Vegetationndash;environment relationships are defined with the aid of principal-components analysis and canonical correlation analysis. In both the uplands and lowlands a moisture gradient, determined by measuring gravimetric moisture and indicated by organic carbon, is the most important environmental influence on the vegetation. In the uplands this gradient is also associated with snow depth (drifting) and in the lowlands with conductivity. The second environmental gradient in the uplands is associated with depth to permafrost and its soil textural correlates. Thus soil texture, independent of its effect on soil moisture status, influences the distribution of plant communities. In the lowlands the second environmental gradient is less clear but is also associated with depth to permafrost and, in addition, elevation and CaCO3equivalent. Canonical correlation analysis shows that the components extracted by principal-components analysis of the vegetation data did not conform to the important trends of variation in the environmental data. Principal-components analysis is nevertheless an essential means of data reduction prior to the application of canonical correlation. The statistical model used in the study has potential advantages over the independent use of either principal-components analysis or canonical correlation.
展开▼
