Species distribution modelling
Whilst all the individual SDM techniques yielded good results (AUC
> 0.8), RF had the best performance (AUC = 0.98) compared
to SRE which had the poorest performance (AUC= 0.84) (Table 5). All
models had TSS scores of above 0.4 and thus were all included in the
final Ensemble model.
The final Ensemble model had good predictive ability as indicated by TSS
scores of 0.80 and an AUC value of 0.95. Analysis into the importance of
the environmental variables showed that the driest quarter was the most
important environmental variable followed by mean temperature of the
coldest quarter and maximum temperature of the warmest month. Adjusted
annual precipitation was the least important variable (Table 6).