.|  Baltimore Ecosystem Study
Forest Patch Change in the Gwynns Falls Watershed
  • Mary L. Cadenasso, University of California at Davis, Davis, California
  • Steward T.A. Pickett, Cary Institute of Ecosystem Studies, Millbrook, New York
  • Elizabeth M. Cook, Cary Institute of Ecosystem Studies, Millbrook, New York
The purpose of this project is to quantify the temporal and spatial changes of forest cover in the Gwynns Falls watershed in Baltimore, Maryland. The Gwynns Falls watershed is the focal watershed of the BES and is located along an urban-to-rural gradient. While the focus of the project is on the change of forest patches across years, the forest change in relation to distance from the urban core is also examined. Three study sites in the watershed are used to represent the urban to rural gradient; Rognel Heights near the urban core, McDonough in a suburban landscape, and Glyndon in a rural area of the watershed. The cause of the change in forest cover was also quantified by categorizing the other landscape covers in each year. The land cover categories include forest, agriculture, open field (recreation and cemeteries), development (residential and commercial), roads, and bare ground.
Motivating Questions
  1. Does the total area of forest in the Gwynns Falls watershed change over time? Does this temporal change differ as the distance from the urban core is increased?
  2. How do forest patches change in configuration (number, area, edge length, etc) over time in the Gwynns Falls Watershed? And, what are the dynamics associated those changes – loss and gain of forest area and the causes of change, including urbanization, development, and agriculture?
  3. Are there differences in how forests in urban, suburban and rural areas change through time?
  4. What are the major landscape changes (rise of suburbia, loss of farmland, etc) that correlate with the change in forest patches over time?
The forest cover was digitized using imagery from 1914, 1938, 1957, 1971, 1999, and 2004. Forest patches for each year were delineated as a separate layer on top of the digital imagery to a resolution of 0.5 ha. Using ArcMap and ArcView a variety of spatial analyses were completed on the forest patches across the whole watershed and within the three representative study sites. Because the extent of each year’s imagery differed, the whole watershed analyses only encompass the area for which data was available for all years, represented by the black outline. Analyses included examining the forest patch dynamics and configuration, examining how forest patch dynamics are affected by distance from the urban core, and quantifying why forest patches changed in size in each year to year transition. The specific forest patch dynamics examined include amount of forested area, number of forested patches, mean and median patch size, as well as edge density and mean patch edge in each year.

Preliminary Results
The amount of total forested area, within the area of the watershed encompassed by all data sets, remained relatively constant. However, there is a large increase in the number of forest patches and the amount of forest edge between 1914 and 2004. The cause of forest patches decreasing in size between 1914 and 1938 was dominated by new agricultural fields, whereas the decrease in forest patch size between 1938 and 1957 was predominately due to new development (residential and commercial). New development remained the leading cause of forest patches losing area after 1957.