Citizen science needs fancy statistics to detect the impacts of climate change

Background:  Climate change is causing a range of effects in plants and animals. One of the most noticeable is the colonisation of new areas as the environment warms to a point where animals are able to persist where once they could not. However, the sources of data used to detect these kinds of patterns tend not to be systematically collected and so present unique challenges during analysis. In particular, a lot of existing data on sightings of animals that are used to detect trends under climate change originate from enthusiastic amateurs who make a note of which species they see and where.

What we did: I analysed a series of different methods that have been used to control for the effects of recorder effort bias in the detection of range shifts.  This recorder effort bias occurs when there are far more recorders looking for animals in a later period and so the chance of discovering those extreme populations increases. Thus range shifts could simply be an artefact of increased sampling. I demonstrate that the methods that have been used before vary in the detection of range shifts and that some make more sense than others. I follow this up with a case study on range shifts in British Odonata and make recommendations concerning the most appropriate methods.

Importance: Climate change is an important issue and we need cutting-edge analytical tools if we are to properly assess its impacts on the world. I hope that this paper has contributed to this aim.


This is part of a series of short lay summaries that describe the technical publications I have authored.  This paper, entitled “Accounting for recorder effort in the detection of range shifts from historical data”, was published in the journal Methods in Ecology and Evolution in 2010. You can find this paper for free online at the publisher.

Image credit: Ken Slade, CC BY-NC 2.0, http://bit.ly/1qAae4a