Conjecture, probabilism, and provisional knowledge in renaissance meteorology