Development of a High-Resolution Indoor Radon Map Using a New Machine Learning-Based Probabilistic Model and German...

Area of Science:

• Environmental Health
• Geospatial Analysis
• Risk Assessment

Background:

• Radon is a carcinogenic, radioactive gas that accumulates indoors, posing health risks.
• Accurate indoor radon concentration data is vital for public health and identifying high-risk areas.
• Current national-scale estimations often lack spatial resolution and may not represent target populations accurately.

Purpose of the Study:

• To develop a model-based approach for more realistic indoor radon concentration estimation.
• To achieve higher spatial resolution in indoor radon distribution mapping than traditional methods.
• To improve the accuracy of indoor radon assessment, even with non-representative survey data.

Main Methods:

• A multistage modeling approach using quantile regression forest with environmental and building data.
• Estimation of the probability distribution function of indoor radon for each floor level.
• Application of Monte Carlo sampling for population-weighted, floor-level prediction combination.

Main Results:

• Indoor radon in German dwellings follows a lognormal distribution.
• Significant portions of the population are exposed to elevated radon levels (12.5% exceed 100 Bq/m³, 2.2% exceed 200 Bq/m³).
• Lower indoor radon concentrations are generally found in large cities compared to rural areas due to floor-level population distribution.

Conclusions:

• The proposed model provides accurate indoor radon concentration estimates with high spatial resolution.
• This approach effectively accounts for variations in floor level and soil radon concentration.
• The method enhances the understanding of indoor radon exposure variability and population impact.