The terra-cli README includes examples for both us-east1 and us-central1.
It would be good for us to make an effort to standardize on us-central1.
Many projects that already exist in the life sciences in the US have either used US multi-regional buckets or us-central1 (in order to reduce storage costs.
Terra Community Workbench (app.terra.bio) has always defaulted to running VMs in us-central1 (workflows or notebook VMs).
The problem with arbitrarily using regions around the US is that if project A stores data in us-central1 and project B stores data in us-east1, then any cross-analysis of these two datasets is going to end up unnecessarily incurring GCP network egress charges as data needs to leave one region or the other.