Edge Effects On Plant Community Structure On Land-bridge Islands In The Thousand Island Lake In East China

The edge effect is one of the major consequences of habitat fragmentation, which affects ecosystem structure and function by changing abiotic and biotic factors along the edge-to-interior gradient. Land-bridge islands formed by dam construction are ideal natural labs for studying habitat fragmentation because of their simultaneous isolation, clear boundaries and homogeneous matrix (water). The Thousand Island Lake, which is located in the Chun'an County of Zhejiang Province, has more than 1000 islands that were formed by the damming of the Xin'an River in 1959. We selected 29 undisturbed or less disturbed islands (20 of them are smaller than lha) as long-term monitoring plots to study plant community dynamics. Islands smaller than lha were sampled in their entirety, while transects were established on islands greater than lha. Sampling area ranged from 0.5ha to lha depending on the total area of each island. Plots were initially divided into 10m×10m quadrats, then further divided into 5mx5m quadrats during the plant census. The woody individuals (DBH≥1cm) were mapped, tagged and identified to species level, and their DBH, height and location in the quadrats were measured. Each island was divided into 5m x 5m quadrats along the edge-to-interior gradient. All 29 islands were divided into five area categories, and the small islands(<1ha) were divided into eight slope aspects in order to compare how plant community structure along the edge-to-interior gradient was affected by area and slope aspect. The results showed that:(1) A total of 186,688 tree and shrub plants were measured on 29 islands, belonging to 36 classes,66 orders, and 87 species. All the species were divided into three functional groups according to the lifehabit, including 34 shade intolerant and 53 shade tolerant species; 32 evergreen and 55 deciduous species; 43 tree and 44 shrub species. Among the 87 species, the genera with the widest distribution and greatest number of species were Loropetalum chinense, Pinus massoniana, and Vaccinium carlesii.(2) The results of the Pearson correlation analysis between the basic characteristics of the community parameters (number of species, plant density, Shannon-Wiener index, average height and average DBH of plants) suggested that other indicators showed a significant correlation (p<0.01) in addition to species numbers and Shannon-Wiener index (p>0.05). The species numbers and Shannon-Wiener index had no obvious trend along the edge-to-interior gradient, while the plant density, average height and average DBH had two thresholds, which were 40-50m and 80-100m. When the distance from the edge was more than 100m, the three parameters tended towards stability. Plant density increased with the distance from the edge reaching the maximum at 50m, then gradually decreased until 100m from the edge. The average height of plants had a sharp increase before the 50m, then followed a slow upward trend before reaching a maximum value at 100m. The average DBH of the plants showed the opposite trend as plant density. The Spearman correlation analysis was conducted between the five community characteristic parameters and the distance gradient and slope aspect. The results showed that species number and distance gradient, slope aspect, Shannon-Wiener index and slope aspect had no significant correlation (P>0.05), while the other three parameters and distance gradient, slope aspect showed a significant correlation (P<0.01).(3) The number of species and Shannon-Wiener index of three functional groups were not affected by edge effects, but at the same distance gradient, the two indices of shade tolerant, evergreen and tree plants were higher than that of shade intolerant, deciduous and shrub plants. The plant density, average height and average DBH of the three functional groups showed strong response to edge effects. The plant density of shade intolerant species showed a gradually decreasing trend from the island edge to the interior, while the shade tolerant species showed the opposite trend from the island edge to 50m away. The average height and average DBH increased with the distance from the edge. The plant density of evergreen and tree plants increased from the edge to 50m away, and then decreased until 100m. The deciduous and shrub plants showed continuous decrease trend from the edge to 100m away.(4) Among of five area categories, the species number, Shannon-Wiener index, plant density, average height and average DBH had no obvious changes from the edge of islands to the interior; at the same distance gradient, the species number and average height increased with island area more than 50ha. When the island area was less than lha, the distance from the edge to the center of the island was no more than 30m, so the five parameters had no significant changes from the edge to the interior because edge effects affected the entire area. When the island area was greater than lha, the five parameters had no significant differences among the five island area categories, which suggested that different island areas had no influence on the edge effect.(5) The number of species, plant density, Shannon-Wiener index, average height and average DBH showed no obvious trend from the edge of islands to the interior between eight distinct slope aspects on all small islands, giving additional evidence that the entire area of small islands are effectively edge habitats.