Have you looked at the ->get method of WWW::Mechanize? In fact, you already use it yourself when loading the initial page. Maybe consider using it also when you want to download another page.

You can also look at the ->follow_link method in the same documentation. But note that ->follow_link will only work for following one link on a page.