[Application of geostatistical analysis in study of macrophytes community's pattern]

In this paper, the Log-log semivariogram and theoretical models of Geostatistics and fractal theory were applied to study the mosaic spatial pattern of macrophytes community and its different characters on different scales in littoral zone of the Baoan Lake, Hubei Province. The macrophytes in the littoral zone with a square area of 25,600 m2, which was divided in to 1,024 grids, were investigated, and the macrophytes species in every grid were recorded. If the species was discovered in the grid, the variable of the grid had a value of 1, otherwise, it was 0. The community had a value of the total species discovered in the grid. By the method, the sampling results of Vallisneria spiralis, Ceratophyllum oryzetorum, Myriophyllum spicatum and the community were made into four worksheets. With help of the software GS+ for Windows to treat these worksheets, we got their log-log semivariogram, isotropic model, anisotropic model, isotropic fractal dimension and anisotropic fractal dimension. According to the results of isotropic models, the exponential models of Vallisneria spiralis and Myriophyllum spicatum indicated their contagious distribution with a high proportion of spatial structure, which indicated that their distribution changed lesser at microscale. In their range of 48.6 m and 34.2 m respectively, they had a spatial autocorrelation. Isotropic models of Ceratophyllum oryzetorum and the community were linear, and their proportions of spatial structure were low, indicating their great change at microscale and their moderate contagious distribution. The isotropic fractal dimension and anisotropic fractal dimension results indicated the dominant pattern of Vallisneria spiralis in the direction of NE-SW, the dominant patterns of Myriophyllum spicatum and Ceratophyllum oryzetorum in the direction of N-S, and the community's dominant pattern in the direction of NW-SE. Compared with other analysis tools, the method had advantages in deciding whether the community components had a spatial autocorrelation and in quantitatively classifying the aggregation density. It provides us a new way to quantitatively describe the pattern of the macrophytes community.