For this week’s assignment, I used two museum data aggregators, GBIF and iDigBio, to find existing museum data for coho salmon, Oncorhynchus kisutch. I chose this species because I know that there will be several occurrences and therefore it will should be relatively straight forward to compare the two aggregators’ collection of data. I also did a good bit of work with O. kisutch when I was working on a research project in Washington.
GBIF returns 75,428 occurrences for O. kisutch while iDigBio returns 1,861. I think the main explanation for the difference in records is that GBIF aggregates biological specimen data as well as observations and checklists whereas iDigBio appears to only aggregate biological specimen data. At least, this is what I could gather from their website. I will say that GBIF has a graph that graphs occurrences per basis of record which shows that 2.7% of their records were based on preserved specimen which would indicate that around 2,036 records are specimen. This is pretty close to 1,861, so unsurprising.
Both sites have similar dates of collection ranges. On iDigBio there were 2 records pre-1880 and on GBIF there were 6. On interesting feature on GBIF is that you can see a timeline of occurrences per year (Figure 1). From this graph, we can see that the majority of occurrences are occurring around the 1980s with a large spike around 2005. I think this is probably due to the advancement of the Information Age.
Both aggregators show a pretty close approximation of the species range. O. kisutch can be found in coastal waters from Alaska to the Pacific Northwest to Monterrey Bay, California (Crawford and Muir 2008). They can also be found in Japan and Russia at similar latitudes (Crawford and Muir 2008). They’ve also been introduced in a number of places around the world especially in the lower 48 and especially in the Great Lakes. Both maps sort of represent this well. I would be interested to find out the relative abundance of coho in Japan/Russia compared to the U.S. Both aggregators show a great deal more records in the U.S. compared to Japan/Russia. I wonder if this actually represents the population sizes in these areas or perhaps an aggregator bias simply because there are more records in the U.S.
This is sort of not super relevant, but I did find one interesting set of records. In the iDigBio information, there appeared to be one recordset that occurred in the North Sea near the UK. This is not extremely surprising as it could be possible that there are incidents where some farmed fish escape a net pen and then get collected (or some similar situation). However, when I looked at the locality data, it clearly says that it was collected from the Sashin Creek in Sitka, Alaska. Upon closer inspection the Lat/Long is listed as 56.352/0 which would indeed place it in the North Sea. However, the correct Lat/ Long is likely 56.352/-134.705. There were three records with this (mis?)-information.
One interesting feature in iDigBio is that you can view the recordsets. For example, it shows that there are 190 results in the UAM Fish Collection in Arctos. Just for fun, I thought I would check this. I went to Arctos and entered the species name and put UAM:Fish in the GUID Prefix field. This indeed returned 190 records.
General conclusions: GBIF has a great deal more records because they include different types of records such as observations. iDigBio has fewer records, but if you are only interested in physical preserved specimen, it might be the way to go. Both have different tools and features so depending on which is more relevant to your needs, each one of them has unique benefits. I’ve used both for various projects and sometimes I use a combination of a few different databases depending on my specific needs for that project.
Q: Do you have much experience using aggregators or is this your first time exploring them?
Crawford, S. S., and A. M. Muir. 2008. Global introductions of salmon and trout in the genus Oncorhynchus: 1870-2007. Reviews in Fish Biology and Fisheries 18:313–344.