Mapping the geogenic radon potential for Germany by machine learning

Petermann, Eric; Meyer, Hanna; Nussbaum, Madlene; Bossew, Peter

doi:10.1016/j.scitotenv.2020.142291

Cited by 46 publications

(24 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By evaluating all RF performances (highest R 2 , lowest RMSE), we achieved the best results with the GWL and SM at the LP site, and the lowest variation in the predictions were found at the UP site. As a potential reason, lower performance might be caused by the test data, including, incomplete spatial coverage of the mapping area [65] and, more likely, incomplete coverage of the UAV maps. UAV mapping of an entire study area cannot be maintained constantly during field campaigns due to wind, battery, or camera failures, particularly for the Tetracam camera, which is sensitive to operational conditions.…”

Section: Model Performancementioning

confidence: 99%

Mapping the Groundwater Level and Soil Moisture of a Montane Peat Bog Using UAV Monitoring and Machine Learning

et al. 2021

View full text Add to dashboard Cite

One of the best preconditions for the sufficient monitoring of peat bog ecosystems is the collection, processing, and analysis of unique spatial data to understand peat bog dynamics. Over two seasons, we sampled groundwater level (GWL) and soil moisture (SM) ground truth data at two diverse locations at the Rokytka Peat bog within the Sumava Mountains, Czechia. These data served as reference data and were modeled with a suite of potential variables derived from digital surface models (DSMs) and RGB, multispectral, and thermal orthoimages reflecting topomorphometry, vegetation, and surface temperature information generated from drone mapping. We used 34 predictors to feed the random forest (RF) algorithm. The predictor selection, hyperparameter tuning, and performance assessment were performed with the target-oriented leave-location-out (LLO) spatial cross-validation (CV) strategy combined with forward feature selection (FFS) to avoid overfitting and to predict on unknown locations. The spatial CV performance statistics showed low (R2 = 0.12) to high (R2 = 0.78) model predictions. The predictor importance was used for model interpretation, where temperature had strong impact on GWL and SM, and we found significant contributions of other predictors, such as Normalized Difference Vegetation Index (NDVI), Normalized Difference Index (NDI), Enhanced Red-Green-Blue Vegetation Index (ERGBVE), Shape Index (SHP), Green Leaf Index (GLI), Brightness Index (BI), Coloration Index (CI), Redness Index (RI), Primary Colours Hue Index (HI), Overall Hue Index (HUE), SAGA Wetness Index (TWI), Plan Curvature (PlnCurv), Topographic Position Index (TPI), and Vector Ruggedness Measure (VRM). Additionally, we estimated the area of applicability (AOA) by presenting maps where the prediction model yielded high-quality results and where predictions were highly uncertain because machine learning (ML) models make predictions far beyond sampling locations without sampling data with no knowledge about these environments. The AOA method is well suited and unique for planning and decision-making about the best sampling strategy, most notably with limited data.

show abstract

Section: Model Performancementioning

confidence: 99%

Mapping the Groundwater Level and Soil Moisture of a Montane Peat Bog Using UAV Monitoring and Machine Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The construction of the GRP of an area is based on the analysis of the spatial distribution of some proxy geological information (e.g., lithological types, U, Th and Ra content, the Rn emanation coefficient from rocks, soil/rock permeability, faults, etc.) that can be related to in situ radon production and migration processes [ 10 , 11 , 17 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 ]. The resulting GPR map defines the spatial distribution of radon risk from sub-surface sources; information that can then be used for land-use zoning and strategic indoor radon monitoring purposes.…”

Section: Methodsmentioning

confidence: 99%

Radon Hazard in Central Italy: Comparison among Areas with Different Geogenic Radon Potential

Giustini

Ruggiero

Sciarra

et al. 2022

IJERPH

View full text Add to dashboard Cite

Radon (222Rn) is a natural radioactive gas formed in rocks and soil by the decay of its parent nuclide (238-Uranium). The rate at which radon migrates to the surface, be it along faults or directly emanated from shallow soil, represents the Geogenic Radon Potential (GRP) of an area. Considering that the GRP is often linked to indoor radon risk levels, we have conducted multi-disciplinary research to: (i) define local GRPs and investigate their relationship with associated indoor Rn levels; (ii) evaluate inhaled radiation dosages and the associated risk to the inhabitants; and (iii) define radon priority areas (RPAs) as required by the Directive 2013/59/Euratom. In the framework of the EU-funded LIFE-Respire project, a large amount of data (radionuclide content, soil gas samples, terrestrial gamma, indoor radon) was collected from three municipalities located in different volcanic districts of the Lazio region (central Italy) that are characterised by low to high GRP. Results highlight the positive correlation between the radionuclide content of the outcropping rocks, the soil Rn concentrations and the presence of high indoor Rn values in areas with medium to high GRP. Data confirm that the Cimini–Vicani area has inhalation dosages that are higher than the reference value of 10 mSv/y.

show abstract

“…The Wismut mining company (SAG/SDAG Wismut) produced a total of 220,000 tons of uranium during its operation period from 1946 to 1990, making it the third largest uranium producer worldwide [2]. The underlying radon potential (dimensionless variable) 1 m underground is indicated by the color code in the map [3]. The area of the uranium mines has naturally high radon concentrations in the soil and air.…”

Section: Introductionmentioning

confidence: 99%

“…The study is conducted at the Federal Office for Radiation Protection (BfS) and is still ongoing [4]. Mortality data are regularly evaluated every 5 years The underlying radon potential (dimensionless variable) 1 m underground is indicated by the color code in the map [3]. The area of the uranium mines has naturally high radon concentrations in the soil and air.…”

Section: Introductionmentioning

confidence: 99%

The German Uranium Miners’ Biobank—A Biobank for OMICs Radiation Research

et al. 2022

View full text Add to dashboard Cite

Systematic bio- and databanks are key prerequisites for modern radiation research to investigate radiation response mechanisms in the context of genetic, environmental and lifestyle-associated factors. This report presents the current status of the German Uranium Miners’ Biobank. In 2008, the bio- and databank was established at the Federal Office for Radiation Protection, and the sampling of biological materials from former uranium miners with and without lung cancer was initiated. For this purpose, various biological specimens, such as DNA and RNA, were isolated from blood samples as well as from formalin-fixed paraffin-embedded lung tissue. High-quality biomaterials suitable for OMICs research and the associated data on occupational radiation and dust exposure, and medical and lifestyle data from over 1000 individuals have been stored so far. Various experimental data, e.g., genome-wide SNPs, whole genome transcriptomic and miRNA data, as well as individual chromosomal aberration data from subgroups of biobank samples, are already available upon request for in-depth research on radiation-induced long-term effects, individual radiation susceptibility to lung cancer and radon-induced fingerprints in lung cancer. This biobank is the first systematic uranium miners´ biobank worldwide that is suitable for OMICs research on radiation-exposed workers. It offers the opportunity to link radiation-induced perturbations of biological pathways or processes and putative adverse outcome(s) by OMICs profiling at different biological organization levels.

show abstract

Mapping the geogenic radon potential for Germany by machine learning

Cited by 46 publications

References 58 publications

Mapping the Groundwater Level and Soil Moisture of a Montane Peat Bog Using UAV Monitoring and Machine Learning

Mapping the Groundwater Level and Soil Moisture of a Montane Peat Bog Using UAV Monitoring and Machine Learning

Radon Hazard in Central Italy: Comparison among Areas with Different Geogenic Radon Potential

The German Uranium Miners’ Biobank—A Biobank for OMICs Radiation Research

Contact Info

Product

Resources

About