12 Conclusion
We have seen the process of ecological niche modeling using R, focusing on viper species in the Iberian Peninsula as a case study. We covered topics such as data collection, preprocessing predictors and presence data, model building, ensembling, and projection to future climate scenarios. The general steps we covered were:
Data Collection: Gather presence data for the species of interest along with environmental variables like bioclimate varibles and NDVI.
Data Preprocessing: Process the data by checking for duplicates in presence record and other erroneous coordinates; making sure the predictors are fully geographically align and that correlations among variables are adequate; prepare a training and projection area.
Model Building: Use the biomod2 package to build ecological niche models, selecting appropriate pseudo-absences, algorithms, and resampling strategies.
Model Evaluation: Evaluate model performance using metrics like True Skill Statistic (TSS) and Receiver Operating Characteristic (ROC) curves.
Model Ensembling: Ensemble multiple models to create a more robust prediction by combining different pseudo-absences sets, algorithms, and resampling strategies.
Projection: Project the ensembled models to future climate scenarios to assess potential distribution changes under different climate conditions.
Spatial Analysis: Analyzs spatial patterns in the model predictions, for instance, identify sympatric zones, and assess changes in contact zones over time.
At this point, you should understand that:
- Ecological niche modeling provides valuable insights into species distributions and responses to environmental changes.
- Model performance depends on various factors such as data quality, variable selection, algorithm choice, and vary between model evaluation metrics.
- Ensembling multiple models improves prediction accuracy and robustness by capturing a wider range of situtations and simplify the analyses of complex model results.
- Future climate projections highlight potential shifts in species distributions, and might provide a good insight for conservation planning and management strategies.
- Spatial analysis is an important process in the ecological modelling an, in this example, was essential to identify areas of overlap and potential interactions between species.