Estimating potential illegal land development in conservation areas based on a presence-only model

Area of Science:

• Environmental Science
• Conservation Biology
• Geographic Information Systems

Background:

• Conservation areas face escalating threats from human land use.
• Accurate prediction of illegal land development is vital for effective conservation.
• Traditional methods are limited by selection bias and unreliable absence data.

Purpose of the Study:

• To introduce a novel presence-only maximum entropy (MAXENT) model for estimating potential illegal land development.
• To address the limitations of traditional methods in recognizing illegal land use.
• To support decision-making in conservation area management.

Main Methods:

• Application of the presence-only maximum entropy (MAXENT) model.
• Utilizing MAXENT's principle to avoid introducing inaccurate pseudo-absence samples.
• Validation of model robustness using high AUC scores (over 0.80).

Main Results:

• The MAXENT model demonstrated superior performance compared to the presence-absence random forest model.
• Topographic conditions and proximity to transportation networks were key drivers of illegal land development.
• A probability map identified significant forest, farmland, grassland, and water bodies at high risk.

Conclusions:

• The proposed MAXENT method effectively estimates potential illegal land development in conservation areas.
• Findings highlight the importance of considering topography and transportation networks in land use planning.
• The study provides a valuable tool for governmental bodies to prioritize conservation efforts in vulnerable regions.