| Marine benthos are characterized by narrow activity and special territory distribution and can be used to indicate the quality of water and the polluted condition of sediments, which plays an important part in the biologic indication of marine pollution. Because of the adaptability of organism and the diversity and complexity of ecosystem, benthic researchers have cited some mathematical formulae and some statistic analysis graphs in order to show the perplexing changes of community structure resulted from pollution. But when we deal ecological data gotten with deferent methods, we may get very deferent results, which plays a direct part on how to explain the practical ecology meaning and how to assess ocean environmental quality. Multivariate analytical techniques can deal with multiattribute ecological data very effetely. Based on benthic community species abundance and four kinds of environmental factors data in Spring, 2000, the author analyzed and compared some important multivariate analytical techniques at present, and then designed a set of technique procedure of multivariate analysis which is fit for dealing with ecological data. Finally the theoretic research results were used to explore the reasons that have brought the spatial and temporal variations of intertidal benthic community structure in Futian, Shenzhen Bay.In general, the main contents of this research can be summarized as the following.Firstly, Mi)S ordination methods can reflect the continual gradient of changes between different community structure. Even if it doesn't anatomies very well, its integral effect is better than that of hierarchical cluster analysis. Results of hierarchical cluster analysis can be taken as reference of MDS ordination. When we used those two multivariate statistic analytical techniques to study the spatio-temporal variation laws of benthic community structure at Shenzhen Bay, the results indicated that: 1) There was not obverse difference in intertidal benthic community structure between Futian, Shenzhen and MaPi, HongKong in Shenzhen Bay in Spring, 2000. However, there was continual gradient change in community structure from outer bay to inner bay. 2) There was obvious deference with the respect of benthic community structure between station C and stations Al, A2, A3 in transect A and station B in transect CBA in Shenzhen Bay. There being no benthos at station C in Summary, 2000, it indicated that station C had been polluted very heavily. 3) Deferent seasons and deferent tide sites are the most two important factors which play influence on benthic community structure at stations Al, A2, A3 and B3 in Shenzhen Bay.Secondly, if there is multicollinearity within some environmental variables when we apply multivariate analytical methods to explore the relationship between environmental factors and community structure, this multicollinearity problem should be eliminated in advance. The method is to replace those multicollinear variables withtheir linear combination. The general formula is:~ ~ Xfl>~IxThirdly, CCA, which taken the dimensional variables with the character ofIIcontinuum generated from MDS ordination as the new biological variables, can indicates the relationship between environmental factors and community structure more objectively than discriminant analysis, which based on the grouping results of hierarchical cluster analysis. The results of multivariate analysis showed that there was no obvious relationship between benthic commumty structure and environmental factors, which included heavy metals (Pb~ Cd~ Zn~ Cu and Cr), organic matter, salinity and sediment particle sizes(fine-sand and clay) in Shenzhen Bay.Fourthly, software named BCD has been designed in this paper to computed Bray-Curtis dissimilarities. The results of BCD software can be written to the data windows of other general statistic analysis software from their data connector.Fifthly. the most important contribution of this paper is that a set of technical procedure of multivariate analysis has bee...
|
PDF Full Text Request