Global mean estimation using a self-organizing dual-zoning method for preferential sampling.

Area of Science:

• Environmental Science
• Geostatistics

Background:

• Accurate global mean estimation is critical for environmental monitoring and resource assessment.
• Traditional methods struggle with preferential sampling, where sample points are not uniformly distributed.

Purpose of the Study:

• To develop and validate a novel global mean estimation method for preferential sampling.
• To improve the accuracy and stability of mean estimations compared to existing techniques.

Main Methods:

• A self-organizing dual-zoning method was employed to divide the study area.
• Stratified sampling was applied, considering spatial and feature distribution of sampling points.
• The proposed method was tested using manganese (Mn) concentrations in Jilin Province, China.

Main Results:

• The proposed method demonstrated superior accuracy and stability across varying Feature Deviation Index (FDI) values and sample sizes.
• Relative errors for the proposed method ranged from 0.14% to 1.47%, significantly lower than the direct arithmetic mean method (4.83%-8.84%).
• Existing methods (direct arithmetic mean, polygon, cell) showed sensitivity to FDI and sample size variations.

Conclusions:

• The developed method provides more reliable global mean estimates, especially with preferential sampling.
• This approach offers a robust alternative for geostatistical analysis in environmental studies.
• Consideration of both spatial and feature space improves estimation accuracy.