Atik, ŞeymaYetiş, HakanDenizli, HalukEvrendilek, Fatih2021-06-232021-06-2320161018-46191610-2304https://hdl.handle.net/20.500.12491/9016https://www.webofscience.com/wos/woscc/full-record/WOS:000373523600016This study aims to model spatiotemporal variability of indoor radon (Rn-222) concentrations measured for one year from May 2012 to May 2013 in the built environment of Abant Izzet Baysal University. There exist a few studies about data driven modeling of spatiotemporal dynamics of indoor radon and their validation. Mean indoor radon concentration varied spatially between 14 +/- 8.5 Bq/m(3) and 28.5 +/- 17.5 Bq/m(3) and on a monthly basis between 37.3 +/- 21.6 Bq/m(3) in September and 13.1 +/- 7.7 Bq/m3 in April, and on a seasonal basis between 23.4 +/- 18.4 Bq/m3 for the summer period of June to September and 13.3 +/- 7.9 Bq/m(3) for the spring period of April to May. The best-fit multiple non-linear regression (MNLR) model developed in this study elucidated 57.9% (R-adj(2)) of the spatiotemporal variability, with a cross-validation derived predictive power of 57.1% (R-cv(2)). The two-way interactions among the temporal predictors of hour and month, air temperature, relative humidity, and location were most influential in predicting indoor radon levels. Parsimonious versus data hungry empirical non-black-box models appear to be of great practical importance to the quantification, monitoring, and mapping of short and long-term local, regional, or global spatiotemporal dynamics of indoor and outdoor radon concentrations.eninfo:eu-repo/semantics/closedAccessIndoor RadonIndoor Air QualityAir Quality MonitoringEmpirical ModelingSpatiotemporal DynamicsArticle253823829WOS:000373523600016Q4