Reply to comment by Roberto Rondanelli and Richard S. Lindzen on Variations of tropical upper tropospheric clouds with sea surface temperature and implications for radiative effects