Diurnal to annual variations in the atmospheric water cycle

This dissertation examines aspects of diurnal to annual variability in the atmospheric water cycle in observations and global numerical weather prediction models. Investigations begin with an in-depth evaluation of variance at a single time scale, followed by a comprehensive analysis of a particular water cycle component, and finally a complete description of the balances and exchanges of water cycle components across time scales.Comparisons of global and regional reanalyses reveal significant differences in the amplitude and phase of water and energy components' diurnal cycles, with parameterization and land-surface errors propagating throughout the system. Evaluations of the 6-hour to 1-year spectra of global precipitation data from reanalyses and high-resolution precipitation products also indicate significant model biases in capturing the sub-seasonal variability of precipitation as well as disagreement among observation-based products. Component interactions in the atmospheric water cycle reveal distinct characteristics that are unique to particular locations and time scales, including a clear separation between thermodynamic and dynamic controls of variability.Together, these experiments reveal considerable differences in the physical mechanisms that govern the atmospheric water cycle at different temporal and spatial scales. Parameterization sets that have been tuned for performance on a given frequency are often inadequate for simulations that require more complete statistical representations of hydrometeorological processes, and simulated water cycle components are still far from observations. Methodologies introduced in this study help isolate key obstacles that prevent the accurate simulation of global hydroclimate, and identify unique processes that are vital to understanding regional anomalies.