Approach to spatialize local to long-range atmospheric metal input (Cd, Cu, Hg, Pb) in epiphytic lichens over a meso-scale area (Pyrénées-Atlantiques, southwestern France)

International audienceGeographically based investigations into atmospheric bio-monitoring usually provide information on concentration or occurrence data and spatial trends of specific contaminants over a specified study area. In this work, an original approach based on geographic information system (GIS) was used to establish metal contents (Hg, Cu, Pb, and Cd) in epiphytic lichens from 90 locations as atmospheric bio-monitors over a meso-scale area (Pyrénées-Atlantiques, southwestern France). This approach allows the integration of the heterogeneity of the territory and optimization of the sampling sites based on both socioeconomical and geophysical parameters (hereafter defined as urban, industrial, agricultural, and forested areas). The sampling strategy was first evaluated in several sites (n = 15) over different seasons and years in order to follow the temporal variability of the atmospheric metal input in lichens. The results demonstrate that concentration ranges remain constant over different sampling periods in “rural” areas (agricultural and forested). Higher variability is observed in the “anthropized” urban and industrial areas in relation to local atmospheric inputs. In this context, metal concentrations in lichens over the whole study show that (1) Hg and Cd are homogeneous over the whole territory (0.14 ± 0.04 and 0.38 ± 0.26 mg/kg, respectively), whereas (2) Cu and Pb are more concentrated in “anthropized” areas (9.3 and 11.9 mg/kg, respectively) than in “rural” ones (6.8 and 6.0 mg/kg, respectively) (Kruskall-Wallis, K(Cu) = 13.7 and K(Pb) = 9.7, p < 0.00001). They also showed a significant local enrichment for all metals in many locations in the Pays Basque (West) mainly due to metal and steel industrial activities. This confirms the local contribution of this contamination source over a wider geographic scale. A multiple linear regression model was applied to give an integrated spatialization of the data. This showed significant relationships for Pb and Cu (adjusted r2 of 0.39 and 0.45, respectively), especially with regards to variables such as industry and road densities (source factors) and elevation or water balance (remote factors). These results show that an integrated GIS-based sampling strategy can improve biomonitoring data distribution and allows better differentiation of local and long-range contamination