Unable to scale down atlas cluster's storage after deleting documents and running compact

I was doing a one-time large data test of our API webapp in our testing environment, and as a result our devtesting Atlas cluster autoscaled its storage from 8GB to 32GB, which is fine.

But now that I’ve deleted the added data and run compact on the collection, I want to scale the storage tier back down to 8GB.

The data usage is sitting at 6.15 GB, but when trying to scale down I get the error “The selected disk size (8.0 GB) is smaller than the amount of data currently used (16.04489517211914 GB)”

What else is necessary to truly reclaim the space to allow a storage scale-down?