Week 9: Species Distribution Modeling¶
Core Questions¶
- What does a species distribution model estimate?
- How do presence/absence, presence/background, and pseudoabsence designs differ?
- Why might logistic regression and random forests produce different suitability maps?
Concepts¶
- Species-environment relationships.
- Presence/absence, presence/background, and pseudoabsence data.
- Logistic regression.
- Random forests.
- Habitat suitability.
- Prediction versus inference.
- Model evaluation and thresholding.
Python Tools¶
- scikit-learn.
- rasterio or rioxarray.
- pandas.
- matplotlib.
Applied Lab¶
Students combine occurrence data, climate, elevation, and land cover covariates to fit logistic regression and random forest models, then compare predicted suitability rasters.
Deliverable¶
Submit a first species distribution model with methods, evaluation metrics, prediction maps, and a comparison of model behavior.