Macrobenthos habitat potential mapping using GIS-based artificial neural network models

Area of Science:

• Ecology
• Marine Biology
• Geographic Information Systems (GIS)

Background:

• Tidal flats are crucial ecosystems supporting diverse macrobenthos communities.
• Accurate habitat mapping is essential for effective conservation and management of these dynamic environments.
• Traditional methods for habitat assessment can be labor-intensive and time-consuming.

Purpose of the Study:

• To develop and validate a method for generating macrobenthos habitat potential maps.
• To utilize artificial neural networks (ANN) combined with GIS and remote sensing for habitat prediction.
• To assess habitat potential for specific macrobenthos species in the Hwangdo tidal flat, Korea.

Main Methods:

• Field collection of macrobenthos samples.
• Compilation of eight environmental control factors into a spatial database using remote sensing and GIS analysis.
• Application of an artificial neural network model to generate habitat potential maps.
• Validation of maps against surveyed species locations.

Main Results:

• Habitat potential maps were generated for Macrophthalmus dilatatus, Cerithideopsilla cingulata, and Armandia lanceolata.
• A strong correlation was observed between the predicted potential maps and actual species distribution.
• Validation achieved average accuracies of 74.9% for M. dilatatus, 78.32% for C. cingulata, and 73.27% for A. lanceolata.

Conclusions:

• The integrated approach of GIS, remote sensing, and ANN is effective for mapping macrobenthos habitat potential in tidal flats.
• The developed method provides a reliable tool for ecological assessment and conservation planning in similar coastal environments.
• Accurate habitat potential mapping supports informed decision-making for tidal flat ecosystem management.