This dissertation describes spatio-temporal properties of rainfall. Rainfall in space was modeled by a precipitation areal reduction factor (ARF) using a NEXRAD image. The storms are represented as ellipses, which are determined by maximizing the volume of rainfall. The study investigated 18,531 storms of different durations that took place in different seasons and regions of Texas. Statistical analysis was carried out to find a relationship between ARFs and predictor variables (storm duration, area, season, region, and precipitation depth). The stochastic model for temporal disaggregation of rainfall data was evaluated across Texas. The hourly historic data from the selected 531 hourly gauges in Texas were used to evaluate the modelAcA?A?s performance to reproduce hourly rainfall statistics. Spatial trends in performance statistics or spatial patterns among gauge characteristics (e.g. period of record, precipitation statistics) were examined by cluster analysis. Since no spatial trends or patterns were identified, the state database is used and verified for a selection of gauges. The method was further applied to estimate intensity-duration curves for hydrologic applications. To obtain basic information on the spatial and dynamic patterns of rainfall over an area, it is necessary to identify and track a storm objectively. Automated algorithms are needed to process a large amount of radar images. A methodology was presented to overcome the identification and tracking difficulties of one-hour accumulated distributed rainfall data and to extract the characteristics of moving storms (e.g., size, intensity, orientation, propagation speed and direction, etc.). The method presented in this dissertation allows the user to better understand the precipitation patterns in any given area of the United States, and yields parameters that describe storm dynamic characteristics. These parameters can then be used in the definition of synthetic dynamic storms for hydrologic modeling.
