Table 4.
Mean niche similarity metrics compared to true locality models
| Trial | I | D | RR |
|---|---|---|---|
| BAMONA (n = 230) | |||
| County centroid | 0.9829 | 0.8860 | 0.9188 |
| County average | 0.9870 | 0.8961 | 0.9247 |
| 50/50 True locality/county‐centroid mix | 0.9915 | 0.9209 | 0.9439 |
| 50/50 True locality/county‐average mix | 0.9906 | 0.9176 | 0.9349 |
| OdonataCentral (n = 283) | |||
| County centroid | 0.9837 | 0.8917 | 0.9210 |
| County average | 0.9802 | 0.8810 | 0.9132 |
| 50/50 True locality/county‐centroid mix | 0.9925 | 0.9273 | 0.9475 |
| 50/50 True locality/county‐average mix | 0.9924 | 0.9264 | 0.9459 |
The similarity statistic I is a metric used to test if models generated from different populations are identical. Schoener's D is a metric describing the level of niche overlap. Relative rank, RR, is an estimate of the probability that a pair of rasters agree in the relative ranking of any two patches of habitat regardless of the suitability values. See text for interpretation.