| The rapid urbanization process has a significant impact on species diversity and distribution, community types and structure, even population regeneration and community succession. Since1992, Qingdao has experienced rapid and intense urbanization, and formed a special urban space expansion mode while the remaining natural vegetation is preserved well on hills in urban area. This study summarized the urban development process and analyzed the status of land use of Qingdao, according to which the urban-rural gradients were divided, built a vegetation classification system for Qingdao urban forests, and analyzed the community assemblage rules and dynamics of Qingdao urban forests. Moreover, fragmentation effects on species diversity were studied on remaining natural forests in hills scattered in urban region. We hope this study can provide the theoretical support and scientific basis for "Qingdao Forest City" and landscaping construction. The main conclusions are as follows:(1) Generally, Qingdao urban space expansion mode was divided into two phases: axial belt expansion mode before1992and multi-center expansion mode after1992. The land use type analysis based on2009remote sensing image showed that, the regions where impervious land percentage greater than90%are consistent with the earliest urbanized area before a century; the regions where impervious land percentage between70%and90%are the extension of old urban area and new urban centers; the regions where impervious land percentage between40%and70%are the extension of new urban area; the regions where impervious land percentage lower than40%are hill forests and farm land in villages and towns. According to these, urban-rural gradients were divided into high-density urban areas, medium-density urban areas, low-density urban areas and non-urban areas, in Qingdao.(2) Three transects were established along urban-rural gradients for urban forests in hills (HF), parks (PF) and aside roads(RF), respectively. The surveys showed that 111genera,52families,189species of woody plants were recorded in336plots, totally. In56%level,51community types were divided based on clustering analysis. We established a four-level vegetation classification system for Qingdao urban forest. In the first level, urban forests were divided into two groups:remain natural forests and plantations. In the second level, both of two groups were classified into vegetation type groups by life form composition. In the third level, all vegetation type groups were classified into formation groups by constructive species’ affinity. Finally, all formation groups were classified into formations by dominant species. In summary, there were one vegetation type group, two formation groups and three formations for the remaining natural forests and four vegetation type groups, fifteen formation groups and fifty formations for plantations.(3) Species richness and individual density along urban-rural gradients were different among HF, PF and RF. In PF, there were no significant variations for both richness and individual density. In HF and RF, from high-density urban areas to non-urban areas, it showed an increasing tendency in richness and individual density for native species, but decreased tendency in richness for alien species with consistent individual density. Because the increasing rate of native species was greater than the decreasing rate of alien species in HF, it showed an increasing trend of species richness and individual density, totally. In contrast, there was no significant variation in RF due to the similar relative variation rates between the increasing trend of native species and decreasing trend of alien species.(4) Multiple regression and canonical correspondence analysis (CCA) were used to determine the effects on species richness and species composition, by using sixteen influencing factors that divided into three types:natural environment (NEFs), community structure (CSFs) and human activities (HAFs). Native species richness was influenced by all three types factors in which HAFs was the most important for both of HF and PF, and was only determined by CSFs for RF. Alien species richness was determined by HAFs in HF and determined by NEFs&CSFs in PF and RF. In addition, the explanations of variation of native species richness decreased along HF-PF-RF with increased human management intensity. Inversely, the explanations of variation of alien species richness increased. All factors had little contribution on explaining the community species composition according to CCA analysis, in which human disturbance intensity was the most effective factor.(5) Population size-structure was analyzed for main species in51communities and classified into five basic regeneration patterns:reverse-J type, L-type, sporadic type, unimodal type and unibar type. According to frequency in five basic regeneration patterns, regeneration potential of each species was determined, and then the short-term dynamics of communities were predicted, based on which communities were summarized into four groups:self-sustaining community, disturbance-dependent sustaining community, coexist species dominance replacing community and post-colonized species dominant community.(6) For seedlings, richness and individual density mainly depends on CSFs and HAFs. Density of individual taller than1.5m was the greatest influencing factor in the former, which suggested a significant contribution of seed production on seedling recruitment. Human disturbance intensity was the greatest influencing factor in the latter, which suggested forest management and utilization had negative effects on seedling recruitment. In addition, the explanations of variation of seedling richness decreased for native species but increased for alien species, along remain natural forests-plantations with increased human management intensity, which could be explained by higher native species richness in HF and higher alien species richness in RF.(7) To explore the effects of habitat fragment on remnant natural forests, CCA was used to analysis the relationship between species composition (present-absence data) and four factors:area, distance to species pool, human disturbance intensity and the isolation degree. The length of vector in CCA biplot was used as weights for each influencing factor, to calculate the sum, named isolated island index (Ⅲ). Then, the relationships between III and species richness were simulated by linear model, logarithmic curve, power curve and S curve. The results indicated that relationship between III and richness was significant negative for native species, but not significant for alien species richness, and relationship between III and richness was significant negative for bird dispersal species and wind dispersal species, but not significant for gravity dispersal species. Moreover, logarithmic curve was the best simulating model for any life form group, except for evergreen broadleaved species, shrub species and coniferous species which cannot be simulated by all four models. All those results indicated that the negative fragmentation impact on Qingdao urban forest which belongs to warm temperate deciduous broad-leaved forest region was mainly embodied in tree native deciduous broad-leaved species with great dispersal ability. |
PDF Full Text Request