Spatial And Temporal Trends Of Organic Pollutants In Vegetation From Remote And Rural Areas

Abstract

Persistent organic pollutants (POPs) and polycyclic aromatic hydrocarbons (PAHs) used in agricultural, industrial, and domestic applications are widely distributed and bioaccumulate in food webs, causing adverse effects to the biosphere. A review of published data for 1977-2015 for a wide range of vegetation around the globe indicates an extensive load of pollutants in vegetation. On a global perspective, the accumulation of POPs and PAHs in vegetation depends on the industrialization history across continents and distance to emission sources, beyond organism type and climatic variables. International regulations initially reduced the concentrations of POPs in vegetation in rural areas, but concentrations of HCB, HCHs, and DDTs at remote sites did not decrease or even increased over time, pointing to a remobilization of POPs from source areas to remote sites. The concentrations of compounds currently in use, PBDEs and PAHs, are still increasing in vegetation. Differential congener specific accumulation is mostly determined by continent-in accordance to the different regulations of HCHs, PCBs and PBDEs in different countries-and by plant type (PAHs). These results support a concerning general accumulation of toxic pollutants in most ecosystems of the globe that for some compounds is still far from being mitigated in the near future.

Figure 1. Continental distribution of POPs in vegetation around the globe.

The map shows the proportion of each family of compounds reviewed for all sampling locations with published data. The boxplots compare the concentrations of each family of POPs across continents. Interquartile ranges (25^th and 75^th percentile) are shown by the height of the boxes, and the horizontal lines represent medians (50^th percentile). Whiskers range from the 10^th to 90^th percentiles. Concentrations are in ng·g⁻¹ dry weight. The map was generated using the R packages rgdal, maptools, and mapplots. Permission is granted to Nature Publishing Group to publish the image in all formats under an Open Access license.

Figure 2. Variability of POPs and PAHs in vegetation from remote, rural, urban, and industrial environments.

Box plots of the logarithm of the concentration of each family of compounds. Small letters refer to significant differences among land uses (Tukey’s HSD, P < 0.05). Different letters indicate significant differences. Interquartile ranges (25^th and 75^th percentile) are shown by the height of the boxes, and horizontal lines represent medians (50^th percentile). Whiskers range from the 10^th to 90^th percentiles, and values outside this range are indicated by small squares. Concentrations are in ng·g⁻¹ dry weight.

Figure 3. Multilevel model of the selected effects of abiotic and biotic variables on the concentrations of POPs in plants.

Constant and varying effects by congeners are included for the predictor variables included in the lowest-AIC models. All models accounted for spatial correlation structure and nugget (which were significant in all cases, P < 0.001) and were run using the restricted maximum likelihood criterion for optimizing parameter estimates. MAT, mean annual temperature; MAP, mean annual precipitation.

Figure 4. Individual relationships among HCB, ΣHCHs, and ΣDDTs (ng·g⁻¹ dry weight) and year.

The equations for rural and remote areas are: (A) HCB rural: y = −0.04× + 91, r² = 0.19, df = 1354, P < 0.0001, and HCB remote: y = 0.07× − 140, r² = 0.43, df = 1,95, P < 0.0001; (D) ΣHCHs rural: y = −0.07× + 148, r² = 0.31, df = 1,356, P < 0.0001, and ΣHCHs remote: y = 0.08× − 16, r² = 0.01, df = 1,143, P = 0.2; (G) ΣDDTs rural: y = −0.01× + 23, r² = 0.01, df = 1,356, P = 0.09, and ΣDDTs remote: y = 0.01× − 13, r² = 0.00, df = 1,155, P = 0.4; and (J) ΣPCBs rural: y = −0.02× + 50, r² = 0.11, df = 1,102, P < 0.001, and ΣPCBs remote: y = −0.03× + 71, r² = 0.24, df = 1,124, P < 0.0001. Panels (B), (E), (H) and (K) and (C), (F), (I) and (L) show the estimated curve for kernel density for both rural and remote regions, respectively, before and after 2000.

Figure 5. Individual relationships among ΣPBDEs and ΣPAHs (ng·g⁻¹ dry weight) and year.

The equations for rural and remote areas are: (A) ΣPBDEs rural: y = 0.05× − 98, r² = 0.48, df = 1,91, P < 0.0001, and ΣPBDEs remote: y = 0.03× − 64, r² = 0.05, df = 1,18, P = 0.3; and (B) ΣPAHs rural: y = 0.06× − 125, r² = 0.32, df = 1,192, P < 0.0001, and ΣPAHs remote: y = −0.01× + 28, r² = 0.02, df = 1,47, P = 0.3.

Figure 6. Coefficients from the multilevel model giving the significant congener-specific responses to continent and plant type*.

The values are the species-specific changes of slope plus the estimates for fixed effects. *Lichens as plant-like composite organisms.