Table 5.
Details of the best standard spatial (BSS) models in each trial for grain yield (GY) and plant height (PH)
| Trial | BSS model for GYa | BSS model for PHa |
|---|---|---|
| BIL05 | Lin(r) + AR1xAR1 + n | Lin(r) + AR1xAR1(c) + n |
| DAB05 | R + AR1xAR1 + n | Lin(c) + AR1xAR1 + n |
| DYS05 | R + Spl(r) + AR1xAR1 + n | Lin(r) + AR1xAR1 + n |
| HER05 | Lin(r) + Spl(c) + AR1xAR1 + n | Spl(r) + Lin(c) + AR1xAR1 + n |
| JIM05 | C + AR1xAR1 + n | R + AR1xAR1 + n |
| BIL06 | Spl(c) + AR1xAR1 + n | C + AR1xAR1 |
| CEP06 | R + AR1xAR1 + n | Spl(c) + AR1xAR1 + n |
| DAB06 | C + Spl(r) + AR1xAR1 + n | AR1xAR1 + n |
| GON06 | Spl(c) + AR1 + n | R + C + AR1 |
| HER06 | R + Spl(r) + Lin(c) | R + C + Lin(c) + AR1xAR1 + n |
| BIL07 | Lin(c) + AR1xAR1 + n | R + Spl(c) |
| CLE07 | R + Lin(r) + AR1xAR1 + n | Lin(c) + AR1xAR1 + n |
| DYS07 | Spl(c) + AR1xAR1 + n | Lin(c) + AR1xAR1 + n |
| HER07 | R + Spl(r) + Spl(c) + AR1xAR1 + n | Lin(r) + AR1xAR1 + n |
| BIL08 | R + C + Spl(c) | C |
| DAB08 | R + AR1xAR1 + n | Lin(c) + AR1 + n |
| DAL08 | R + C + Lin(c) | C |
| HER08 | – | AR1 |
| KIL08 | R + Lin(r) + AR1xAR1 + n | AR1xAR1 + n |
| LIV08 | R + C + AR1xAR1 + n | AR1xAR1 + n |
| SPR08 | Spl(c) + AR1xAR1 + n | – |
aSpl(·): cubic smoothing spline indexed by row (r) or column (c) positions; Lin(·): linear regression on row (r) or column (c) positions. R: random row effects; C: random column effects; AR1 and AR1xAR1: correlated residuals modelled as one- and two-dimensional first-order autoregressive process, respectively; n: spatially independent residual (nugget effect). Note that all the models included a fixed block effect