Residential outdoor landscapes constitute the most frequently encountered visual environment for urban residents, yet systematic evaluation of their aesthetic quality and ecological service value remains lacking. This study developed an integrated three-stage framework combining deep learning semantic segmentation, the Scenic Beauty Estimation method, and machine learning regression modeling to assess public visual aesthetic preference for residential outdoor landscapes, understood as one measurable component of the cultural ecosystem services provided by urban residential green spaces. A PSPNet model performed pixel-level semantic segmentation on 2816 field-collected landscape images from 32 residential communities in central Chengdu (mIoU = 0.763), extracting 13 element categories to construct a residential landscape feature indicator system. Public aesthetic preference scores were obtained through 276 valid questionnaires. Pearson correlation analysis identified the Green Visual Index (GVI) as the strongest positive predictor (r = 0.57), while the Interference Factor Index (IFI) showed the most pronounced negative correlation (r = − 0.46). The random forest model achieved R2 = 0.796, substantially outperforming the stepwise OLS regression (adjusted R2 = 0.517). SHAP interpretability analysis revealed a marginal contribution saturation effect for GVI at approximately 35–40
As urban ecological conservation receives increasing attention, coordinating green infrastructure (GI) spatial structure with biodiversity conservation has become an important issue in urban ecological planning. However, a highly structured GI does not necessarily correspond to high bird habitat (BH) suitability, and existing research has paid relatively little attention to the spatial congruence and mismatch relationships between the two, and how these can be further translated into ecological network planning strategies. To address this issue, this study employs Hangzhou as a case study. Green infrastructure assessment (GIA) was applied to evaluate GI spatial structure across three dimensions: scale, morphology, and pattern. The Maximum Entropy (MAXENT) model is utilized to assess BH suitability. The four-quadrant model is applied to identify spatial congruence and mismatch zones between GI and BH, based on which three optimization objectives are proposed. Using source patch extraction- resistance surface construction-circuit theory, multi-objective ecological networks (ENs) are developed. Finally, structural resilience assessments of connectivity and robustness are conducted for three ENs using connectivity indices and network attack simulations. The principal findings of this study include: (1) The spatial structure of GI and BH suitability generally exhibit a positive spatial correlation; with some regions displaying high or low values align, while other regions demonstrate mismatched relationships. (2) The three ENs exhibit distinct differences in source area size and composition, corridor number and length, and network topology, reflecting the spatial organizational characteristics associated with different objectives. (3) EN structures that achieve dual-objective coordination exhibit optimal resilience. This study clarified the spatial relationship between GI structure and BH suitability and provided methodological insights for integrating biodiversity conservation with the allocation of blue-green spaces to support sustainable urban development.
The rapid expansion of urban areas, along with the increase of impervious surfaces and encroachment on floodplains, has resulted in significant challenges related to flood and climate risk, water quality, biodiversity, and preservation of essential ecosystem services. An effective approach to mitigate these impacts lies in Nature-based Solutions (NbS), which have emerged as promising strategies for enhancing both environmental sustainability and human well-being, and they have been demonstrated to play a pivotal role in climate change adaptation and mitigation. This research aims to quantify the potential effects of NbS on water- and climate-related ecosystem services over a specific site in Legnano city which is a peri-urban zone, western the Milan metropolitan area in the North-West of Italy. The study adopts an integrated approach to urban regeneration, examining how the implementation of NbS can enhance the provision of various ecosystem services. Specifically, pluvial flood storage, heat island mitigation and biodiversity are quantified as primary benefits in the study area. The paper emphasizes the improvement of service indicators through the NbS approach that can be applied to all urban areas, fostering the development of resilient cities and ensuring the well-being of both citizens and ecosystems.
In China, driven by policies that promote the concentrated and contiguous preservation of traditional villages, the concept of regional and holistic conservation has gained broad acceptance among academia and practitioners. As heritage-preserving entities, traditional villages can be conceptualized as populations sharing and embodying traditional Chinese culture. This study adopts a population-ecology perspective, analyzing 413 traditional villages in Jiangxi Province, to examine village clustering patterns and population relationships, identify traditional village clusters, and propose cluster-based conservation measures. Ripley’s K function, kernel density analysis, entropy analysis, gravity models, and hierarchical nearest-neighbor clustering were employed to examine village clustering patterns and population relationships. The most pronounced clustering effect in Jiangxi’s traditional villages occurs within a 30 km radius. Villages with strong and those with weak competitiveness are spatially intermixed. Villages exhibiting strong symbiosis are concentrated in the Luling Cultural Zone, Linchuan Cultural Zone, and Huizhou Cultural Zone. Jiangxi Province contains nine traditional-village clusters, including the Fuliang Cluster and Wuyuan Cluster, primarily distributed in central and eastern Jiangxi Province. Each cluster region should leverage traditional culture to rebuild competitive advantages, enhance symbiosis through key nodes, and strengthen the circulation of key factors to enhance agglomeration effects. Following the ‘population characteristics-population relationships-population evolution’ framework, the delineation of traditional village clusters based on population traits and relationships, and the proposal of conservation strategies, offer a potentially effective pathway for protecting traditional village clusters.
This study aimed to investigate the difference in bird collisions related to stickers on the glass of noise barriers in Ansung and Pyeongtaek cities, Gyeonggi province, South Korea. We surveyed for bird collisions at each noise barrier in every three weeks from September 2023 to August 2024. We measured landscape variables (e.g., coverage of paved areas, buildings, agricultural land, bare ground, woody vegetation, and surface of water bodies) within a 500 m radius from the noise barriers using QGIS. The coverage of all landscape variables did not differ between noise barriers with and without stickers. The number of carcasses and species was higher in noise barriers without stickers than in those with stickers. The size of the glass, land cover, shape, and season were highly related to the number of carcasses and species of collided birds. The results of our study suggest that constructing noise barriers with stickers on the glass may be an effective method to reduce bird collisions. Further research is required to investigate the conservation implications of collisions concerning bird community and their habitats.
