PCA and multidimensional visualization techniques united to aid in the bioindication of elements from transplanted Sphagnum palustre moss exposed in the Gdańsk City Area

Abstract

Goal, Scope and Background

During the last decades, a technique for assessing atmospheric deposition of heavy elements was developed based on the principle that samples of moss are able to accumulate elements and airborne particles from rain, melting snow and dry deposition. Despite a broad interest in bioindication there are still ongoing works aimed at the preparation of a standard procedure allowing for a comparison of research carried out in various areas. This is why the comparison of living and dry moss of the same species and growth site seems to be interesting, logical and promising. A most reliable approach seems to be the application of bioindication connected with multivariate statistics and efficient visualization techniques in the interpretation of monitoring data. The aim of this study was: (i) to present cumulative properties of transplanted Sphagnum palustre moss with differentiation into dry and living biomaterial; (ii) to determine and geographically locate types of pollution sources responsible for a structure of the monitoring data set; (iii) to visualize geographical distribution of analytes in the Gdańsk metropolitan area and to identify the high-risk areas which can be targeted for environmental hazards and public health.

Materials and Methods

A six month air pollution study based on Sphagnum palustre bioindication is presented and a simplified procedure of the experiment is given. The study area was located at the mouth of the Vistula River on the Baltic Sea, in Gdańsk City (Poland). Sphagnum palustre was selected for research because of its extraordinary morphological properties and its ease in being raised. The capability of dry and living moss to accumulate elements characteristic for anthropogenic and natural sources was shown by application of Principal Component Analysis. The high-risk areas and pollution profiles are detected and visualized using surface maps based on Kriging algorithm.

Results

The original selection of elements included all those that could be reliably determined by Neutron Activation Analysis in moss samples. Elimination of variables covered the elements whose concentrations in moss were lower than the reported detection limits for INAA for most observations or in cases where particular elements did not show any variation. Eighteen elements: Na, Ca, Sc, Fe, Co, Zn, As, Br, Mo, Sb, Ba, La, Ce, Sm, Yb, Lu, Hf, Th, were selected for the research presented.

Discussion

Two runs of PCA were performed since, in the first-run a heavy polluted location (Stogi — ‘Sto’) understood as outlier in the term of PCA approach was detected and results in the form of block diagrams and surface maps were presented. As ensues from the first-run PCA analysis, the factor layout for both indicators is similar but not identical due to the differences in the elements accumulation mechanism. Three latent factors (‘phosphatic fertilizer plant impact’, ‘urban impact’ and ‘marine impact’) explain over 89% and 82% of the total variance for dry and living moss respectively. In the second-run PCA three latent factors are responsible for the data structure in both moss materials. However, in the case of dry moss analysis these factors explain 85% of the total variance but they are rather hard to interpret. On the other hand living moss shows the same pattern as in first-run PCA. Three latent factors explain over 84% of the total variance in this case. The pollution profiles extracted in PCA of dry moss data differ tremendously between both runs, while no deterioration was found after removal of Stogi from data set in case of living moss. Performance of the second-run PCA with exception of Stogi as a heavy polluted location has led to the conclusion that living moss shows better indication properties than dry one.

Conclusions

While using moss as wet and dry deposition sampler it is not possible to calculate deposition values since the real volume of collected water and dust is hard to estimate due to a splash effect and irregular surface. Therefore, accumulation values seam to be reasonable for moss-based air pollution surveys. Both biomaterials: dry and living Sphagnum palustre show cumulative properties relative to elements under interest. Dry moss has a very loose collection of the atmospheric particles, which can also easily get lost upon rinsing with rainwater running through exposed dry moss material. The living moss may, on the contrary, incorporate the elements in its tissue, thus being less susceptible to rinsing and thus better reflecting the atmospheric conditions. Despite the differences in element uptake and uphold capabilities dry and living moss reflect characteristic anthropogenic and natural profiles.

Visible differences in impacts’ map coverage exist mostly due to the accumulation mechanisms differentiating dry from living moss. However, in case of each indicator ‘phosphatic fertilizer plant impact’ is recognized as the strongest pollution source present in examined region.

Recommendations and Perspectives

General types of pollution sources responsible for a structure of monitoring data set were determined as high-risk/low-risk areas and visualized in form of geographic distribution maps. These locations can be targeted for environmental hazards and public health. Chemometric results in the form of easy defined surface maps can became a powerful instrument in hands of decision-makers working in the field of sustainable development implementation.