Absence of evidence from citizen science does not equal evidence of absence
Citizen science is a great tool for early detection of pests, diseases, invasive species and other rare events. One reason it is so useful is because anyone can get involved, and so act as eyes and ears for detection right across the country, for example as highlighted by the UK’s Tree Health and Plant Biosecurity Expert Taskforce.
This week we, at CEH, published a paper (with colleagues from Butterfly Conservation in the UK and Naturalis Biodiversity Center in the Netherlands) that looks at the importance of citizen science for early detection, using the oak processionary moth as a case study. Oak processionary moth has been accidentally introduced to parts of London and is spreading. This is important because its caterpillars have hairs which can cause rashes and breathing problems (see the Forest Research website for advice) and in extreme cases there can be so many caterpillars they can cause defoliation of oak trees.
Left: The oak processionary moth caterpillars can cause rashes and sometimes breathing problems, so follow advice from the Forestry Commission if they are found. (Photo: James Kitson, University of Hull). Right: The adult oak processionary moth which has been recorded at light traps, especially by naturalists in the Netherlands where this moth is well established. (Photo: Les Hill, Butterfly Conservation).
The conclusion of our paper is that many citizen science projects (including those looking for pests, diseases and invasive species) receive ‘presence only’ records. This means we cannot distinguish between where people are looking but not recording the item of interest, and where people are just not looking.
Therefore, in order to make best use of citizen science we need to:
- respond rapidly where the species is recorded with ‘presence only’ citizen science records
- be cautious about making conclusions based on where we have not received records (and think carefully on how citizen science projects can be improved to help provide this information).
Detect and respond as early as possible
If we could detect new outbreaks of pests, diseases and invasive species as early as possible, then we are better placed to undertake appropriate action. This is where citizen science can play a part.
Volunteers in citizen science have been described as a “standing army” and a resource “ready to act as the need arises”. Records from volunteers can be acted upon and enable agency staff or conservation groups to respond rapidly. Indeed, sightings from the general public have been credited with helping to stop Colorado beetle becoming established in the UK, stopping avian flu becoming established in wildfowl populations in Europe, and enabling rapid responses to Oriental chestnut gall wasp and Asian longhorn beetle in the UK.
If a sighting is reported, then it is vital that it is confirmed and acted upon quickly and we recommend that systems should be put in place to speed up this flow of data (examples of this include Recording Invasive Species Counts and Observatree).
But there is a problem: absence of records does not equal absence of presence
Encouraging people to submit sightings is superb, but there is a big problem with these so-called ‘presence only’ records: what about the places where nothing has been reported? Does this mean that the species is definitely not present? Of course not!
What about the places where nothing has been reported? Does this mean that the species is definitely not present? Of course not!
The trouble is that it is often impossible to distinguish between:
- an absence of records because of the absence of recorders or recording
- an absence of records because of the absence of the species.
(The lack of a record could also be because people are looking but did not detect it, even though it is present, as recently written about by our colleagues.)
In our paper we considered whether naturalists recording moths could help in the early detection of the oak processionary moth because they often send in records of all the moths they see.
- We took data from the 4.5 million records in the Noctua database submitted by moth recorders in the Netherlands (where oak processionary moth is already established) and calculated the chance of any recorder finding oak processionary moth.
- We then applied this result to the pattern of recording moths in the UK, based on data from the 20 million records in the UK National Moth Recording Scheme database, to work out what chance moth recorders had of recording oak processionary moth if it was established in their area.
- We found that oak processionary moth usually had a chance of being detected by moth recorders (if it was present), but often that chance was small.
Naturalists recording moths at a light trap. Photo: Michael Pocock
An important conclusion was that these datasets were special because we could actually calculate the chance of a recorder detecting the species we were interested in. With this result we can say how confident we are that no detections = not present. If we wanted to, we could then target extra recording effort to under-recorded, high risk places. Of course, people other than moth recorders have found and reported the oak processionary moth in the past, but without more information it is practically impossible to determine their coverage (and so the likelihood that they would find the oak processionary moth if it was present).
Citizen science is a fantastic way of obtaining information about rare events (such as invasive species and diseases) but we need to think carefully about the best use of citizen science.
Full paper reference: Pocock, M J O, Roy, H E, Fox, R, Ellis, W N & Botham, M. (2016) Citizen science and invasive alien species: Predicting the detection of the oak processionary moth Thaumetopoea processionea by moth recorders. Biological Conservation. doi:10.1016/j.biocon.2016.04.010