Tree diversity mediates the distribution of longhorn beetles (Coleoptera: Cerambycidae) in a changing tropical landscape (southern Yunnan, SW China)

Abstract

Cerambycidae) have been used to identify sites of high biological diversity and conservation value in cultivated landscapes, but were rarely studied in changing landscapes of humid tropics. This study was conducted in a region of southern Yunnan, China, which was dominated by natural rainforest until 30 years ago, but is successively transformed into commercial rubber monoculture plantations since that time. The objectives were to investigate longhorn beetle species diversity and distribution in the major land use types of this landscape and to estimate the effects of an expected expansion of rubber plantations on the longhorn beetle assemblages. The results showed that tree species diversity (181 species in total) and longhorn beetle diversity (220 species in total) were closely related with no significant differences between the tree and longhorn beetles assemblages shown by similarity distance analysis. There was a highly positive relationship between the estimated species richness of longhorn beetles and the number of tree species. Individual numbers of longhorn beetles and trees were also highly positive related at the sampling sites. Non-metric multidimensional scaling revealed that the degree of canopy coverage, succession age and tree diversity explained 78.5% of the total variation in longhorn beetle assemblage composition. Natural forest sites had significantly higher numbers of species and individuals than any other type of habitat. Although young rubber plantations bear the highest longhorn beetle diversity outside forests (half of the total number of longhorn beetle species recorded in total), they can not provide permanent habitats for most of these species, because they develop into closed canopy plantations with less suitable habitat conditions. Therefore, along with an expected expansion of rubber cultivation which largely proceeds at the expense of forest areas, the habitat conditions for longhorn beetles in this region might decrease dramatically in future.

Figure 1. Quantitative similarity cluster analysis of the tree species at the sampling localities (site codes see Table 1), generated from the Bray-Curtis index using UPGMA through stratigraphically constrained clustering.

The dendrogram shows four subgroups clearly classified by habitat categories.

Figure 2. Quantitative similarity cluster analysis of the longhorn beetles at the sampling localities (site codes see Table 1), generated from the Bray-Curtis index using UPGMA through stratigraphically constrained clustering.

The dendrogram shows three subgroups clearly classified by habitat categories.

Figure 3. Number of (A) species and (B) individuals of longhorn beetles in the four subgroup produced by habitat categories.

Box and whisker plots illustrate the 5th, 25th, 50th (median), 75th, and 95th percentiles, and the means as the dashed line. Different letters indicate significant differences (P<0.05; Mann-Whitney U-Test). Small circle indicate outliners.

Figure 4. NMS ordination of longhorn beetle assemblages at the 13 sample sites: (♦) rubber plantations; (△) grassland and shrubland; (□) rice field fallows; (•) forest.

Blue small circle means longhorn beetle species. Variables with Pearson correlation coefficient at P<0.05 are shown. Cumulative variation in the original dataset explained by ordination is 78.5% (Axis 1 = 41.0%, Axis 2 = 19.2%, Axis 3 = 18.4%, Final stress = 9.58, Final instability = 0.00001).

Figure 5. The polynomial regression between the diversity index of trees and longhorn beetles at 13 sampling location within Naban River Watershed National Nature Reserve.

(A) Number of tree species and estimated Chao1 species richness of longhorn beetles; (B) Number of tree species and collected species richness of longhorn beetles; (C) Number of tree individuals and collected species individuals of longhorn beetles.