A GIS Method to Summarize Changes Among Classes During a Time Series With an Application to Land Cover in Western Bahia, Brazil

ABSTRACT Petabytes of raster maps exist that show numerous classes across a series of many time points. Scientists need new methods to summarize the temporal changes so scientists can understand broad patterns and have guidance for how to look deeper into particular places, trajectories, classes, and time intervals. Our manuscript addresses this need by presenting an innovative method that summarizes the temporal changes in terms of a map of four trajectories and bar charts of five components: quantity, allocation exchange, allocation shift, alternation exchange, and alternation shift. An application illustrates the method using 16 land cover classes at 7 time points from 1990 to 2020 in western Bahia, Brazil. Results reveal a substantial component of alternation shift, which exists when pixels transition from savanna to temporary crops to soybean. Readers may apply the method using the app available for free at https://github.com/antoniovfonseca/summarize‐change‐components .