DelSole, TimothyYan, Xiaoqin2016-09-282016-09-282016https://hdl.handle.net/1920/10449One of the most significant uncertainties in climate change projections is the sensitivity of the climate system to increasing greenhouse gas concentrations. Attempts to estimate this sensitivity based on observations over the past century have failed to reduced this un- certainty, primarily because of uncertainties in the cooling effect of aerosols, which have cancelled some of the warming induced by greenhouse gases. This study attempts to improve estimates of aerosol cooling by exploiting new statistical techniques and by identifying variable combinations that are more effective at estimating the response to climate forcings than single variables. The exploration of variable combinations is facilitated using a new measure called potential detectability, which quantifies the extent to which the response to climate forcing can be detected in a model. It is shown that joint temperature-precipitation information over a global domain provides the most accurate estimate of aerosol forced responses in climate models, compared to using temperature, precipitation, or sea level pressure individually or in combination. Unfortunately, observational errors in precipitation are too large to permit estimation of aerosol-induced climate changes. Repeating this estimation using only land-data, where reliable rain gauge data are available, succeeds in estimating aerosol cooling, but is only modestly improved by including precipitation data.122 pagesenCopyright 2016 Xiaoqin YanClimate changeStatisticsAtmospheric sciencesAnthropogenic aerosolsClimate changeClimate detection and attributionOptimal fingerprintingDissertation