Fragmentation Trajectories as a Review of Existing and Proposed Single-valued Fragmentation Indices

Two related single-valued landscape fragmentation indices D and F are proposed, based on patch aggregation, shape complexity, and percent of the focal pixels on the landscape, and are computed using Fragstats metrics on a colonization landscape continuously fragmented over 36 years. The same was done for two existing single-valued fragmentation indices, i.e., the Matheron index based solely on normalized unlike joins, and the Normalized Hypsometric Curve (NHMC) index from GUIDOS Toolbox. All were plotted chronologically, and also against percent non-forest (%nf) of the landscape, and the trajectories were compared for behavior. The NHMC index starts high even if deforestation is low, and continues increasing even further as deforestation continues, while the other three indices all start close to zero and increase gradually. F mimics D very closely, and the Matheron index only behaves differently from F and D at the end of the data range. The deviation may be due to patch aggregation, which the Matheron index does not consider. An accepted single-valued fragmentation index computed from Fragstats landscape metrices could allow for cross-study comparisons relating fragmentation with any other attribute on or of the landscape, hopefully advancing the science of fragmentation in landscape ecology as cross-study generalizations would now be possible.