Table 2 Akaike information criterion (AIC), estimates of variance components (Genetics, Spatial, Residual), heritabilities (Heritability), and average of the correlation coefficients and root mean square errors (RMSE) from the cross-validation analyses for each of the three spatial mixed models studied (see text for models’ abbreviations). Standard errors are in brackets. The corresponding traits analyzed are total height for sites 2.708.3 and 2.707.1, and diameter for site 3.704.2. The lowest AIC and RMSE values, and highest correlations values are highlighted in italics
From: Performance of alternative spatial models in empirical Douglas-fir and simulated datasets
Trial | Â | Modela | ||
|---|---|---|---|---|
| Â | block | AR | splines | |
2.708.3 | AIC | 104,886 | 104,454 | 104,523 |
Genetic | 2507 (172) | 2484 (182) | 2514 (173) | |
Spatial | 1889 (320) | 1786 (172) | 1649 (341) | |
Residual | 4625 (107) | 4089 (114) | 4359 (111) | |
Heritability | 0.35 (0.02) | 0.38 (0.002) | 0.37 (0.02) | |
Correlation coefficients | 0.54 (0.04) | 0.57 (0.03) | 0.56 (0.03) | |
RMSE | 23.65 (1.09) | 23.1 (1.06) | 23.22 (1.03) | |
2.707.1 | AIC | 64,179 | 63,722 | 63,905 |
Genetic | 1723 (180) | 1882 (224) | 1770 (217) | |
Spatial | 2050 (403) | 4804 (378) | 4693 (890) | |
Residual | 11,353 (262) | 7998 (163) | 10,430 (292) | |
Heritability | 0.13 (0.01) | 0.19 (0.02) | 0.15 (0.02) | |
Correlation coefficients | 0.38 (0.03) | 0.49 (0.03) | 0.44 (0.04) | |
RMSE | 33.02 (1.21) | 31.07 (1.15) | 32.03 (1.24) | |
3.704.2 | AIC | 33,830 | 33,799 | 33,798 |
Genetic | 7833 (971) | 7719 (978) | 7731 (1170) | |
Spatial | 286 (137) | 907 (354) | 578 (268) | |
Residual | 18,155 (799) | 18,060 (839) | 18,070 (824) | |
Heritability | 0.30 (0.03) | 0.30 (0.03) | 0.30 (0.04) | |
Correlation coefficients | 0.24 (0.07) | 0.26 (0.07) | 0.27 (0.07) | |
RMSE | 45.38 (1.91) | 45.07 (2.04) | 45.05 (2.08) | |