pres - Tamu.edu
Transcription
pres - Tamu.edu
Multivariate Ordination Antony Rodger Edwin Lopez Todays Presentation Why use multivariate ordination? What is “ordination”? Distance metrics (Ecological distances) Common transformations Types of ordination Examples of multivariate ordination analyses Why do it?? What is ordination ??? Ordination methods geometrically arrange sites so that distances between them in the graph represent their ecological distances. In an ordination graph, sites are plotted so that distances between them in the graph reflect the ecological differences between them. Ordination Visualize complex data in few dimensions Find patterns and combinations of the variables that can be use in subsequent analysis The goal of ordination is to find axes of the greatest variablility in the community composition (the ordination axes) for a set of samples and to visualize (using an ordination diagram) Large and messy matrices of Variables Multivariate Analysis Distance Metrics (Ecological Distance) • • • • • Euclidean Distance Bray-Curtis Distance Kulczynski Distance Hellinger Distance Chi-Square Distance Euclidean Distance Pricipal Component Analysis Redundancy Discriminant Analysis Bray-Curtis Distance Kulczynski Distance Hellinger Distance Chi-Square Distance Correspondence Analysis Canonical Correspondence Analysis Common Transformations Bray-Curtis, Kulczynski to minimize impact of dominant species Square-root Log-transformation Remove rare species Proportions (species profiles) Presence/Absence Hellinger Chi-square to minimize impact of rare species Euclidean to minimize impact of species abundance Community composition data that usually contains many zero values Ordination • Constrained Direct gradient analysis • • Use information from both the species and the environmental matrices Attempt to explain difference in species composition between sites by differences in environmental variables The aim of constrained ordination is to find the variability in species composition that can be explained by the measured environmental variables Types Unconstrained Indirect gradient analysis Are only based on the species matrix Multivariate Ordination Analyses Unconstrained Constrained Principal Component Analysis (PCA) Redundancy Discriminant Analysis (RDA) Correspondence Analysis (CA) Canonical Correspondence Analysis (CCA) Non-Metric Multidimensional Scaling (NMDS) Distance-based redundancy analysis (dbRDA) Principal Coordinates Analysis (PCOA, MDS) Canonical analysis of principal coordinates (CAP) Discriminant Analysis (DA) Correspondence Analysis (CA) Canonical Correspondence Analysis (CCA) Canonical Correspondence Analysis (CCA) References Kindt, R. and R. Coe. 2005. Tree diversity analysis. A manual and software for common statistical methods for ecological and biodiversity studies. Nairobi: World Agroforestry Centre (ICRAF). http://www.worldagroforestry.org/downloads/publications/PDFs/B13695.pdf Leps, J. and P. Smilauer. 2003. Multivariate analysis of ecological data using CANOCO. Cambridge University Press, New York. http://www.planta.cn/forum/files_planta/multivariate_analysis_of_ecological_dat a_using_canoco_390_173.pdf Additional Resources Vegan Package Vegan: An Introduction to Ordination http://cc.oulu.fi/~jarioksa/opetus/metodi/ Multivariate Statistics Summary http://ordination.okstate.edu/ Community Analysis http://ecology.msu.montana.edu/labdsv/R/labs/ Ordination Methods for Ecologists http://cc.oulu.fi/~jarioksa/opetus/metodi/vegantutor.pdf R Lab for Vegetation Ecologists http://cran.r-project.org/web/packages/vegan/vignettes/intro-vegan.pdf Multivariate Analysis of Ecological Communities in R: vegan tutorial http://cran.r-project.org/web/packages/vegan/vegan.pdf http://www.umass.edu/landeco/teaching/multivariate/schedule/summary.handouts.pdf Ecologically Meaningful Transformations http://adn.biol.umontreal.ca/~numericalecology/Reprints/Legendre_&_Gallagher.pdf