Data comparison Bioaccumulation In bioaccumulation studies a key
Data comparison - Bioaccumulation -
In bioaccumulation studies, a key issue is the interpretation of pollutant accumulated by biological matrices in terms of a deviation from an “unaltered”, “natural” reference. Several approaches have been suggested to quantitatively assess such deviation: • Comparison with the minimum values measured in the survey area; • Comparison with Background Element Content values (BECs); • Use of “interpretative scales”. All these methods have advantages and drawbacks, and each method shows particular aspects of the phenomenon under study.
N. collection sites in Italy whose samples have that specific concentration value Nimis & Bargagli, 1999 Chromium n=654 ppm Percentile class 20° 75° 50° 95° 90° 98°
The “Mortal Sins” No distinction was made among lichen species used; No methodological control on data sources; Sample pre-processing? Acid attack? No references for bioaccumulation data; The dataset was never published, i. e. nobody could control the data used in this statistical analysis. ? ? ?
The “Mortal Sins” No distinction was made among lichen species used; No methodological control on data sources; Sample pre-processing? Acid attack? No references for bioaccumulation data; The dataset was never published, i. e. nobody could control the data used in this statistical analysis. ? ? ? Objective Develop BRAND NEW, SPECIESSPECIFIC interpretative scales for bioaccumulation data.
Preliminary actions ü Dataset retrieving (Nimis & Bargagli 1999 + data up to 2009) Filling missing methodological information: - Sample acid digestion (TOTAL vs PARTIAL) Finding (and killing) ‘weird’ data… - Typing errors; - Missing units of measurement; - Wrong units of measurement… ü Further data collection Bioaccumulation data after 2009!
Preliminary actions ü Dataset retrieving (Nimis & Bargagli 1999 + data up to 2009) Filling missing methodological information: - Sample acid digestion (TOTAL vs PARTIAL) Finding (and killing) ‘weird’ data… - Typing errors; - Missing units of measurement; - Wrong units of measurement… ü Further data collection Bioaccumulation data after 2009! Revised dataset is ready…
The dataset Record (N) 32186 data (2626 raws) Elements 42 Ag, Al, As, B, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cs, Cu, Fe, Ga, Hg, In, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Pd, Rb, S, Sb, Sc, Se, Sn, Sr, Te, Ti, Tl, U, V, W, Zn, Zr Species 5 E. prunastri, F. caperata, P. sulcata, P. furfuracea, X. parietina Admin. regions 18 % data 4. 7% 39. 5% 55. 8% Molise Calabria Sicilia Basilicata Campania Toscana Marche Umbria Abruzzo Lazio Friuli Venezia Giulia Veneto Liguria Trentino Alto Adige Emilia Romagna Lombardia Nord Italia Piemonte Val d'Aosta 0 100 200 300 400 500 600 700 800 N
The dataset Temporal coverage 1980 -2017 % data 1980 -1985 1986 -1990 1991 -1995 1996 -2000 2001 -2005 2006 -2010 2011 -2015 2016 -2020 0. 9% 12. 6% 5. %5 12. 2% 15. 2% 27. 3% 9. 2% 17. 1% Digestion through years Frequencies of data obtained through partial (P) and total (T) digestion of samples collected in Italy in the last decades.
Criteria for data selection: Enhancing methodological homogeneity - Selecting only element content data deriving from TOTAL acid digestion (HF-based) of samples; - Only data produced in the last decade (2008 - 2018); - N ≥ 3 administrative regions; - N > 40 records. 2 lichen species: Flavoparmelia caperata (Fc) ← 11 elements Xanthoria parietina (Xp) ← 10 elements
Data distributions – species comparison vs Fc Xp Data counts As Cr 500 400 300 200 100 0 1500 1000 500 0 0 5 10 15 0 50 Pb 100 150 400 600 Zn 600 800 600 400 200 400 600 0 200 Element content (µg g-1)
What to do? Bioaccumulation capacity is undoubtedly species-specific (at least for some elements) → Species-specific scales are obviously needed!
What to do? Bioaccumulation capacity is undoubtedly species-specific (at least for some elements) → Species-specific scales are obviously needed! Old scales: “naturality/ALTERATION”… “Alteration” with respect to?
What to do? Bioaccumulation capacity is undoubtedly species-specific (at least for some elements) → Species-specific scales are obviously needed! Old scales: “naturality/ALTERATION”… “Alteration” with respect to? Again… A benchmark is needed. BEC values! However, NO available BECs for Xanthoria parietina and Flavoparmelia caperata…
Finding BECs Remember yesterday?
Finding BECs Remember yesterday? Do not despair! BECs are under our nose.
Finding BECs are inside our Revised Dataset! Nimis & Bargagli, 1999 Class 1: “Absence of alteration”: upper threshold: 20 th %ile 25 th %ile was selected as BEC threshold (instead of 20 th).
Finding BECs are inside our Revised Dataset! Nimis & Bargagli, 1999 Class 1: “Absence of alteration”: upper threshold: 20 th %ile 25 th %ile was selected as BEC threshold (instead of 20 th). From ecotoxicology (and geochemistry)… “[…] the 25 th percentile of the data distribution is chosen as an estimate of the concentration of a chemical in soil at which few significant effects are observed in the soil biota. ” Gaudet CL, Bright D, Adare K, Potter K, (2001). Approach to Deriving Canadian Soil and Sediment Quality Guidelines. In: Species sensitivity distributions in ecotoxicology, Posthuma, L. , Suter II, G. W. , & Traas, T. P. (Eds. ), CRC press.
BEC dataset construction Numerical example: 25 th %ile assessment E. g. n = 20 bioaccumulation data for a generic element in a certain species.
BEC dataset construction Numerical example: Outlier removal q 3 + 3 IQR Outliers q 3 + 1. 5 IQR BEC calculation
Review-based BECs Background values
Review-based BECs Background values … Field-based assessment’s better… * ? ? N > 35 N > 10* Data selection Dataset dimension Revised bioaccumulation dataset 32186 data 42 elements 5 species Total digestion Last decade data 7201 data 39 elements 4 species Criteria: • ≥ 3 regions/element • > 40 data/element 3390 data 11 elements 2 species BEC data • Data ≤ 25 th %ile • OLs removal 931 data 11 elements 2 species
From absolute to dimensionless ‘Bioaccumulation scales’ Ok, we have BECs, but we still need an interpretative tool… Numerical example: E. g. n = 20 bioaccumulation data for a generic element in two lichen species: E. g. - X. parietina Dimensionless!
From absolute to dimensionless ‘Bioaccumulation scales’ X. parietina Numerical example: E. g. n = 20 bioaccumulation data for a generic element in two lichen species. F. caperata
From absolute to dimensionless ‘Bioaccumulation scales’ N observations Testing B ratio distributions for the two species X. parietina F. caperata Bioaccumulation ratio Bioaccumulation is species-specific… HOWEVER also the background is species-specific, so that the ratio between bioaccumulation values and the corresponding background tends to “flatten” species differences.
From absolute to dimensionless ‘Bioaccumulation scales’ By using the B ratio, we can merge data for : - Different elements different species In this way we obtain a ‘B ratio’-dataset, cumulative for the two species. From this dataset we can calculate the %iles (25 th, 50 th, 75 th, and 95 th) and use these as thresholds for BIOACCUMULATION CLASSES. Unique, 5 -class, dimensionless BIOACCUMULATION SCALE that also takes into account the species-specificity.
From absolute to dimensionless ‘Bioaccumulation scales’ Bioaccumulation scale Unique, 5 -class, dimensionless scale that takes into account species-specificity VS Naturality / Alteration scale Multi-specific, 7 -class absolute scale
- Slides: 28