Responses of the Climate System to Opposing Solar and CO2 Forcings

Abstract

This dissertation analyzes various aspects of the response of Earth’s climate to a simultaneously imposed increase in the atmospheric CO2 concentration and decrease in the solar constant, based on an analysis of experiment G1 of the Geoengineering Model Intercomparison Project. The analysis focuses on changes in meridional temperature gradients, tropical precipitation, the seasonal cycle, cloud fraction, and the atmosphere’s radiative balance. I propose and evaluate a theory for the amount of solar constant reduction required to maintain global mean temperature under increased CO2. The atmosphere transports less heat from the tropics to the poles in G1 than in preindustrial conditions, which implies that the phenomenon of polar warming and tropical cooling in G1 must be directly caused by the imposition of a net negative forcing in the tropics and a net positive forcing in high latitudes. Damping of the polar warming by the reduction in energy transport helps to explain why most of the polar amplification of warming in quadrupled-CO2 experiments is canceled in G1. Changes in cross-equatorial energy transport by the atmosphere help account for the inter-model spread in tropical rain shifts in G1. Radiative effects of changes in clouds are the largest source of inter-model spread in changes in meridional energy transport. The seasonal migration of the tropical precipitation median is damped in G1, due to preferential cooling of the summer hemisphere by the solar reduction. The amplitude of the seasonal temperature cycle is reduced in much of the troposphere due to reductions in insolation and water vapor concentrations. There is an increase in seasonal temperature amplitude in parts of the stratosphere. The required solar constant reduction to achieve energy balance in G1 is between 3.2% and 5.0%, depending on the model, and is uncorrelated with the models’ equilibrium climate sensitivity, while a formula from the experiment specifications based on the models’ effective CO2 forcing and planetary albedo is well correlated with but consistently underpredicts the required solar reduction. I propose instead that the required solar reduction should be equal to the sum of the instantaneous CO2 forcing and the shortwave and longwave radiative adjustments to the combined forcing. To test this hypothesis, I analyze changes in cloud fraction and the atmospheric profile of temperature and humidity, and I quantify the radiative effects of these changes at the top of atmosphere. Low cloud fraction decreases in all models in G1, likely due to reductions in boundary layer inversion strength over the ocean and plant physiological responses to increased CO2 over land. High cloud fraction increases in the global mean in most models. Among the various radiative adjustments, there are strong warming effects due to reductions in low cloud fraction and upper tropospheric and stratospheric cooling. Reductions in water vapor roughly offset the tropospheric temperature effect, while cloud changes have little effect in the longwave. Taken together, the sum of the diagnosed radiative adjustments and the CO2 instantaneous forcing predicts the required solar forcing in G1 to within about 6%. The theory presented here for the required solar constant reduction should make it easier to run the G1 experiment in the future, since a better initial guess can be made which reduces the required amount of tuning. Also, the theory demonstrates that implementing solar geoengineering in the real world would require advance knowledge of the rapid adjustments to the geoengineering forcing, which could only be obtained through large-scale outdoor field tests. Doing such experiments would raise ethical questions, which have been discussed elsewhere. The extent to which the G1 experiment tells us about how clouds and other atmospheric properties respond differently to solar versus CO2 forcings depends on the extent to which rapid adjustments to the two forcings add linearly. Answering this question would require analysis of solar-forcing-only experiments in a multi-model framework.