Modeling and predicting mean indoor radon concentrations in Austria by generalized additive mixed models

Oliver Alber*, Christian Laubichler, Sebastian Baumann, Valeria Gruber, Sabrina Kuchling, Corina Schleicher

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Radon is a noble gas that occurs naturally as a decay product of uranium. Aside from smoking, radon is considered to be one of the major causes of lung cancer. Indoor environments, where radon can accumulate and potentially reach high concentrations, are of a particular concern. A mixed effects additive model along with a data-driven cross validation model selection method is applied to model the mean indoor radon concentration of dwellings in Austria. For this model a prediction approach is introduced, which enables the mapping of indoor radon potential to identify radon areas in Austria. The data used for modeling was collected in monitoring campaigns for private dwellings in Austria from 2013 to 2019. The proposed method allows policy makers to identify regions with high indoor radon concentrations and enables them to meet regulatory requirements or prioritize radon protection measures. The currently published Austrian radon map and the delineation of radon areas in Austria is based on this proposed method.
Original languageEnglish
Pages (from-to)3435-3449
Number of pages15
JournalStochastic Environmental Research and Risk Assessment
Volume37
Issue number9
Early online date19 May 2023
DOIs
Publication statusPublished - Sept 2023

Keywords

  • Austrian radon map
  • Building factors
  • Generalized additive mixed models
  • Predicted mean indoor radon concentration
  • Stratified step-wise forward k-fold cross validation
  • Transformation of predictions

ASJC Scopus subject areas

  • Water Science and Technology
  • General Environmental Science
  • Safety, Risk, Reliability and Quality
  • Environmental Engineering
  • Environmental Chemistry

Fingerprint

Dive into the research topics of 'Modeling and predicting mean indoor radon concentrations in Austria by generalized additive mixed models'. Together they form a unique fingerprint.

Cite this