I noticed that some site_id entries share the exact latitude and longitude information. One example is DAMO and DOBE.
Since the latitude and longitude have up to four decimal points, the theoretical precision should be around 10 meters (as far as I know). The nest counts between each respective site_id pair mostly vary drastically (which seems not to be explained by the error estimate).
What is the expected precision of the latitude and longitude information? Is each site_id a truly unique location?
Hi @nik.las, here’s some more background on the site location from the experts:
Some of the sites are so close together that they have the same lat/long to within the precision of the locations. To the best of our biological understanding of the species, these are reproductively isolated populations even though they are only maybe 100 m apart. DOBE for example is a small island off the coast of DAMO, which is a site on the edge of a slightly bigger island. We believe DOBE was colonized by penguins from DAMO but have no way of knowing for sure, though at this point we assume all populations (following colonization and any immigration immediately following) are reproductively isolated.
These locations are only just now getting sorted and some of them may not be correct to the meter since the Antarctic is still largely unknown, many areas are poorly mapped, and its difficult to get good GPS readings at high latitudes sometimes (fewer satellites available often). That said, I’m quite confident that having spent several years cleaning up the locations that the locations are well within the resolution of the covariate data we might look at. In other words, the location might be off by 20m but the predictor data resolution is at a 25 km grid size.