Comparing Selections of Environmental Variables for Ecological Studies: A Focus on Terrain Attributes

Area of Science:

• Ecology
• Geomorphology
• Conservation Science

Background:

• Terrain attributes are vital abiotic surrogates for species distribution and habitat mapping.
• Subjective selection of terrain attributes can result in suboptimal ecological models and maps.
• Effective variable selection is critical for reliable conservation and management decisions.

Purpose of the Study:

• To assess the impact of subjective terrain attribute selection on ecological applications.
• To compare the performance of different terrain attribute combinations in habitat mapping and species distribution modeling.
• To highlight the importance of rigorous variable selection for ecological decision-making.

Main Methods:

• Seven distinct terrain attribute selections were used for mapping benthic habitats.
• Habitat maps were generated using unsupervised classification of biophysical characteristics.
• Species distribution models for sea scallops were created using MaxEnt software.
• Performance metrics were used to quantify and compare the effectiveness of different variable combinations.

Main Results:

• A specific combination of terrain attributes (relative position, slope, orientation, topographic mean, rugosity) outperformed others.
• This optimal combination yielded superior results for both habitat mapping and species distribution modeling.
• Differences in accuracy (up to 47%) and spatial distribution (up to 58%) were observed between variable selections.

Conclusions:

• Subjective terrain attribute selection significantly impacts map accuracy and spatial representation.
• Careful, evidence-based variable selection is essential for reliable ecological applications.
• Inaccurate maps derived from suboptimal variables have critical implications for conservation and management strategies.