodes in a DC using the cass-operator.

The size parameter of the CassandraDatacenter determines how many nodes are deployed in a DC. Here's an example configuration of a DC called dc1 with size: 3 nodes:

kind: CassandraDatacenter
metadata:
  name: dc1
spec:
  clusterName: cluster1
  serverType: cassandra
  serverVersion: 3.11.6
  size: 3
  ...

To scale up the number of nodes, you will need to edit your YAML configuration and re-apply it with the kubectl command. The pod(s) will be added to the DC provided there are enough Kubernetes worker nodes in your Kubernetes cluster. For details, see Scaling up a datacenter.

Note that scaling down is not supported as documented in the cass-operator release notes. Cheers!