Using Kriging incorporated with wind direction to investigate ground-level PM_2.5 concentration

Area of Science:

• Environmental Science
• Atmospheric Science
• Geostatistics

Background:

• Traditional interpolation methods for particulate matter (PM) data often neglect wind direction.
• Accurate spatial distribution of PM2.5 is crucial for environmental and health assessments.

Purpose of the Study:

• To develop and validate a novel interpolation algorithm, Wind-direction Ordinary Kriging (Win-OK), that incorporates wind direction.
• To compare the performance of Win-OK against Ordinary Kriging (OK) for PM2.5 spatial distribution estimation.

Main Methods:

• Developed the Win-OK algorithm, which prioritizes upwind measurements for predictions.
• Applied Win-OK and OK to analyze hourly PM2.5 concentrations in Xinxiang city, China.
• Utilized leave-one-out cross-validation to calculate Root-Mean-Square Error (RMSE) and standard deviation for performance evaluation.

Main Results:

• Win-OK demonstrated more stable and accurate predictions of PM2.5 spatial distribution compared to OK.
• OK exhibited high RMSE values and significant deviations from measured data with a Gaussian semi-variance model.
• Win-OK, particularly with a spherical semi-variance model, proved to be the most accurate method.

Conclusions:

• The Win-OK algorithm significantly improves the prediction accuracy of ground-level PM2.5 spatial distribution.
• Incorporating wind direction into interpolation is essential for more reliable PM2.5 monitoring and analysis.
• Win-OK offers a more robust approach for understanding air quality dynamics.