Table 3 Predictive ability of the simulation results. The number that the model considered as “correct” and showed the lowest AIC among 4 models was shown. In total, 100 runs were conducted for each simulation, but some of the runs were not considered as “correct”, thus, the total number was not always 100. The best performance for each month is highlighted by bold italic text.
Month | kNDVI | Landscape | AND | OR |
|---|---|---|---|---|
May | 69 | 0 | 21 | 10 |
June | 1 | 0 | 4 | 0 |
July | 55 | 0 | 13 | 32 |
August | 37 | 0 | 10 | 53 |
September | 0 | 77 | 0 | 23 |
October | 1 | 2 | 3 | 94 |
November | 2 | 0 | 3 | 0 |