Guest blog by Liz Duffel, Georeferencing Digitiser
Most specimens within the Museum collection have locality information, showing where the specimen was found, on the accompanying label(s). When we are digitising our specimens, we can use that locality information for georeferencing – the process used to give the locality of a specimen geographical coordinates, so that it can be plotted on a map.
This is important because it allows for mapping and modelling, which underpins research on anything from species distributions and relationships, to environmental changes or targeting conservation practices.
For example, Steve Brooks and his team compared butterfly collection records – dates and locations collected – to the Central England Temperature record, which has daily temperature readings dating back to 1772. By comparing the geolocation data to the temperature records, they found that between 1880 and 1970 all but one of the 51 butterfly species were affected by yearly changes in climate, and 47 of these species emerged earlier when spring temperatures were higher.
We currently have 893,615 georeferenced specimens on the Museum’s Data Portal. The number that we can process each week depends on many factors, for instance where the specimens were collected – UK locations for example are much easier for us than, say, Belize. Also the amount and type of information added by the curator to the label (e.g. 30ft upstream from the bridge vs just the name of a town.) The amount of problem solving needed per specimen can vary significantly.
Here’s an everyday example:
Brockham Hills sounds like a place that would be easy to find, however a quick search of Streetmap and Geonames show no results. Streetmap shows a result for Brockham Hill Barn in Hampshire and Geonames shows a result for a village called Brockham in Surrey. Immediately, we know further investigation is needed.
Brockham Hill Barn is tempting, as there is also a Brockham Hill Lane and some Brockham Hill Cottages. Streetmap shows us some nice contour lines that indicate a hill. However, the label doesn’t say Brockham Hill Barn. We have to be very careful to geolocate what is on the label, not what we think is on the label. If you get a label for Dover it is very tempting to say, I know that, it’s in Kent, but there is, albeit a tiny one, another Dover in Wigan. So Hampshire looks good, but I need more proof.
I widen the search, and search for ‘Brockham’. I want to know how many are there and are they hilly? This search gives me two, this one (as Geonames picked up) in Surrey and another in Cumbria.
Looking at the Cumbria Brockham, there aren’t really any hills to speak off. However, the Surrey Brockham, turns out is very close to Box Hill, which is a very popular moth collecting site, and it is noticeable that on the hills there is a Brockham Warren.
It is sometimes possible to get clues from the distribution of the moth. I take a look on the National Biodiversity Network’s Gateway and find that the moth, Aspitates gilvaria, otherwise known as the straw belle, is a distinctly southern species.
With the species distribution evidence, we can almost completely rule out Cumbria. My final part of the puzzle would usually be to check the collector and to see if I can find out some information on where they lived and collected. However, on this occasion this is a dead end, as this information isn’t on the label.
As the result of this particular search still has some uncertainty, I label the specimen ‘needs scrutiny’ so that it will be passed on to our Moth Curator, Geoff Martin for a final check. Geoff is an absolute wealth of knowledge on his collections and often can clear up tangles quickly and effectively.
Even with all these methods, and with curators’ input, some records simply do not have enough information to geolocate them and so get assigned to the ‘ungeoreferancable’ pile. Sometimes no amount of research or knowledge can throw up a satisfactory answer.
As we digitise our collections of over 80 million objects, we want to geolocate the specimens whenever possible. We provide open access to data on our Data Portal so that scientists around the world can access the data they need to conduct their research, build models or maps and answer some of society’s biggest questions about extinction rates, climate change and protecting food sources.
If you are in the habit of collecting specimens, please do remember to label their location clearly. This won’t only help us to complete geolocation now, but will ensure that the specimen is accessible for future research.
Visit the Museum’s Data Portal to see what’s already available online.
A really interesting blog showing the steps taken to confirm geo-referencing. If only you had a collectors field book to help! I’m impressed though that even without collector information you were able to narrow the locality down to one, albeit slightly uncertain, locality.
Hi Siobhan, Thank-you for your comment, I’m glad you enjoyed the post. We chose an uncertain location for this post so that we could demonstrate how complex the geolocation process can be. Although we narrowed the locality down, in this instance this goes back to the curator for a further check. The notes about the search will join the location record so that any researcher in the future looking at the geolocation of the specimen will know about the process of determining the location.