Anny Anselin

Species richness coincidence: conservation strategies based on predictive modelling

The present-day geographic distribution of individual species of five taxonomic groups (plants, dragonflies, butterflies, herpetofauna and breeding birds) is relatively well-known on a small scale (5 × 5 km squares) in Flanders (north Belgium). These data allow identification of areas with a high diversity within each of the species groups. However, differences in mapping intensity and coverage hamper straightforward comparisons of species-rich areas among the taxonomic groups. To overcome this problem, we modelled the species richness of each taxonomic group separately using various environmental characteristics as predictor variables (area of different land use types, biotope diversity, topographic and climatic features). We applied forward stepwise multiple regression to build the models, using a subset of well-surveyed squares. A separate set of equally well-surveyed squares was used to test the predictions of the models. The coincidence of geographic areas with high predicted species richness was remarkably high among the four faunal groups, but much lower between plants and each of the four faunal groups. Thus, the four investigated faunal groups can be used as relatively good indicator taxa for one another in Flanders, at least for their within-group species diversity. A mean predicted species diversity per mapping square was also estimated by averaging the standardised predicted species richness over the five taxonomic groups, to locate the regions that were predicted as being the most species-rich for all five investigated taxonomic groups together. Finally, the applicability of predictive modelling in nature conservation policy both in Flanders and in other regions is discussed.

Predicted insect diversity declines under climate change in an already impoverished region

Being ectotherms, insects are predicted to suffer more severely from climate change than warm-blooded animals. We forecast possible changes in diversity and composition of butterflies, grasshoppers and dragonflies in Belgium under increasingly severe climate change scenarios for the year 2100. Two species distribution modelling techniques (Generalised Linear Models and Generalised Additive Models), were combined via a conservative version of the ensemble forecasting strategy to predict present-day and future species distributions, considering the species as potentially present only if both modelling techniques made such a prediction. All models applied were fair to good, according to the AUC (area under the curve of the receiver operating characteristic plot), sensitivity and specificity model performance measures based on model evaluation data. Butterfly and grasshopper diversity were predicted to decrease significantly in all scenarios and species-rich locations were predicted to move towards higher altitudes. Dragonfly diversity was predicted to decrease significantly in all scenarios, but dragonfly-rich locations were predicted to move upwards only in the less severe scenarios. The largest turnover rates were predicted to occur at higher altitudes for butterflies and grasshoppers, but at intermediate altitudes for dragonflies. Our results highlight the challenge of building conservation strategies under climate change, because the changes in the sites important for different groups will not overlap, increasing the area needed for protection. We advocate that possible conservation and policy measures to mitigate the potentially strong impacts of climate change on insect diversity in Belgium should be much more pro-active and flexible than is the case presently.

Can we predict the distribution of heathland butterflies with heathland bird data

National or regional conservation strategies are usually based on available species distribution maps. However, very few taxonomic groups achieve a full coverage of the focal region. Distribution data of well-mapped taxonomic groups could help predict the distribution of less well-mapped groups and thus fill gaps in distribution maps. Here, we predict the distribution of five heathland butterflies in Flanders (north Belgium) using typical heathland bird distribution data as predictor variables. We compare predictions with those using only biotope or a combination of both biotope and bird data as variables. In addition, we test the transferability of ‘bird’, biotope and combined models to the Netherlands, an ecologically similar region. Transferability was tested in three separate sandy regions in the Netherlands at different distances from the region in which the models were built. For each of the five heathland butterflies, we applied logistic regressions on ten random model sets and tested the models on ten random evaluation sets within Flanders. We used the area under the curve (AUC) of the receiver-operating characteristics (ROC) plots to estimate model accuracy. Overall, bird models performed significantly better than biotope models but were not significantly different from the combined models in Flanders. In the Netherlands, the transferred biotope and the combined models performed better than the transferred ‘bird models’. We conclude that on a local scale, birds can, to some extent, serve as proxies for biotope quality, but that biotope models are more robust when transferred to another region.