On surface versus atmospheric forcing in regional climate simulations

Abstract

<p>This dissertation, a collection of three papers, performed a number of numerical experiments to contrast surface and atmospheric forcing in regional climate simulations. Results of the first paper showed that a transient increase in soil moisture enhanced domain total rainfall. The increase in soil moisture enhanced local rainfall when the lower atmosphere was thermally unstable and relatively dry, but it decreased the rainfall when the atmosphere was humid and lacked sufficient thermal forcing to initiate deep convection;It was found in the second paper that as integrations continued without reinitialization, locations of weather systems drifted downstream because of the positive model bias on wind speeds, implying the necessity of periodically reinitializing the model. The simulated domain-averaged variables were not sensitive to model reinitialization, suggesting the feasibility of dividing long regional climate simulations into a set of shorter segments that could be run in parallel;The results of the third paper suggested that the current landuse increased domain-averaged rainfall by less than 1% over the pre-settlement landuse, and the hypothetical future uniform cropland increased total rainfall by 3% under the normal condition. The current and future landuse tended to reduce rainfall over the east-central U.S. and to increase rainfall over the western U.S., especially in the flood year;The following conclusions may be drawn as to surface versus atmospheric forcing: (i) The mean properties over the continental scale are determined mainly by large-scale atmospheric forcing, and surface forcing has little impacts on them. (ii) With given large-scale atmospheric forcing, surface forcing has noticeable roles to redistribute atmospheric variables by either shifting specific patterns around or modulating them. (iii) The displacement of weather patterns can result in large differences in climate variables on local scales, especially for rainfall.</p>