LAUSR

Abstract Ecological restoration programs (ERPs) have greatly expanded forested areas in China to mitigate environmental problems. Unfortunately, this general increase in forest cover has not revealed clear changes in the states of fragmentation and loss of connectivity of forested habitats, and few studies have investigated the effects of ERPs on the spatial patterns of forests. Taking the Three Gorges Reservoir Area as a case study, we assess and quantify how the dynamics of the spatial patterns of forest following massive ERPs from 1990 through 2015 by morphological spatial pattern analysis (MSPA). We examine possible landscape structure thresholds in the forest restoration gradient. The results indicate that dramatic decreases in forest fragmentation and patchiness have occurred since ERP implementation. Among the seven MSPA forest types, core forests show the highest increase of 513.91 km2·yr−1. ERPs promote the defragmentation of forests via two pathways: (1) Old forests are preserved and expanded by building new forests around existing patches and closing perforations within forests; and (2) new core forests are gradually created in areas far from older forests. However, isolated forest patches surrounded by cropland production areas formed after the conversion of croplands into forests have appeared across the study area. Furthermore, at least two points of nonlinear changes in the landscape structure with increasing forest cover are identified. These nonlinear trajectories can be used as indicators of critical shifts in ecosystem states, and relevant thresholds support the identification of specific conservation actions. Our findings demonstrate the importance of ERPs for the defragmentation of forest patches and provide a basis to establish a more effective arrangement of patches necessary to mitigate fragmentation effects. Given that landscape structural changes are not expected in ERP design, future ERPs should focus on remaining isolated forest patches and on maintaining the health and sustainability of landscapes.