D. J. Leu


A technique has been designed and tested that permits the quantitative assessment of historic marsh conditions from black and white aerial photography. Utilizing a computerized scanning and writing Rotating Drum Microdensitometer (RDM) system, baseline marsh data has been derived by combining present day field data with computer analysis of present day and historical aerial photography. Quantitative information on the marsh grasses studied (Spartina alternifora - tall and short growth forms, Salt Hay - a mixture of Spartina patens and Distichlis spicata), their distributional patterns, and surface water locations were established.

The RDM technique relates subtle optical density differences in a film emulsion to actual field conditions. Algorithms are then used to classify and map the marsh variables being studied. In this investigation, photographs were scanned using picture elements (pixels) as small as 25 µm per side. For 1:12,000 scale photographs, this translates to a ground equivalent area of 0.96 square feet.

The RDM system measures the optical density of each pixel and assigns it one of 256 different density levels. This is approximately an order of magnitude greater than the human eye's ability to differentiate subtle shades of gray. Once computer analysis is completed, the RDM system then prints the classification categories on to an emulsion, producing a film image that maps the marsh variables.

Classification maps were produced by the RDM system for two geographically separate saline marshes. Each marsh was in excess of 100 acres. Using a Zoom Transfer Scope, the RDM classification maps were transferred and compared to historic wetland maps produced from manual interpretation of color infrared photographs and then extensively field surveyed. The wetland maps were produced by the New Jersey Department of Environmental Protection.

Except for a few small areas where S. alterniflora was misclassified, vegetative cover classes were correctly classified. The vegetational boundaries and land/water interfaces were located with great accuracy.

Date of this Version