Changing Landscape Land Cover Accuracy Assessment
Summary
| YEAR | OVERALL ACCURACY | OVERALL ACCURACY | KAPPA |
| Primary Reference | Primary &Secondary Reference | Primary Reference | |
| 1985 | 86.45% | 91.82% | 0.8307 |
| 1990 | 85.00% | 91.09% | 0.8213 |
| 1995 | 83.55% | 90.27% | 0.8059 |
| 2002 | 82.64% | 89.18% | 0.7969 |
| 2006 | 81.91% | 88.18% | 0.7895 |
| 2010 | 82.18% | 88.91% | 0.7919 |
| 2015 | 81.82% | 88.55% | 0.7878 |
Overall Accuracy Primary Reference - The overall percent accuracy based on just the Primary Reference point. This point identifies the perceived land cover, based on reference imagery, at the center of the Landsat pixel. This is the typical accuracy reporting using ERDAS IMAGINE.
Overall Accuracy Primary & Secondary Reference - The overall percent accuracy based on combining the Primary and, if that is incorrect, the Secondary Reference points. The Secondary Reference is the next likely land cover category based on the land cover features in the immediate surrounding area.
Kappa - Expresses the proportionate reduction in error generated by a classification process compared with the error of a completely random classification. It is generally thought to be a more robust accuracy measure than simple overall accuracy calculation, since κ takes into account the agreement occurring by chance.