Table 2.
Validation statistics of univariate regression models for AGB estimation (significance at 0.05 level (2-tailed)) by canopy roughness at the two locations (the best model per experimental test site is highlighted in boldface).
| Site | Model | R 2 | RMSE (g/m2) | RRMSE (%) | ASE (%) | MPSE (%) | |
|---|---|---|---|---|---|---|---|
| 1 | Linear | y = 3187.2x − 2938.3 | 0.506 | 26.273 | 16.74 | 3.19 | 14.73 |
| Power | y = 274.22x20.061 | 0.488 | 26.468 | 16.87 | 1.58 | 14.37 | |
| Exponential | y = 3E − 07E20.622x | 0.487 | 26.853 | 17.11 | -0.26 | 14.15 | |
| Polynomial | y = 2764x2 − 2189x − 324.27 | 0.506 | 26.272 | 16.74 | 3.15 | 14.72 | |
| Logarithmic | y = 3099.3ln(x) + 247.71 | 0.506 | 26.274 | 16.74 | 3.13 | 14.72 | |
|
| |||||||
| 2 | Linear | y = −1705.5x + 1751.2 | 0.502 | 16.706 | 13.28 | 2.08 | 11.86 |
| Power | y = 60.331x−14.87 | 0.531 | 16.736 | 13.31 | 1.04 | 11.84 | |
| Exponential | y = 3E + 08e−15.5x | 0.531 | 19.129 | 15.21 | -5.65 | 12.62 | |
| Polynomial | y = 7886.9x2 − 16844x + 9114.6 | 0.503 | 16.691 | 13.27 | 2.12 | 11.90 | |
| Logarithmic | y = −1637ln(x) + 46.94 | 0.502 | 16.703 | 13.28 | 2.05 | 11.86 | |
R 2: coefficient of determination; RMSE: root mean square error; RRMSE: relative RMSE; ASE: average systematic error; MPSE: mean percent standard error.