Urban forests are increasingly a focus of interest as urbanized populations grow and urban areas expand. Urban forests change as trees are planted, grow, die, and are removed. These processes alter a city's tree cover over time, but this inherent dynamism is poorly understood. Better understanding of how tree cover is a variable land cover component will enhance knowledge of the urban environment and provide new perspectives for management of urban resources.

In this study, tree cover variability within a major urban center was observed over a 20 year period. Changes in tree cover proportion were measured in the District of Columbia between 1984-2004 utilizing highly calibrated satellite remote sensing data. Testing of alternate methodologies demonstrated that an approach utilizing support vector regression provided most consistent accuracy across land use types. Tree cover maps were validated using aerial photography imagery and data from field surveys.

Between 1984-2004, the city-wide tree cover remained between 22.1(+/-2.9)% and 28.8(+/-2.9)% of total land surface area. The District of Columbia did not experience an overall increase or decrease in total tree canopy area. Spatial patterns of tree cover variability were investigated to identify local scale changes in tree cover and connections with urban land use. Within the city, greatest variability was observed in low density residential zones. Tree cover proportion in these zones declined 7.4(+/-5.4)% in the years between 1990-1996 and recovered after 1996.

Changes in tree cover were observed with high resolution aerial photography to determine relative contribution from fluctuation in the number of standing trees and changes in crown sizes. Land cover conversion removed dense tree cover from 50.2 hectares of the city's land surface between 1984-2004.

The results demonstrate that tree cover variability in the District of Columbia occurred primarily within low population density residential areas. Neighborhoods within these zones were analyzed to identify factors correlated with tree cover. Implications of the results include enhanced understanding of the possible impact of urban forest management, and how a focus on low density residential zones is appropriate in setting goals for expansion of urban tree cover.