The Value of Social Media in Developing Ecological Niche Models for Australian Native Bees

Australia has over 2,000 species of native bees but only a handful are described in detail. What is often missing from their descriptions is the environmental niche in which they live and the extent of their geographic range. While there are established online repositories often used for environmental niche models (ENMs), the value of social network sites (SNSs) has been frequently overlooked in developing accurate models. This study compared environmental niche models using traditional data and a combination of social media data and traditional data. With the aim to evaluate how social media data effects the predicted environmental niche. Using the Maxent algorithm, I modelled the potential niche of nine species of native Australian bees. This study compared two models for each species, one consisting of only Atlas of Living Australia (ALA) data (traditional dataset) and the other a combination of ALA and geo-tagged data from Flickr photos (social media dataset). The two models for each species determined if the inclusion of Flickr data changed the predicted niche range and the environmental predictors (climatic variables downloaded from Worldclim) deemed as important in predicting that range. I show that Flickr data can add significant value in developing ENMs for species which have low occurrence data in online repositories. All species modelled experienced a change in their niche range between the two models, with six of the nine species modeled exhibiting a significant interaction between environmental predictor and model type in predictor importance. I also discovered that the geographic location of the additional occurrence points was more important in determining changes in niche range than the number of occurrence points. My results show that geo-tagged Flickr images can enhance our understanding of what drives environmental niche ranges in native bees and has implications for many other species. The inclusion of social media data can aid in discovering new species distribution ranges, which may then lead to improved model accuracy and thus improved conservation efforts across different taxa, particularly those with limited distribution information.