2.1 Rule 1: a site must be sampled during at least 3 winters
The models take both a site and a winter effect into account. When a site is sampled during only one or two winters, then model has a hard time separating the site and winter effect. This can lead to an unstable model. Therefore it is safer to restrict the data to sites with a sufficient number of visits over several winters.
Table 2.1 displays how many sites have data from 1, 2, 3, … different winters. 37 sites have data from less than 3 different winters and are currently ignored.
winters | sites |
---|---|
1 | 13 |
2 | 24 |
3 | 14 |
4 | 10 |
5 | 12 |