Identification and Restoration Potential of Urban-Rural Forest-Water Composite Ecosystems

<sec><title>Objective</title> Shanghai, a densely populated megacity with a dense water network, faces challenges such as fragmented forest – water configurations, weak coordination among ecosystem functions, and the lack of standardized frameworks for integrated blue – green spatial planning. These issues constrain ecological capacity and urban planning effectiveness. In response to the urgent need for promoting “forest – water composition” and “water – green integration” as key directions in ecological spatial governance, this study proposes a technical framework for identifying forest – water composite zones and evaluating their restoration potential, aiming to provide spatially explicit ecological management and restoration strategies. </sec><sec><title>Methods</title> Taking Shanghai as a case study with its complex hydrological background and dynamic land use, this study uses high resolution satellite images from 2021 to 2023 were processed on the Google Earth Engine platform. Forests and water bodies were extracted using Sentinel-derived Dynamic World land cover products and spectral indices (EVI, NDVI, MNDWI, and LSWI). Seasonal water extent was delineated from maximum water distribution between April and September (2021−2023) to capture hydrological dynamics. A multi-step classification and evaluation system was constructed. Water – forest adjacency was quantified using a spatial adjacency index (shared perimeter over total water perimeter), and water body complexity was measured through the perimeter – area ratio. Dynamic characteristics of vegetation and water cover were calculated at monthly and seasonal scales to establish the seasonal (<italic>I</italic>_s) and monthly (<italic>I</italic>_m) dynamic indices. Based on spatial proximity and temporal variability, forest – water composite zones were classified into three categories: 1) Ecological integration zones (adjacent with sedimentation); 2) basic forest – water interface zone (adjacent but weak dynamics); 3) water – forest zones to be restored (sedimentation but lacking forested edges). Spatial patterns are analyzed using Getis-Ord Gi* statistics and nearest neighbor analysis. Restoration potential is assessed through dynamic indicators, spatial adjacency, and available surrounding land. </sec><sec><title>Results</title> The results reveal distinct spatial differentiation. The ecological integration zones, basic forest – water interface zones, and water – forest zones to be restored respectively occupy an area of 128.90 km², 447.54 km², and 25.56 km², representing 19%, 66% and 4% of Shanghai's total water area. Ecological integration zones are primarily distributed in outer districts such as Qingpu and Chongming, corresponding to sediment-rich lakes and wide rivers with forest margins. Basic forest – water interface zones are more evenly spread but concentrated in central districts (e.g., Huangpu, Yangpu, Xuhui), where adjacency exists but dynamic transformation is minimal due to shoreline hardening. Water – forest zones to be restored are typically located at the margins of open water and disturbed lands, including abandoned ponds and silted tributaries. Dynamic analysis shows the highest ecological fluctuation in zones to be restored (<italic>I</italic>_s = 0.41; <italic>I</italic>_m = 0.26), suggesting strong seasonal responsiveness and vulnerability; ecological integration zones exhibit moderate variability (<italic>I</italic>_s = 0.30; <italic>I</italic>_m = 0.15), indicating stable connectivity with restoration potential; basic forest – water interface zone remain largely static (<italic>I</italic>_s = 0.04; <italic>I</italic>_m = 0.01), often due to artificial modification. Urban – rural gradient analysis reveals significant heterogeneity. Suburban districts such as Songjiang and Jiading host larger composite patches, with significant clustering (<italic>p</italic> &amp;lt; 0.01), implying high restoration opportunities. In contrast, central areas show fragmented, random distributied patches. Statistical tests confirm no significant relationship between forest – water composite level and water area or shape complexity (<italic>p</italic> &amp;gt; 0.05), indicating that composite potential is primarily driven by anthropogenic regulation and policy interventions rather than natural morphology. </sec><sec><title>Conclusion</title> This research establishes a standardized, scalable classification and evaluation framework for forest – water composite ecosystems, applicable to complex urban landscapes. Through spatial disaggregation and dynamic assessment, the research uncovers the multi-scalar heterogeneity and ecological transformation patterns of Shanghai’s forest – water systems, enabling precise zoning, targeted restoration, and evidence-based planning. The research further proposes a governance model based on “core – corridor – reserve” spatial logic: Preserving ecological integration zones as biodiversity-rich ecological cores, enhancing basic forest – water interface zones as green – blue corridors, and prioritizing water – forest zones to be restored through adaptive restoration tailored to hydrological and vegetative feedbacks. In central urban areas, vertical ecological integration technologies (e.g., sponge structures, terrace planting) are recommended to overcome spatial constraints, whereas in suburban districts, horizontal corridor expansion is prioritized. The proposed methodological system responds to the urgent need for spatially explicit, process-informed planning tools in water-rich, development-intensive cities. By integrating structural and dynamic metrics, this framework advances understanding of composite ecosystem resilience and provides a practical toolset for restoration prioritization under future climate and land use scenarios. The findings have broader implications for ecological governance in deltaic and river-network cities, offering a transferrable reference for implementing synergistic blue – green infrastructure strategies. </sec>