An assessment of the potential impact of a downward shift of tropospheric water vapor on climate sensitivity

This work uses an energy balance climate model (EBCM) with explicit infrared radiative transfer, parametrized tropospheric temperature and humidity profiles, and separate stratosphere, troposphere, and surface energy balances, to investigate claims that a downward redistribution of tropospheric water vapor in response to surface warming could serve as a strong negative feedback on climatic change. A series of sensitivity tests is carried out using: (1) a variety of relationships between total precipitable water in the troposphere and temperature; (2) feedbacks between surface temperature and the vertical distribution of tropospheric water vapor at low latitudes; and (3) feedback between surface temperature or meridional temperature gradient and lapse rate. Fixed relative humidity (RH) enhances the global mean surface temperature response to a CO₂ doubling by only 50% compared to fixed absolute humidity, giving a response of 1.8 K. When water vapor is assumed to be redistributed downward between 30°S–30°N such that a 1 K surface warming reduces total precipitable water above 600 hPa by 10%, the global mean surface air temperature response is reduced to 1.2 K. Assuming a stronger downward redistribution in relation to surface temperature change has a rapidly diminishing marginal effect on global mean and tropical surface temperature response, while slightly increasing the warming at high latitudes due to the parametrized dependence of middle-to-high latitude lapse rate on the meridional temperature gradient. A modest downward water vapor redistribution, such that absolute humidity in the upper troposphere at subtropical latitudes is constant as total precipitable water increases, can reduce the tropical temperature sensitivity to less than 1 K, while increasing the equator-to-pole amplification of the surface air temperature response from a factor of about three to a factor of four. However, it is concluded that whatever changes in future GCM response might occur as a result of new parametrizations of subgrid-scale processes, they are exceedingly unlikely to produce a climate sensitivity to a CO₂ doubling of less than 1 K even if there is a strong downward shift in the water vapor distribution as climate warms. Received: 23 February 1998 / Accepted: 1 November 1999