Keywords
aerial photography, barrier island, change detection, epsilon bands, fragmentation, salt marsh
Start Date
1-7-2006 12:00 AM
Abstract
Barrier islands and coastal salt marshes are complex ecosystems that move and change throughtime in response to many factors. For example, hurricanes bring strong winds, rain, and storm surge whichcan greatly change the distribution of surficial deposits. Through time the islands can migrate and inletschange their positions. A study of back-barrier salt marshes in southeastern North Carolina, USA, wasconducted to map changes in the distribution and fragmentation characteristics of the marshes. Topsail Islandand Masonboro Island were chosen as comparative study sites since Topsail has had increasing urbanizationsince the 1930s while Masonboro is a protected, undeveloped, island. By gathering, rectifying, interpreting,and digitizing historical aerial photography (from 1938 to 2002) we computed the rate of change in the backbarrierland cover types as well as used GIS spatial analysis tools to compute the degree of fragmentationthrough time and place. Results have been mixed where the marshes behind Masonboro Island are mostaffected by storms while Topsail Island marshes have changed mostly due to urbanization and inletlocation/migration.To quantify the significance of this historical change, a series of tests were designed and conducted todescribe the amount of spatial variability and accuracy of the rectified photographs, the digitized polygons,and the quantification of change. A digitizing accuracy assessment was conducted where 140 randomlychosen locations were identified on the aerial photographs and compared with the digitized data. Using anerror matrix, the overall accuracy was greater than 80 percent which was acceptable. Second, we measuredthe impact that the degree of crenulation, or curviness of the digitized marshes had on the change detectionresults. To compute this we used a subset of the study area, used progressive smoothing functions (from 5mto 70m), and recomputed the change detection matrices. Results indicated that the interpretation of thephotographs and the resulting digitization was not a factor in the computation of change.Third, we incorporated a fuzziness test (using derived epsilon bands) into the GIS data to identify andquantify real changes in the marsh habitats versus positional changes or sliver polygons. Results indicate thatrectification of aerial photography (although an RMS error of less than 1), photointerpretation, and digitizingcan lead to some erroneous results however by using fuzziness techniques we can minimize the errors andpredict which areas are changing through time. Statistically, the removal of small polygons of change usingthe epsilon band method did not alter the general outcome of the change detection analyses, however it is aworthwhile data processing method.
Sensitivity Analysis of Historical Aerial Photography: A Case Study of Barrier Island Marshes
Barrier islands and coastal salt marshes are complex ecosystems that move and change throughtime in response to many factors. For example, hurricanes bring strong winds, rain, and storm surge whichcan greatly change the distribution of surficial deposits. Through time the islands can migrate and inletschange their positions. A study of back-barrier salt marshes in southeastern North Carolina, USA, wasconducted to map changes in the distribution and fragmentation characteristics of the marshes. Topsail Islandand Masonboro Island were chosen as comparative study sites since Topsail has had increasing urbanizationsince the 1930s while Masonboro is a protected, undeveloped, island. By gathering, rectifying, interpreting,and digitizing historical aerial photography (from 1938 to 2002) we computed the rate of change in the backbarrierland cover types as well as used GIS spatial analysis tools to compute the degree of fragmentationthrough time and place. Results have been mixed where the marshes behind Masonboro Island are mostaffected by storms while Topsail Island marshes have changed mostly due to urbanization and inletlocation/migration.To quantify the significance of this historical change, a series of tests were designed and conducted todescribe the amount of spatial variability and accuracy of the rectified photographs, the digitized polygons,and the quantification of change. A digitizing accuracy assessment was conducted where 140 randomlychosen locations were identified on the aerial photographs and compared with the digitized data. Using anerror matrix, the overall accuracy was greater than 80 percent which was acceptable. Second, we measuredthe impact that the degree of crenulation, or curviness of the digitized marshes had on the change detectionresults. To compute this we used a subset of the study area, used progressive smoothing functions (from 5mto 70m), and recomputed the change detection matrices. Results indicated that the interpretation of thephotographs and the resulting digitization was not a factor in the computation of change.Third, we incorporated a fuzziness test (using derived epsilon bands) into the GIS data to identify andquantify real changes in the marsh habitats versus positional changes or sliver polygons. Results indicate thatrectification of aerial photography (although an RMS error of less than 1), photointerpretation, and digitizingcan lead to some erroneous results however by using fuzziness techniques we can minimize the errors andpredict which areas are changing through time. Statistically, the removal of small polygons of change usingthe epsilon band method did not alter the general outcome of the change detection analyses, however it is aworthwhile data processing method.
