ETL is not a one time process. It is used to make periodical updates to
dimensional as well as fact data.

Usually dimensions contains lookup data and has less tendency to change but
if it changes and you want to maintain the change history then surrogate key
will help with it. Read about SCD (slowly changing dimension) for more details.

Also surrogate key helps uniquely identify a row in the table. Many times
ETL Extraction process is designed to pull data from multiple sources and
there's possibility that same rows can get pulled from different sources. If you
use key column same as source you are likely to get the Primary Key violation
error. Surrogate key will let you pull all the data and you can clean it during
Transformation/Cleansing process.