The Impact of Changes in Water Vapor, Clouds, and Snow Cover on Elevation Dependent Warming

Abstract:

There are many high-elevation regions that are experiencing greater rates of warming than their lower elevation surroundings. Using a combination of ground-based observations, satellite retrievals, and CMIP5 model simulations, we examine effects of specific humidity (q), snow cover, and cloud cover and optical thickness on surface downward longwave (DLR) and shortwave (SW) radiation, and quantify them using a neural network analysis. The sensitivity of DLR to changes in q is non-linear and larger when the atmosphere is drier, which is more likely at higher elevations during winter and at night. The magnitude of q is the primary determinant of the DLR-q sensitivity, although observations also reveal that finer-scale topography has an impact on the diurnal variation of this sensitivity. Although DLR increases with increasing cloud cover, clouds have limited effect on the sensitivity of DLR to changes in q. The sensitivity of DLR to changes in cloud optical thickness is generally small but does increase as clouds become semi-transparent. The sensitivity of surface SW fluxes to changes in cloud cover depends significantly on cloud optical thickness. This has implications for the prediction of trends in incoming SW fluxes as cloud cover changes are usually the only trend that is measured. The CMIP5 model simulations for the 21^st century show that the snow-albedo effect contributes to enhanced warming in the Tibetan Plateau and surrounding high-elevation regions. The albedo decreases more at higher elevations than at lower elevations owing to the retreat of the zero-degree isotherm and the associated snow line. The treatment of snow cover, snow melt and the associated snow-albedo feedback may have an impact on the climate sensitivity among models. Elevation dependent warming is found in all but one of the CMIP5 models, and the degree of enhancement is weakly correlated with their global sensitivities.