Abstract:

The combined effects of rapid urbanization and climate change have triggered dramatic fragmentation of natural landscapes and continuous shrinkage of ecological spaces, severely threatening regional ecological security. Ecosystems form structured networks via material cycles, energy flows, and information transmission, with connectivity and resilience key to sustaining biodiversity and functions. Ecological networks enhance species survival by integrating fragmented habitats and facilitating migration, making their protection crucial for regional stability. Current research follows the “source identification-resistance surface construction-corridor extraction” paradigm but has limitations: focusing on static patterns, lacking analysis of resilience dynamics under disturbances; provincial studies rarely explore provincial networks' responses to multi-source disturbances. Jiangsu, a highly urbanized region, faces fragmented landscapes and uneven distribution, with existing resilience assessments failing to reveal disturbance-induced failure mechanisms. Therefore, this study uses land use data from five periods (2000, 2005, 2010, 2015, 2020) and applies MSPA to analyze pixel-level ecological structure and identify key patches for ecological source selection. Woodlands, grasslands, and water areas serve as MSPA foreground (others as background) converted to binary raster data. GuidosToolbox 3.3’s 8-neighborhood analysis yielded 7 functional landscape types, with core areas as potential sources. These are filtered via minimum area threshold, selecting core areas over 3 km² as potential ecological sources. Furthermore, with Conefor 2.6, distance threshold (150 km) and connectivity probability (0.5) are set to calculate PC and dPC indices; core patches with dPC≥0.1 are selected as network sources. Considering regional ecological resistances, land use, slope, road distance, construction land distance, and NDVI were chosen as resistance factors. AHP is used to determine indicator weights, assign resistance values, and derive the ecological network's basic resistance surface via raster calculation to evaluate obstacles to material cycles, energy flows, or biological migration between sources. GIS tool Linkage Mapper series tools are used to calculate the minimum cost paths, pinch point areas, and “barrier point” areas between sources in circuit theory. An ecological network resilience evaluation system is constructed by integrating structural and functional dimensions. Based on complex network theory, ecological sources are abstracted as nodes and ecological corridors as edges to build the topological structure of the ecological network. Pajek software is used to calculate three evaluation indicators: degree centrality, closeness centrality, and betweenness centrality, to quantitatively assess the importance of each ecological source. These evaluation indicators are normalized, and the entropy weight method is used to determine the weight of each indicator to obtain the comprehensive importance of each node. However, with the continuous advancement of urbanization, factors such as urban expansion and transportation corridor construction will damage the landscape ecological pattern; in addition to damage caused by human activities, extreme natural disasters may also lead to fragmentation, degradation, or even disappearance of ecological sources. From the perspective of disaster science, disasters often have spatial proximity effects and cascading effects. Since the destruction and degradation of ecological sources or nodes not only reduce the ecological effect of the node but also affect the migration probability of species from other sources to this source, under the background of source degradation, the connected ecological corridors may also degrade. Therefore, this study simulates and constructs scenarios where nodes in the ecological network face disaster or impact risks and cause partial failure of the ecological network under the dual disturbance conditions of natural and social systems, and uses the Python network analysis library NetworkX to simulate disturbance scenarios to predict the network’s ability to withstand potential risks and quantitatively evaluate ecological network resilience. The Habitat Quality module of the InVEST model is used to quantify habitat quality, and the Zonal Statistics as Table tool in ArcGIS is used to count the habitat quality values of ecological pinch points and barrier point areas. The top 10% of the values are selected as simulated new ecological sources to optimize the ecological network, calculate network resilience, evaluate the optimization effect of the ecological network, and then propose protection and restoration countermeasures based on the optimization simulation results of ecological network resilience. The research results show that: 1) From 2000 to 2020, the area of core patches (potential ecological sources) in Jiangsu Province had a decrease of about 5% of the total ecological landscape area in the corresponding year, with a total reduction of 1250.75 km², a decrease of 11.53%. The overall landscape structure showed a trend of shrinkage and fragmentation; the number of ecological source patches screened through comprehensive evaluation methods such as ecological source screening by the minimum area threshold method and landscape connectivity evaluation showed a basically decreasing trend, and their distribution was significantly correlated with regional topography, geomorphology, and hydrological conditions, with the overall landscape structure showing a trend of shrinkage and fragmentation; the ecological network pattern had the characteristic of “dense in the south and sparse in the north”. During the study period, the number of corridors generally showed a decreasing trend, and the increase in ecological pinch points was relatively large. 2) Affected by urbanization and climate disasters, under random attack scenarios, the resilience of ecological networks in different years showed a gradually slow decline trend. When the node removal rate was between 0% and 10%, the overall network resilience was relatively stable; when the removal rate was between 10% and 95%, the resilience decreased significantly; under targeted attack scenarios, the network resilience showed a significant unbalanced attenuation characteristic when subjected to targeted attacks. When the node removal rate was about 10%, the overall resilience dropped by more than 50%, indicating that important nodes are crucial to maintaining network stability. 3) 39 pinch point areas and 1 barrier area in 2020 were selected as "stepping stones" to optimize the ecological network of Jiangsu Province. The optimized ecological network showed a decrease in the absolute value of the slope, a slowdown in resilience loss, and an increase in the coefficient of determination when facing attacks. When the proportion of attacked nodes was in the range of 20%-50%, the network resilience value tended to be flat, indicating that the optimized ecological network has better resilience and buffering capacity against attacks. Finally, the study further puts forward optimization suggestions from the aspects of classified protection of key ecological sources, optimization of ecological network connectivity, and protection and restoration of important ecological nodes.

Key words: ecological network resilience, disturbance scenario simulation, complex network, circuit theory, Jiangsu Province

摘要：

快速城镇化和气候变化使得生态系统服务功能及生态网络韧性发生着剧烈的变异，如何对不确定的自然与社会系统双重扰动下生态网络韧性进行评估对于区域可持续发展至关重要。以复杂网络理论和电路理论为基础，从结构与功能两个维度构建生态网络韧性评价体系，测度随机攻击与目标攻击情景下2000－2020年江苏省生态网络韧性，基于生态网络韧性优化模拟结果提出保护与修复对策。结果表明：1）研究期间江苏省生态源地数量减少35个，面积占比下降1.15%，破碎化加剧，廊道平均长度增加0.43 km，夹点数量增加271个，生态网络总体呈现“南密北疏”格局；2）随机攻击情景下，生态网络韧性变化趋势均呈线性衰减趋势，当移除率介于10%－95%时，韧性明显下降；目标攻击情景下，当节点移除率约10%时，整体韧性骤降超50%，呈显著的非均衡衰减特征；3）选取39个夹点区域和1个障碍区域作为“踏脚石”进行优化，优化后的生态网络由144个斑块和255条廊道组成，在面对攻击时斜率绝对值降低0.17，当攻击节点的比例在20%－50%范围内时，生态网络韧性值趋于平缓，生态网络韧性损失减缓、决定系数增加。最后提出应依据生态价值与区域功能对生态源地实施“分级管控、分类施策”的保护策略，并通过增设“踏脚石”节点、修复强化生态廊道以优化网络连通性与冗余度，同时对关键节点与障碍点开展针对性保护与修复。

关键词: 生态网络韧性, 扰动情景模拟, 复杂网络, 电路理论, 江苏省