Table 3.
Detailed results of dbRDA significance tests for the global dbRDA analysis (after removing non-significant terms) and the marginal test, considering 32 source populations of C. mauritiana
| dbRDA | Partial dbRDA | |||||
|---|---|---|---|---|---|---|
| % of variance explained | d.f. | P-value | % of variance explained | d.f. | P-value | |
| Global analysis | 85.03 | 7 | <0.001 | |||
| Residuals | 14.97 | 24 | – | – | – | |
| Marginal test | ||||||
| Geography | 47.56 | 21.37 | ||||
| geo1 | 22.30 | 1 | <0.001 | 10.54 | 1 | <0.001 |
| geo2 | 15.24 | 1 | <0.001 | 7.01 | 1 | <0.001 |
| geo3 | 6.21 | 1 | 0.006 | 2.51 | 1 | 0.002 |
| geo4 | 3.81 | 1 | 0.048 | 1.31 | 1 | 0.013 |
| Environment | 23.66 | 7.20a | ||||
| CDD | 9.75 | 1 | <0.001 | 4.27 | 1 | <0.001 |
| Tmean | 6.06 | 1 | 0.008 | 2.43 | 1 | 0.014 |
| PET | 7.85 | 1 | 0.002 | 3.32 | 1 | 0.004 |
| Residual | 28.76 | 24 | – | 28.76 | 24 | – |
CDD mean number of consecutive dry days, Tmean mean average daily temperature, PET average monthly potential evapotranspiration.
In the dbRDA analysis, the marginal effects of each constraining factor were tested through 10,000 permutation tests, by removing each term one by one from the model containing all other terms. In partial dbRDA, the pure contribution of each factor was obtained after removing the confounding effects from the other constraining factors. geo1, geo2, geo3, geo4 are uncorrelated geographic factors
aDue to remnant correlations between environmental variables, the estimated contribution of environmental factors in partial dbRDA (7.20%) does not correspond to the sum of individual contributions of each environmental factor (10.02%).