道路建设对西南地区景观格局的影响

为揭示西南地区道路建设对景观格局的影响，基于2015年土地覆被和道路数据，计算了研究区道路核密度与景观指数，在此基础上分析了道路建设的直接、间接和综合影响，探讨了影响的空间异质性和对不同土地覆被影响的特点，指出了各景观指数的适用性。结果表明：(1)道路建设使斑块密度增加，斑块离散化加剧，发生明显的破碎化趋势，同时道路核密度与PD、SHDI、SPLIT变化百分比呈显著正相关，与COHESION、CONTAG、PAFRAC呈显著负相关；(2)道路建设对景观指数的影响程度由大到小依次表现在：破碎化、斑块空间分布形态、聚集连通性，其中PD能很好的表征道路建设对景观破碎化的作用程度，SPLIT在识别破碎化程度高的区域更具优势；(3)不同景观指数具有不同适用性，PAFRAC适用于道路不发达区域，CONTAG适用于道路分布异质性较高的区域，DIVISION适用于人类活动强度不大的区域，COHESION、SPLIT适用于探究道路及伴随的人类活动共同作用的区域，PD、SHDI在探究道路的直接作用时适用性广泛；(4)探究不同土地覆被类型受道路建设的影响，林地优先选取PD、DIVISION;草地优先选取DI...

Effects of road construction on landscape pattern in Southwest China

To reveal the impact of road construction on landscape patterns in Southwest China, the kernel density and landscape index of roads were calculated based on land cover and road data in 2015. On this basis, the direct, indirect and comprehensive impacts of road construction were analysed, the spatial heterogeneity and characteristics of the impacts on different land cover types were discussed, and the applicability of each landscape index was pointed out. The results showed that (1) the road construction increased patch density and intensified patch discretization, leading to an obvious fragmentation trend. The road kernel density index (KD) was positively correlated with the patch density index (PD), Shannon''s diversity index (SHDI), and splitting index (SPLIT) and was negatively correlated with the patch cohesion index (COHESION), contagion index (CONTAG), landscape division index (DIVISION), and perimeter-area fractal dimension index (PAFRAC). (2) The impact of road construction on the landscape index was mostly due to fragmentation, followed by the patch spatial distribution form and aggregation connectivity. The PD could well represent the degree of road construction effect on landscape fragmentation. SPLIT was more advantageous for identifying areas with a high degree of fragmentation. (3) The different landscape indices had different applications. PAFRAC was suitable for the underdeveloped regions of roads. CONTAG was suitable for regions with a highly heterogeneous road distribution. DIVISION was suitable for regions with a low intensity of human activities. COHESION and SPLIT were suitable for exploring regions where roads and accompanying human activities interacted together, and PD and SHDI were widely applicable for exploring roads directly. (4) For exploring the effects of road construction on different land cover types, the PD and DIVISION indices were preferable for forests; DIVISION and SPLIT for grasslands; PD, PAFRAC, COHESION, SPLIT and DIVISION for farmlands; PAFRAC, COHESION and SPLIT for wetlands; PD, PAFRAC, and COHESION for construction lands; and DIVISION, SPLIT and COHESION for other land cover types. Exploring the effects of road construction on landscape patterns is of great practical significance for understanding landscape pattern changes and guiding environmental protection in Southwest China.