Abstract:

Based on data obtained from field investigations, this paper aims to analyze the inter-specific relationships of desert riparian forest plant communities, assess the influence of different quadrat sizes on inter-specific associations, discuss what is an appropriate quadrat size for best results, and present the relationships among populations in desert riparian forests in the middle and lower reaches of the Tarim River using Fisher's exact test. The results showed that 16 species noted in 21 sampling sites formed the principal community of the desert riparian forest. The vertical structure of the forest consisted of a tree layer, shrub layer and herb layer. It was found that 4 sampling sites had 3 layers, 10 had 2 layers and 7 had 1 layer. There were 120 species pairs formed by 16 species in 210 quadrats among which 17 species pairs showed significant and very significant inter-specific associations, accounting for 14.2% of total species pairs. Thirteen species pairs were positively correlated and four species pairs were negatively correlated. The size of the quadrat influenced the inter-specific associations among species pairs, but the most suitable quadrat size varied among species and sampling sites. A quadrat from 8 m×8 m to 30 m×30 m may be the most suitable size for analysis of inter-specific associations of desert riparian forest communities in Tarim River basin. Different species pairs had different minimum quadrat sizes. There was spatial variance in the interaction between different populations of the desert riparian forest. Different patterns and relationships of inter-specific associations were found as the size of the quadrat increased. Generally, four types of changes in species associations were observed with changing quadrat size. For the first type, the correlative significance of species pairs changed from high to low and then to high as the quadrat size changed from the minimum size of 2 m×2 m to the maximum size of 32 m×64 m. These species pairs accounted for 17.6% of the total species pairs. In the second type of variation, the correlative significance of species pairs changed from low to high and then to low with an increase in quadrat size. These species pairs accounted for 41.2% of the total. For the third type, the significant and very significant inter-specific associations occurred in quadrats that ranged in size from 8 m×16 m to 32 m×64 m. The correlative significance of species pairs changed from low to high, and if the quadrat size continued to increase, new relationships would emerge. These species pairs accounted for 29.4% of the total. For the forth type, the significant and very significant inter-specific associations changed from low to high and then from low to high again as the quadrat size continued to increase. These species pairs accounted for 11.8% of the total. Species associations among shrub-herb and herb-herb pairs were more numerous in 17 inter-specific associations, accounting for 76.5% of the total. Species associations between Populus euphratica and arbor and shrub-shrub were distributed independently of each other. Even if inter-specific associations among species-pairs appeared, they were usually negative correlations. Most species-pairs, which were between arbor and herb-shrub and herb-herb, typically had positive correlations.

Key words: Tarim River, Desert riparian forest community, Inter-specific relationship, Fisher'apos, s exact test for contingency table