Ecological Archives E088-115-A5

Wim Vyverman, Elie Verleyen, Koen Sabbe, Koenraad Vanhoutte, Mieke Sterken, Dominic A. Hodgson, David G. Mann, Steve Juggins, Bart Van de Vijver, Vivienne Jones, Roger Flower, Donna Roberts, Victor A. Chepurnov, Cathy Kilroy, Pieter Vanormelingen, and Aaike De Wever. 2007. Historical processes constrain patterns in global diatom diversity. Ecology 88:1924–1931.

Appendix E. Results of the variation partitioning analyses.

   FIG. E1. The unique contribution and overlap between the three classes of environmental variables in explaining mean regional richness (A) and local richness (B and C) showing the overriding effect of historical factors. Variation partitioning analysis was based upon 31 grids (A), 1039 lakes (B) for which the environmental data pH, conductivity, and the concentration of Ca, Mg, K and Na were available, and a subset from B of 681 lakes (C) for which the morphological data lake depth and surface area were additionally available. The variables that were significant during the forward selection procedure and included in the analyses are given in Table E1. Note that none of the temperature related variables significantly explained variations in mRR (A). In a more complete data set for which historical and temperature related variables were available (53 grids), the first (isolation) explains 68.6% (61.3% uniquely) and the latter (the yearly range between warmest and coldest month) accounts for 7.4% (0.1% uniquely and not significant) of the variance. In a more restricted data set for which lake depth and surface area were additionally available (23 grids) only isolation significantly explained (71.5%) of the variance in mRR; none of the limnological (including morphological variables) and temperature related predictors were significant.


TABLE E1. Variables included in the variation partitioning analyses and the amount of variation (P ≤ 0.05) they explained in the model using the forward selection procedure in CANOCO 4.5. "(Q)" denote the quadratic terms of the variables.


Limnological and

Temperature related

Local richness

number of lakes (30.5%)

Ca (2.3%)

(Q)mean temperature (2.2%)

(1039 lakes)

(Q)isolation (1.5%)

Na (1.8%)

mean temperature (1.5%)

(Q)number of lakes (0.9%)

pH (1.5%)

% freshwater (0.8%)

(Q)Ca (0.3%)

(Q)% freshwater (0.5%)

(Q)lake connectivity (0.3%)

lake connectivity (0.2%)

Local richness

(Q)isolation (38.7%)

(Q)pH (2%)

(Q)mean temperature (2.4%)

(681 lakes)

number of lakes (4%)

Ca (1.5%)

yearly range (0.6%)

isolation (1.1%)

(Q)Ca (1.1%)

(Q)yearly range (0.5%)

(Q)lake area (0.8%)

lake area (0.7%)

Na (0.4%)

(Q)max. lake depth (0.3%)


isolation (72%)

variance in Na (4.9%)

(31 grids)

variance in pH (3.7%)

[Back to E088-115]