The canonical correlation coupling and optimization for the structural characteristics of the combined sediment-sea water-plants community system in the Mangrove Forests,Guangxi Province
To effectively observe the complex sediment-sea water-plants community system in the Mangrove Forests,and to accurately assess its integrated eco-environmental quality,a fitted optimal model is needed.The study focused on coupling and optimizing the canonical correlation on the model of the complex sediment-sea water-plants community.Firstly,indices system was proposed by compiling indicators for quantifying ecological characteristics of mangrove plants community,indicators for assessing sea water quality and sediment environmental quality,which was then applied in two Mangrove Nature Reserve,named Shankou and Beilunhekou of Guangxi Province.Then,principal component analysis(PCA)was adopted to optimize structural data sets of canonical correlation,and canonical correlation analysis(CCA)was adopted to explore the multiple relationships among plant community,sea water,and sediment.Finally,the order of indicators was optimized and the ideal CCA model of plant community-sea water-sediment was established by using an integrated method of CCA standardized canonical coefficients,frequency count,and expert knowledge.Results showed that:(1)The ideal fitted 24 indicators with symmetry were reduced from the initial 49 indicators with asymmetry after coupling and optimizing process.Canonical correlations among plant community,sea water and sediment were also transformed from nothing to something.The optimization canonical correlations were showed as the following orders,sediment-sea water>plant community-sediment>plant community-sea water.Meanwhile,sea water and sediment had the highest canonical correlation coefficients(their first and second canonical correlation with its significant test were as follows:c1=0.994,p=0.0001;c2=0.993,p=0.0005,respectively),plant community had very strong canonical correlation with sediment(c1=0.997,p= 0.001;c2=0.0.984,p=0.008,respectively),and effective canonical correlation was also found between plant community and sea water(c1= 0.987,p=0.042;c2=0.902,p=0.423,respectively).(2)As the important life organism between sea water and sediment,plant community's biggest important value,biggest branching number,and biggest DBH were the very useful fitted indicators for sea water and sediment data sets.Meanwhile,the biggest DBH was more important than the total spices number.(3)Both the mean matrix of multiple CCA and increasing samples could improve the significant level on the plant community-sea water-sediment model,whereas the explored matrix for the combined method of the special judgement could both improve the canonical correlation values and their significant level.Integrated method of PCA-CCA-SK could optimize the attributes of indictors very well,and couple the correlation of complex ecosystem effectively,as well as be fitted for the comprehensive eco-environmental quality assessment currently.
canonical correlation analysisdata setsmangrove forestsGuangxi Province
国家科技期刊平台
NSTL
万方数据
