Error propagation in raster data integration: Impacts on landscape composition and configuration

Integrating raster-based categorical maps from multiple sources necessitates the transformation of geometric characteristics to compare maps, as in land change analyses. By projecting maps to a new geographic reference framework and scaling pixel values to a new size, distortions of map information are introduced that can affect the proportion and arrangement of thematic classes across the landscape. Using a sample land cover dataset depicting a heterogeneous landscape, this paper examines these impacts using three common raster-based transformation methods and introduces a new vector-based method that minimizes error propagation. While relative class area was best preserved by a nearest-neighbor resampling method, distortions to the contiguity of thematic classes and the overall fragmentation of the landscape were minimized when using the vectorbased projection and resampling method. Results demonstrate that more than a third of pixel values of a categorical map may be affected by common projection and scaling methods and reinforce the need for careful attention to impacts of error propagation in categorical data transformations.