The positive relationship between species richness and area is a fundamental principle in ecology. However, this pattern deviates on small islands, where species richness either changes independently of area or increases at a slower rate—a phenomenon known as the Small-Island Effect (SIE). While the SIE has been well documented in natural ecosystem, its presence in highly fragmented and disturbed urban ecosystem remains unexplored, posing challenges for urban vegetation conservation. Urban remnant vegetation, isolated by surrounding infrastructures, preserves intact zonal vegetation characteristics, serves as a benchmark for restoring near-natural habitats and offers ideal conditions to test the existence of the SIE in urban area landscapes. In this study, we surveyed 17 remnant vegetation patches in Qingdao City, China. A total of 331 plants attributed to 255 genera in 81 families have been recorded. Firstly, by using six species–area relationship regression models testing the SIE for remnant vegetation with different plant life forms, we found the SIE in only woody plants, with the land surface area threshold ranging from 6.38 ha (tree) to 11.91 ha (shrub). Our finding revealed that the drivers of the SIE in shrubs were landscape shape index, perimeter–area ratio, and the proportion of sealed surfaces within the patch. For trees, the SIE was influenced by the distance to the source of species, GDP, night light intensity, and perimeter–area ratio. This finding justifies that conservation in urban planning, construction, and development should focus not only on protecting large areas but also on maintaining and promoting diverse habitats within these areas. At the same time, reducing anthropogenic disturbance and enhancing the connectivity of green spaces are important for the persistence of metacommunities and can contribute to the local species pool, thus potentially improving the ecological resilience of urban environments.

The positive relationship between species richness and area is a fundamental principle in ecology. However, this pattern deviates on small islands, where species richness either changes independently of area or increases at a slower rate—a phenomenon known as the Small-Island Effect (SIE). While the SIE has been well-documented in natural ecosystem, its presence in highly fragmented and disturbed urban ecosystem remains unexplored, posing challenges for urban vegetation conservation. Urban remnant vegetation, isolated by surrounding infrastructures, preserves intact zonal vegetation characteristics, serves as a benchmark for restoring near-natural habitats and offers ideal conditions to test the existence of SIE in human-altered landscapes. In this study, we surveyed 17 remnant vegetation patches in Qingdao City, China. In total of 331 plants attributed to 255 genus in 81 families been recorded. Firstly, by using six species-area relationship regression models tested SIE for remnant vegetation with different life form (i.e. annual herb, perennial herb, shrub and tree), we found SIE detected in only woody plants, with area threshold ranging from 6.38 ha (tree) to 11.91 ha (shrub). According to stepwise and generalized linear regression mode analysis we found SIE detected in only woody plants, with area threshold 6.38 ha for tree and 11.91 ha for shrub. Further analysis revealed that the SIE in shrub was driven by landscape shape index, perimeter-area ratio, and the proportion of sealed surface within patch. For trees, the SIE was influenced by the distance to the source of species, GDP, night light intensity and perimeter-area ratio. This finding justifies that conservation in urban planning, construction and development should focus not only on protecting large areas, but also on maintaining and promoting diverse habitats within these areas. At the same time, reducing anthropogenic disturbance and enhancing the connectivity of green spaces are important for the persistence of metacommunities and can contribute to the local species pool, thus potentially improving the ecological resilience of urban environments.

The recuperation factors (negative air ion concentration, airborne particulate matter, human comfort index, and acoustic environment index) of coastal green spaces have significant health effects. Most current studies focus on the distribution pattern of single recuperation factors in the forest environment; however, the comprehensive health effects of coastal green spaces are still unknown. To address this, we analyzed the distribution patterns of single and comprehensive health factors in different landscape configurations, landscape compositions, and coastal distances by principal component analysis and systematic clustering. The results show that: (1) coniferous and broadleaf mixed forests exhibit higher integrated health benefits than other landscape compositions; (2) closed and partially closed landscape configurations exhibit enhanced potential for promoting health benefits as opposed to partially open and open spaces; (3) a coastal distance of 150–300 m offers the strongest comprehensive health benefits. These findings collectively suggest that the increased cultivation of closed and partially closed mixed coniferous and broadleaf forest species at a distance of 150–300 m could effectively provide higher comprehensive health effects. Our study complements the ecosystem service of coastal green areas, especially in coastal health ecological services, providing support for coastal rehabilitation landscape planning; and can help to guide tourists in scheduling coastal health activities scientifically.

To enhance regional carbon sequestration capacity and achieve carbon neutrality, it is crucial to accurately estimate the net primary productivity (NPP) of vegetation and its response to changes in land use/land cover (LULC). Based on remote sensing data, the study estimated the NPP of Qingdao, a typical coastal city in eastern China, from 1990 to 2020 using the CASA model and examined its spatiotemporal dynamics. The Geographic Detector was used in the study to measure the impact of anthropogenic and natural factors on variations in NPP. Additionally, the Markov–PLUS model was utilized to forecast future LULC patterns and NPP changes for 2030 under three development scenarios: natural development, cropland protection, and ecological protection. The results indicate considerable regional variation and an initial decrease followed by an increase in Qingdao’s NPP. It is distributed in a strip-like pattern from northwest to southeast, with the regions around Jiaozhou Bay showing the lowest levels. The changes in NPP are driven by multiple interacting factors; among these, LULC, DEM, and slope exhibit significant spatiotemporal correlations with NPP, with LULC playing a dominant role in this process. In addition, interaction analysis demonstrated that the synergistic effects between LULC and other factors constituted the principal driving force behind NPP dynamics, and the interactions between the two factors that contribute show bivariate or nonlinear enhancement effects. The ecological protection scenario produces the greatest NPP values among the three scenarios. Optimizing LULC may significantly increase carbon sequestration capability, as evident in the NPP forecast, which is higher than that of 2020 in all scenarios. This study demonstrates that the ability of vegetation to store carbon can be effectively enhanced through land-use optimization informed by remote sensing technologies. The results provide a scientific basis for adjusting LULC policies and conducting carbon-neutral spatial planning in coastal urban areas.