| The water and energy budgets for the atmosphere and land surface are tightly coupled through latent heat exchanges associated with precipitation and evapotranspiration. Diagnostic studies of water and energy transport in the atmosphere rely on radiosonde data, but the global radiosonde network has poor space-time sampling properties, especially for resolving diurnal and higher frequency periodic processes. These shortcomings can lead to large systematic biases in estimated budgets, which confound comparisons with simulations from global climate models, thus hindering model improvement.; In this dissertation, a holistic view of the coupled land-atmosphere system is taken and three well measured constraints for adjusting and validating atmospheric water and energy budgets are studied. They are: (1) mass balance, through measured surface pressure change; (2) energy balance, through satellite estimated net radiation at the top-of-atmosphere; and (3) water balance, through stream flow. Estimated budgets for the land-atmosphere column defined by the Mississippi River Basin are systematically refined using objective optimization methods. The developed methods address spatial sampling bias through mass balance constraints applied to the relative location of radiosondes along the basin boundary, and diurnal sampling bias through energy balance constraints applied to the mean vertical wind divergence field. Stream flow measurements, which are independent of the first two constraints, show that the refined estimates provide greatly improved closure to the water balance, lending confidence to the adjusted fields.; Using the refined atmospheric fields, the roles of individual transport processes in determining the seasonality of the moisture budget (such as the low-level jets from the Gulf of Mexico and the migrating poleward eddy flux) are better resolved. The results are compared with those simulated from twenty atmospheric general circulation models. The variability of hydrologic cycle simulation among models is large but the model median is in general agreement with observations. The models also tend to overestimate precipitation, evaporation, and the moisture recycling rate at the global scale. Other modeled basins are strongly constrained by net radiation, but the partitioning into dry static energy and latent heat convergence varies with model physics and parameterizations. |
