Uranium distribution in the Variscan Basement of Northeastern Sardinia

We present a detailed map of the uranium distribution and its uncertainties in the Variscan Basement of Northeastern Sardinia (VBNS) at a scale 1:100,000. An area of 2100 km2 was investigated by means of 535 data points obtained from laboratory and in situ gamma-ray spectrometry measurements. These data volume corresponds to the highest sampling density of the European Variscides, aimed at studying the genetic processes of the upper crust potentially triggered by an enrichment of radiogenic heat-producing elements. For the first time the Kriging with Variance of Measurement Error method was used to assign weights to the input data which are based on the degree of confidence associated to the measurements obtained with different gamma-ray spectrometry techniques. A detailed tuning of the model parameters for the adopted Experimental Semi-Variogram led to identify a maximum distance of spatial variability coherent to the observed tendency of the experimental data. We demonstrate that the obtained uranium distribution in the VBNS, characterized by several calc-alkaline plutons emplaced within migmatitic massifs and amphibolite-facies metamorphic rocks, is an excellent benchmark for the study of ‘hot’ collisional chains. The uranium map of VBNS, and in particular the Arzachena minor pluton, confirms the emplacement model based on the recognition of the different petrological associations characterizing the Variscan magmatic processes in the Late Paleozoic. Furthermore, the presented model of the uranium content of the geological bedrock is a potential baseline for future mapping of radon-prone areas.

💡 Research Summary

The paper presents a high‑resolution (1:100 000) map of uranium (U) distribution and its associated uncertainties for the Variscan Basement of Northeastern Sardinia (VBNS), covering an area of 2 100 km². A total of 535 measurement points were collected using two complementary gamma‑ray spectrometry techniques: laboratory analyses of collected rock samples and in‑situ measurements with portable spectrometers. Because the two techniques have markedly different error structures—laboratory data exhibit an average standard error of about ±0.15 ppm while field data show ±0.35 ppm—the authors adopted a geostatistical approach that explicitly incorporates measurement error variance: Kriging with Variance of Measurement Error (KVME). In KVME each observation is weighted by the inverse of its error variance, allowing the more reliable laboratory values to exert greater influence on the interpolated surface while still preserving the spatial information contained in the field data.

Prior to interpolation, an experimental semi‑variogram (ESV) was constructed to characterize spatial autocorrelation. The ESV displayed a rapid rise up to a lag of roughly 5 km and reached a sill at about 15 km, indicating that uranium concentrations are spatially correlated within a 15 km radius. This range aligns with the geological architecture of the study area, where calc‑alkaline plutons, migmatitic massifs, and amphibolite‑facies metamorphic rocks form distinct structural blocks. Model fitting showed that a spherical variogram model provided the lowest root‑mean‑square error (RMSE) compared with exponential or Gaussian alternatives, reflecting a relatively uniform spatial variability across the dataset.

The resulting uranium map reveals pronounced lithological controls. The Arzachena minor pluton exhibits the highest concentrations, ranging from 3.5 to 5.2 ppm, whereas surrounding migmatitic and metamorphic units display lower values between 1.8 and 3.0 ppm. This contrast supports a genetic model in which Late Paleozoic magmatic processes concentrated radiogenic heat‑producing elements within calc‑alkaline magmas, producing a “hot collisional chain” typical of Variscan orogenic belts. The map also highlights secondary enrichment zones along pluton margins, suggesting post‑emplacement hydrothermal alteration or melt‑rock interaction as additional mechanisms for uranium redistribution.

Beyond petrological implications, the authors emphasize the map’s utility for environmental health assessments. Uranium decay produces radon‑222, a leading indoor air pollutant; consequently, regions with elevated uranium are potential radon‑prone zones. The high sampling density (≈1 point per 4 km²) and the quantified uncertainty surface provide a robust baseline for radon risk mapping, allowing authorities to prioritize monitoring and mitigation efforts, especially around the high‑U Arzachena pluton.

In summary, the study achieves three major advances: (1) it delivers the densest uranium dataset ever compiled for a European Variscan segment, (2) it demonstrates the practical advantage of KVME in integrating heterogeneous geophysical measurements, and (3) it links the spatial pattern of uranium to both the tectono‑magmatic evolution of the Variscan basement and to contemporary radon hazard assessment. The methodology and the resulting dataset constitute a valuable reference for future investigations of crustal heat production, geothermal potential, and radiological risk in similar orogenic settings worldwide.