Increased skew_sortkey upon s3 COPY

Question

I have a table that was created via incremental loads that has a lower skew_sortkey value than when I fully load the table via a s3 copy. Both tables have the same sort key and identical DDLs, but the skew_sortkey value is 7-8 times higher doing the s3 copy. Is there a reason for this? How can I improve (reduce) the skew?