Urban land uses and their associated impervious cover increase the quantity and worsen the quality of stormwater runoff, which can seriously impair receiving waters. It is known that there is considerable variability in runoff quantity and quality between rain-to-rain events due to rainfall spatial variability. In addition, runoff presents significant variability between neighborhoods, even if they are affected by the same rain event. It is hypothesized that the variability in stormwater quality between sites is associated with the difference in land uses and surface covers. This research examined the variability in stormwater quality characteristics as contained in the National Stormwater Quality Database (NSQD) for different land use categories and nine selected stormwater constituents (TSS, total zinc, total copper, total lead, total phosphorous, dissolved phosphorous, total nitrogen, TKN, and fecal coliform) at different geographical scales - national, regional (EPA Rain Zone 2), and local levels (Jefferson County, AL). The results of the local data analyses were compared to the results obtained from the national and regional analyses. This research also examined the detailed land development characteristics based on actual local field measurements and explained how this variability affects the variability in stormwater characteristics. The land development characteristics information was collected from Little Shades Creek watershed and five highly urbanized drainage areas situated in Jefferson County, AL (in and near the city of Birmingham). About 170 neighborhoods were surveyed in detailed to determine the actual development characteristics and their variability. This research found that each major land use had unique patterns and mixtures of surfaces. These, in addition to the activities taking place within the land uses, affect the runoff quality and its variability from these areas. It was found that there is less variability in stormwater quality characteristics within each land use category compared to the variability between the land use categories. This finding is also true for land cover areas in that there are lower levels of variations in these area types within each land use compared to between the different land uses. The results from this dissertation analyses can be used as guidance for local stormwater quality monitoring efforts, but the specific results are not all expected to be applicable everywhere. The main focus of this research was in investigating stormwater variability, specifically its sources and how it can be reduced. The general concern with the high variability that is associated with stormwater quality is the uncertainty of being able to meet discharge requirements, even with extensive use of stormwater control practices. This uncertainty can be eliminated, or at least reduced, by a better understanding of sources of this variability. Specifically, appropriate discharge regulations that recognize this variability will assist the stormwater managers to better use their financial resources and to maximize receiving water quality improvements.