城市生态网络是景观生态学应用领域研究的热点和重点之一,识别、评估生境之间的连接是构建生态网络的关键环节。在总结已有连接辨识方法的基础上,提出采用最小费用模型和图论分析相结合的方法,探讨功能性连接的辨识和优先恢复途径。以新西兰基督城为案例,分别利用景观发展强度指数建立阻力面,新西兰鸡毛松( Dacrycarpus dacrydioides)种子最大传播距离作为连接阈值来模拟、评价网络连接。结果表明：在1200 m 距离阈值下,共有408条连接,其重要性分为10类。其中Richmond—Petrie公园,Hansons—Auburn保护地,Centaurus公园—King George保护地是整个生态网络的关键连接；若去除,景观整体连接度将下降31.73%。此外,研究发现连接重要值与两端的源面积之和没有显著相关性,即面积大的源斑块之间的连接不一定对网络构建起关键作用,这一结论还有待进一步证明。针对缺少动物迁移资料的城市环境,改进最小费用模型和网络连接分析的部分参数；可操作性与实用性强,对中国城市区域生态恢复建设、栖息地选择具有借鉴意义。%With rapid urbanization and industrialization, habitat fragmentation and loss are inevitable. Under these circumstances, landscape connectivity and ecological networks have become a focus of applied landscape ecology. A well-connected ecological network is believed to facilitate energy and resource fluxes, species dispersal, genetic exchange and multiple other ecological processes, and to contribute to the maintenance of ecosystem stability and integrity. Identifying and evaluating functional connectivity between habitat patches is the key step in designing and building well-connected ecological networks. Based on a review of literature on linkage identification approaches, our study combined least-cost path modeling with graph-theory based network analysis to simulate, identify, evaluate, and prioritize functional linkages. We made some modifications to conventional least-cost model and network analysis in order to reduce subjectivity and to better suit the specific conditions of urban environments. Least-cost models are commonly used to determine the movement routes of wildlife to optimize conservation of meta-populations in wild environments. Ideally, least-cost models should be parameterized with field data for a specific organism ( usually a small animal);however, these data are difficult and time-consuming to collect, and this has resulted in a large number of studies relying entirely or in large part on collective expert opinion. Different experts may assign different values and therefore affect the reliability of network simulation. Therefore, we adopted the Landscape Development Intensity ( LDI) index, which is a measure of human disturbance of ecosystems, to quantify the relative costs of land use/cover types. In addition, we chose not to use small mammals as the individual surrogate for designing our urban ecological network. Most studies of animal movement patterns have been conducted in continuous forest landscapes, which are very different from patchy urban environments. These are characterized by much smaller and fragmented habitat patches, surrounded by a non-living concrete matrix with various threats hostile to life. Instead, we used native trees as individual surrogates and their seed dispersal pattern, based on a field survey, for devising and simulating a functional urban ecological network. Taking Christchurch, New Zealand, as an example, our study used a climax native forest tree, Dacrycarpus dacrydioides as a surrogate in establishing our urban ecological network. We used the LDI index to build the cost surface for a least-cost model in ArcGIS 10. 0, and the observed maximum seed dispersal distance of D. dacrydioides (1200 m) as the threshold distance for network analysis in Conefor 2.6. Under the 1200 m threshold distance, a total of 408 links were simulated in the study area. These were grouped into 10 importance grades using Hierarchical Cluster Analysis in SPSS 21.0 combined with input from decision-makers. Among these, the three most important links were between Richmond Park and Petrie Park, Hansons Reserve and Auburn Reserve, and between Centaurus Park and King George Reserve. They were identified as the cut-links in a landscape graph, i. e. removal would disconnect critical components in the landscape. Indeed, if these links were absent, the Integral Index of Connectivity ( IIC ) of the whole landscape would be reduced by 31. 73%. Additionally, our study found that there was no linear relationship ( Pearson′s r=0.078, P=0.118) between link importance value and total area of habitat connected. That is, the links between larger habitats are not necessarily more important to the network. Further studies are needed to support this finding. Overall, our case study suggests that the combination of least-cost pathway analysis and graph-theoretical algorithms is an efficient approach for simulating and prioritizing functional connectivity in urban landscapes, and might be valuable for Chinese urban ecological network planning. Furthermore, the energy-consumption based LDI index was shown to be a reliable measurement of relative cost values of land use/cover types. However, it should be pointed out that the LDI index was calculated based on western land use/cover types, which are different from those used in China. It is recommended that when being applied in China, the LDI index should be adjusted to China′s land use/cover types.
展开▼
