Geographic Information System Processing of Remotely-Sensed Data for Analyzing Land Cover Change in Cultural Landscapes

Analysis of larger scale designed and vernacular cultural landscapes often requires plan information that documents landscape change over time. Analysis of these landscapes may serve the purposes of survey documentation, National Register nominations, or cultural landscape management and treatment plans. Vegetative land cover is an important dimension of larger landscapes and is related to spatial boundaries, landscape space and volume, land utilization and production, and design patterns. Vegetative land cover is notoriously difficult to compare across time periods because it is subject to fairly rapid change, can be difficult to map and assess for past time periods, and can require extensive field work. Remote sensing for historic landscape documentation within the time period in which aerial photography has been available (generally from the mid-1930s forward in the United States, but as far back as the early 1920s in some localities) has been used to comparatively map landscape change over time (O’Donnell; Watkins). These methods involve intensive manual interpretation of vegetative cover and because of this are typically used on sites of extremely high significance or on relatively small areas. A range of other disciplines including cultural geographers and ecologists have developed various methods for interpreting vegetation change over time (Bender, et al; Ihse; Baily and Inkpen; Herold, et al). A high-accuracy method for automating aspects of the interpretation of remotely-sensed data to aid in the comparison of the vegetative land cover of historic cultural landscapes for different time periods would increase the understanding of spatial and production patterns in these landscapes and benefit more kinds of research and documentation on a wider range of sites of different levels of significance. The research team has done preliminary work with historic black and white and contemporary panchromatic images to a point that proves a concept: that historic aerial imagery and contemporary imagery can be processed and analyzed with GIS techniques to make valid comparisons between historic and contemporary vegetative land cover on cultural landscapes. The proposed project will robustly develop these methods and test them using three different types of contemporary image data. Primary research questions include the following: which kinds of contemporary imagery will yield most accurate results, which classification and processing routines best interpret vegetative land cover, and to what degree are results corroborated by field investigation? The project goal is to develop and disseminate economical, accurate, and easily replicated methods for analyzing vegetative land cover for different time periods on larger scale cultural landscapes.