Map on predicted deposition of Cs-137 in Spanish soils from geostatistical analyses

Abstract

The knowledge of the distribution of 137Cs deposition over Spanish mainland soils along with the geographical, physical and morphological terrain information enable us to know the 137Cs background content in soil. This could be useful as a tool in a hypothetical situation of an accident involving a radioactive discharge or in soil erosion studies. A Geographic Information System (GIS) would allow the gathering of all the mentioned information. In this work, gamma measurements of 137Cs on 34 Spanish mainland soils, rainfall data taken from 778 weather stations, soil types and geographical and physical terrain information were input into a GIS. Geostatistical techniques were applied to interpolate values of 137Cs activity at unsampled places, obtaining prediction maps of 137Cs deposition. Up to now, geostatistical methods have been used to model spatial continuity of data. Through semivariance and cross-covariance functions the spatial correlation of such data can be studied and described. Ordinary and simple kriging techniques were carried out to map spatial patterns of 137Cs deposition, and ordinary and simple co-kriging were used to improve the prediction map obtained through a second related variable: namely the rainfall. To choose the best prediction map of 137Cs deposition, the spatial dependence of the variable, the correlation coefficient and the prediction errors were evaluated using the different models previously mentioned. The best result for 137Cs deposition map was obtained when applying the co-kriging techniques.