Soil-Geochemical Factors controlling the Distribution and OralBioaccessibility of Nickel, Vanadium and Chromium in NorthernIreland SoilsPalmer, S., Cox, S., McKinley, J., & Ofterdinger, U. (2014). Soil-Geochemical Factors controlling the Distributionand Oral Bioaccessibility of Nickel, Vanadium and Chromium in Northern Ireland Soils. Applied Geochemistry,51, 255–267. https://doi.org/10.1016/j.apgeochem.2014.10.010

Published in:Applied Geochemistry

Document Version:Peer reviewed version

Queen's University Belfast - Research Portal:Link to publication record in Queen's University Belfast Research Portal

Publisher rightsThis is the author’s version of a work that was accepted for publication in Applied Geochemistry. Changes resulting from the publishingprocess, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in thisdocument. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequentlypublished in Applied Geochemistry, Vol 51 (Dec 2014) 10.1016/j.apgeochem.2014.10.010

General rightsCopyright for the publications made accessible via the Queen's University Belfast Research Portal is retained by the author(s) and / or othercopyright owners and it is a condition of accessing these publications that users recognise and abide by the legal requirements associatedwith these rights.

Take down policyThe Research Portal is Queen's institutional repository that provides access to Queen's research output. Every effort has been made toensure that content in the Research Portal does not infringe any person's rights, or applicable UK laws. If you discover content in theResearch Portal that you believe breaches copyright or violates any law, please contact openaccess@qub.ac.uk.

Download date:19. May. 2020

Accepted Manuscript

Soil-geochemical factors controlling the distribution and oral bioaccessibility

of nickel, vanadium and chromium in soil

Sherry Palmer, Siobhan F. Cox, Jennifer M. McKinley, Ulrich Ofterdinger

PII: S0883-2927(14)00242-X

DOI: http://dx.doi.org/10.1016/j.apgeochem.2014.10.010

Reference: AG 3344

To appear in: Applied Geochemistry

Please cite this article as: Palmer, S., Cox, S.F., McKinley, J.M., Ofterdinger, U., Soil-geochemical factors

controlling the distribution and oral bioaccessibility of nickel, vanadium and chromium in soil, Applied

Geochemistry (2014), doi: http://dx.doi.org/10.1016/j.apgeochem.2014.10.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers

we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and

review of the resulting proof before it is published in its final form. Please note that during the production process

errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

��

SOIL-GEOCHEMICAL FACTORS CONTROLLING THE DISTRIBUTION AND ORAL BIOACCESSIBILITY OF NICKEL, VANADIUM AND CHROMIUM IN SOIL Sherry Palmer*,a, Siobhan F. Coxa, Jennifer M. McKinleyb and Ulrich Ofterdingera

aSchool of Planning, Architecture and Civil Engineering; Queen’s University Belfast, BT9 5AG, UK bSchool of Geography, Archaeology and Palaeoecology; Queen’s University Belfast, BT7 1NN, UK *Corresponding author; spalmer04@qub.ac.uk, Tel. +44 (028) 9097 5606, Fax +44 (028) 9097 4278 Additional e-mail addresses; s.cox@qub.ac.uk, j.mckinley@qub.ac.uk, u.ofterdinger@qub.ac.uk

Abstract: Geogenic nickel (Ni), vanadium (V) and chromium (Cr) are present at elevated levels in soils in Northern Ireland. Whilst Ni, V and Cr total soil concentrations share common geological origins, their respective levels of oral bioaccessibility are influenced by different soil-geochemical factors. Oral bioaccessibility extractions were carried out on 145 soil samples overlying 9 different bedrock types to measure the bioaccessible portions of Ni, V and Cr. Principal component analysis identified two components (PC1 and PC2) accounting for 69% of variance across 13 variables from the Northern Ireland Tellus Survey geochemical data. PC1 was associated with underlying basalt bedrock, higher bioaccessible Cr concentrations and lower Ni bioaccessibility. PC2 was associated with regional variance in soil chemistry and hosted factors accounting for higher Ni and V bioaccessibility. Eight per cent of total V was solubilised by gastric extraction on average across the study area. High median proportions of bioaccessible Ni were observed in soils overlying sedimentary rock types. Whilst Cr bioaccessible fractions were low (max = 5.4%), the highest measured bioaccessible Cr concentration reached 10.0 mg kg-1, explained by factors linked to PC1 including high total Cr concentrations in soils overlying basalt bedrock. �

Keywords: trace elements, bioaccessibility, geochemistry, factor analysis, human health risk assessment

�

�

�

�

�

�

�

��

1.0 Introduction

In Northern Ireland, nickel (Ni), vanadium (V) and chromium (Cr) are present in soils at high total

concentrations, in some cases exceeding available statutory soil guideline values (SGVs) or generic

assessment criteria (Environment Agency, 2009; Nathaniel et al., 2009). SGVs are intended for use as

screening values; significant harm may not necessarily be present when a contaminant exceeds its SGV,

but regulators may wish to seek additional evidence to determine if there is a high probability of

significant harm to receptors from soil contaminant exposure under such circumstances (DEFRA,

2012).

Refining the human health risk assessment process with respect to oral exposure to soil borne

contaminants plays an important role in sustainable development approaches. Informed land

management practices and accurate risk communication can save local councils, land owners, and

developers millions of pounds in remediation costs, in turn preserving natural soil resources (NERC,

2012). Soil bioaccessibility testing can be applied to complement a lines-of-evidence approach when

assessing health risks from oral soil contaminant exposure (CIEH, 2009; DEFRA, 2012). In terms of

risk estimation, in vitro oral bioaccessibility testing reduces reliance on total contaminant concentrations

in soil, an approach which commonly will overestimate health risks due to the well-accepted concept

that 100% of soil contamination is rarely bioavailable to receptors when ingested (Ruby et al., 1999;

CIEH, 2009). Compared to in vivo animal bioavailability studies, in vitro testing is also less costly and

can be completed in a shorter time frame while avoiding the ethical concerns of using live animals in

research.

Over the past two decades, a considerable amount of effort has been invested across the scientific

community in developing and validating a suitable standardised extraction method for estimating human

oral contaminant bioaccessibility in vitro (Ruby et al., 1996; Ruby et al., 1999; Oomen et al., 2003;

Wragg and Cave, 2003; Van de Weile et al., 2007; Wragg et al., 2009; NERC, 2012). The Unified

BARGE (Bioaccessibility Research Group of Europe) Method (UBM) is a standardised physiologically

based extraction technique which has been subjected to comprehensive inter-laboratory trials and

validated for arsenic (As), lead (Pb) and cadmium (Cd) using in vivo swine data (Caboche, 2009; Wragg

et al., 2011; Denys et al., 2012).

Previous trace element bioaccessibility research in Northern Ireland has shown that oral bioaccessibility

is not primarily a function of total trace element concentrations in soil (Barsby et al, 2012; Palmer et al.,

2013; Cox et al., 2013), diminishing the suitability of reliance on total soil contaminant concentrations

as a driver for contaminated land management decisions. Following from these findings, specific

geochemical factors accounting for the variability in trace element bioaccessibility across the country

��

require further investigation. Correlation studies exploring relationships between trace element

bioaccessibility and other geochemical variables found statistically significant relationships which

varied across different underlying bedrock types (Palmer et al., 2013; Cox et al. 2013). For example,

sulphur measured as sulphur trioxide (SO3) and loss on ignition (LOI) data showed strong positive

associations with Ni and V bioaccessibility throughout Northern Ireland, whilst calcium appeared to

exert either positive or negative influences over bioaccessibility depending upon the underlying bedrock

type (Palmer et al., 2013).

Although several statistically significant intrinsic correlations between trace element bioaccessibility

and a number of other geochemical variables have been previously confirmed, specific soil-geochemical

associations are difficult to identify from correlation analysis alone. This paper therefore aims to more

explicitly identify such soil-geochemical associations on a regional scale through a combined analysis

approach. Exploratory data analysis (EDA), factor analysis and geostatistics were used to validate and

identify common sources of variability across two data sets (Gooaverts, 1992; Abollino et al., 2011;

Candeias et al., 2011; Giacomino et al., 2011), supporting the research aim of identifying geochemical

factors that account for variability in the distribution and bioaccessibility of Ni, V and Cr in soils across

Northern Ireland. Critically, such findings have applications beyond the study area as the diverse

geology encountered in Northern Ireland extends stratigraphically beyond the immediate region.

2.0 Methodology

2.1 Study Area and Sample Selection

Northern Ireland hosts a wide variety of rock types relative to its small land area of just over 13,800 km2

with most geologic time periods represented by the study area’s diverse lithology (Fig. 1; Wilson, 1972;

Jordan et al., 2007). Bioaccessibility research in this study area therefore provides the opportunity to

increase the global understanding of the relationships between geochemistry and human health within

both the scientific and regulatory communities.

The range of rock types represents three basement terranes: the Grampian, the Midland Valley, and the

Southern Uplands-Down-Longford Terranes. As described in Mitchell (2004), the Grampian Terrane

and associated rocks in the northwest represent the oldest rocks in the country, with their metamorphic

igneous and sedimentary origins spanning the Proterozoic Era. Psammites and semi-pelites are the

dominant rock type, with sandstone and conglomerate also present in this area. Following from the

Proterozoic through to the Devonian, the Midland Valley Terrane hosts Palaeozoic igneous formations

and Late Devonian-Early Carboniferous sedimentary rocks. Red sandstones, limestone and mudstones

comprise the majority of rock types with a minor conglomerate component. Finally, granitic igneous

intrusives and marine sedimentary rocks representing the Lower Palaeozoic Ordovician and Silurian

��

systems are found in the Southern Uplands-Down-Longford Terrane in the southwest. Lithic arenites

and sandstones comprise the majority of bedrock in this part of the country. Portions of the Midland

Valley and Grampian basement rocks in the northeast are covered by a large plateau of igneous

Palaeogene basalts and lava-derived sedimentary clays originating from within the Permian to

Cretaceous systems to a depth of up to 800 meters. Outside of this area, much of the basement rocks in

Northern Ireland remain exposed. In contrast to the diverse range of underlying rock types, superficial

deposits are dominated by glacial till, sand and gravel, and peat, and topsoil pH falls within a relatively

narrow acidic range.

Eleven generic bedrock types as defined by the Geological Survey of Northern Ireland (GSNI; Smyth,

2007) are illustrated in Fig. 1. Geochemical and bioaccessibility data were separated into these groups

prior to EDA with the aim of identifying potential differences in soil geochemistry and trace element

bioaccessibility across different underlying bedrock types. Using generic bedrock types ensures that

sample sizes remain sufficiently robust during analysis and EDA, as defining discrete minor geologic

formations present in Northern Ireland would reduce sample sizes to a prohibitively narrow range of

statistical significance. Such an approach also reduces variance and increases normality in the

distribution of geochemistry data (Jordan et al., 2007; Zhang et al., 2007; Palmer et al., 2013).

Geochemistry data from the Northern Ireland Tellus Survey were provided by the GSNI. In excess of

6800 shallow profile soil samples from 5 – 20 cm depth were gathered during the Tellus Survey and

analysed by X-ray fluorescence spectrometry (XRFS) for major oxides and trace elements. The

complete Tellus Survey field methods and analytical methodology are described in Smyth (2007).

Shallow soil samples from the Northern Ireland Tellus Survey soil archive were used for

bioaccessibility testing and were selected to represent the diversity of rock and soil types across the

study area.

2.2 Bioaccessibility Testing

The UBM is an in vitro extraction procedure for solid matrices designed to mimic the conditions of the

human digestive system. The method incorporates saliva, gastric, and intestinal phases of digestion at a

controlled temperature of 37◦ C. The full methodology is downloadable through the British Geological

Survey (BGS) website (BARGE/INERIS, 2011).

Soil sub-sampling, solution preparation, UBM extractions and inductively coupled plasma mass

spectrometry (ICP-MS) analyses were conducted in the Analytical Geochemistry Facility at BGS,

Kingsley Dunham Centre, Keyworth, Nottingham during 2009 (Barsby et al., 2012) and 2013.

Archived Northern Ireland Tellus soils from the < 2mm fraction were sub-sampled by coning and

��

quartering to ensure representative samples were obtained prior to extraction. The 2009 and 2013

bioaccessibility data sets were joined to ensure the diverse range of rock types present within the study

area was adequately represented.

Fresh digestion fluids were prepared daily and one day prior to use to allow enzymes and pH to

stabilise. Inorganic and organic solutions for each fluid were prepared in separate 500 ml glass

volumetric flasks prior to combining in two litre non-reactive screw top Nalgene bottles to create one

litre of digestive fluid. Additional solid reagents were weighed directly into the Nalgene bottles prior to

adding the inorganic and organic solutions. Digestive fluids were immersed in the same warm water

bath used for the extractions at 37◦C for a least one hour prior to commencing extractions. The pH of

each solution was checked and adjusted to the specifications outlined in Table 1 using either 37% HCl

or 1M NaOH.

2.2.1 Extract Analysis and Quality Control

The UBM has been validated for As, Pb and Cd through in vivo studies (Caboche, 2009; Denys et al.,

2012). It should be noted that, whilst the three trace elements presented in this paper have not been

subjected to in vivo validation under the UBM, data for all trace elements of interest were subjected to

rigorous quality control (QC) and quality assurance (QA) procedures as described below.

Extract analysis was carried out using an Agilent 7500cx series ICP-MS employing an octopole reaction

system in combination with a CETAC autosampler. Sample introduction from the autosampler to the

ICP is coupled to a CETAC flow injection valve. The instrument was calibrated at the beginning of

every analytical run using a minimum of three standards and a blank for each trace element. Multi-

element QC check standards containing the trace elements of interest at 25 μg/L were analysed at the

start and end of each run and after a minimum of every 25 samples.

Accuracy and precision of UMB extract analysis is monitored at BGS by the on-going collation of

BGS102 reference material data (Wragg, 2009). UBM results from ICP analysis were provided in μg/L

with a dilution factor of 100 applied. Data were back calculated to mg kg-1 of soil using Equation A,

taking into account recorded soil sub-sample weights and total fluid volumes inclusive of HCl or NaOH

additions used for pH adjustments.

� ��� ��� � ��� ��⁄ ⁄ �

�

�

� � 1000

(Equation A);

��

where C = analyte concentration in units specified, V = total extraction fluid volume (mL) and W = total

soil weight (g). Soils were pre-dried according to the Tellus methodology (Smyth, 2007); 100% solids

were therefore assumed for calculating final concentrations.

For every seven unknown soil samples, one extraction blank, one duplicate soil and one BGS102

certified reference soil were included in the extraction procedure. Measured values for bioaccessible As

and Pb were within one standard deviation of the certified reference value provided for BGS102.

Overall mean gastric relative per cent difference for the analytes of interest was 11%. Values measured

in blanks above daily detection limits were subtracted from measured sample values for each respective

batch. Measured sample values below daily detection limits were reported as half the daily detection

limit.

2.3 Geochemistry and UBM Data Analysis

Analysis of geochemistry and trace element bioaccessibility data relied upon a combined approach

where intrinsic statistical relationships among variables were evaluated in the context of corresponding

spatial outputs. Multivariate dimension reduction in the form of PCA was applied to link variables to a

narrower range of soil-geochemical components such as specific soil or bedrock types. Conclusions

were subsequently drawn about specific soil-geochemical influences causing variability in the

geographical distribution and bioaccessibility of Ni, V and Cr.

Summary statistics were generated for XRFS geochemistry data in Microsoft (MS) Excel 2010 and IBM

SPSS v.19.0. Boxplot figures were completed in R (R Core Team, 2013). PCA was conducted in R

using the vegan package (Oksanen et al., 2013) and eigenvalue loading plots with eigenvectors were

manually constructed using PCA eigenvalue factor loadings. Geochemistry variables, synonymous with

factors for the purposes of PCA, were selected on the basis of their known potential to influence trace

element bioaccessibility (Stewart et al., 2003; Subacz et al., 2007; Poggio et al., 2009; Meunier et al.

2010; Bradham et al., 2011) and consisted of eight major element oxides (%), pH, loss on ignition (%;

LOI), and total Ni, V, and Cr (mg kg-1).

To address issues of data closure encountered in compositional geochemical data, an additive log ratio

transformation was applied to the data set using the robComposition package in R (Templ et al., 2009).

Titanium dioxide was chosen as the rationing variable for transformation (Reimann et al., 2009).

Values for the thirteen geochemistry factors noted above from 6,862 sample locations were analysed by

PCA using scaled Northern Ireland Tellus XRFS data for shallow soils. Scaling data corrects for

variability in the magnitude of values that will naturally be present in a data set with multiple factors

(Oksanen et al., 2013). To explore relationships between trace element bioaccessibility and Northern

�

Ireland geochemistry, UBM data inclusive of XRFS data from corresponding study set soil sample

locations were also analysed by PCA following the same methods described above.

PCA is an unconstrained ordination method that is useful for dimension reduction in multivariate data.

Data are transposed onto a three dimensional axis where the Euclidian distance between points is

calculated in each direction. Similar to a one-dimensional linear regression, a line of best fit is

calculated accounting for maximum variance between the data points; this vector is the first principal

component (PC1). The second principal component (PC2) is the vector which lies perpendicular to

PC1. Data are rotated in three dimensional space to repeat this process and derive subsequent

components. Eigenvalues represent the proportion of factor variance accounted for by a particular

component, providing the same level of information as an r2 value in linear regression. In addition to

factor eigenvalues, each soil sample location is assigned a unique site score. Site scores give an

indication of individual data positions and resulting relative weightings along component vectors in

three dimensional space. A higher score indicates more sample variance is accounted for by a

prescribed component (Reimann et al., 2009; ESRI, 2010; Oksanen et al., 2013).

Geochemistry variables and PCA site scores were mapped in ArcMap v.10 using ordinary kriging.

Semi-variograms were manually fitted to geochemistry data using an iterative process of varying search

neighbourhoods until the model yielding the least mean prediction error and best fit semi-variogram was

generated. Spatial PCA applies identical ordination methods as described previously, but additionally

generates raster outputs of unique site scores against each principal component to facilitate interpolation

of a site score surface. Final interpolated outputs show sample location variance accounted for by the

specified principal component (ESRI, 2010; Candeias et al., 2011).

3.0 Results

3.1 Regional Trends in Geochemistry

�

Table 2 provides summary statistics for oxides, trace elements, pH and LOI. Ni, Cr and V

concentrations are highest in soils overlying basalt bedrock types with their distributions controlled by a

long-range (a > 70 km) spatial function aligning with the extent of basalts in the northeast of the study

area (Figs. 2 and 3, Table 3). Ni and Cr exhibit low concentrations in soils overlying the 10 other

bedrock types investigated compared to V, which displays more variability in its distribution across the

study area (Figs. 2 and 3b). In particular, soils overlying sandstone and psammite & semi-pelite

bedrock types exhibit high concentrations of V.

Magnesium (Mg), calcium (Ca), manganese (Mn) and iron (Fe) oxides show high relative

concentrations in soils overlying basalt rock types (Figs. 4-6). Mg, Ca and Fe are present in the

�

minerals olivine, plagioclase feldspar and augite, which together account for approximately 90% of the

minerals in underlying basalt bedrock (Hill et al., 2001). Mn is also known to be elevated in basalts as it

readily substitutes for Fe in olivine (Wedepohl, 1978). Aluminium oxide (Al2O3) concentrations also

appear to be controlled by underlying basalt, although Al distributions are not as strongly constrained by

this bedrock type (Fig. 6b).

Spat

ially

, ph

osph

orou

s pe

ntox

ide

(P2O

5) d

ispl

ays

a hi

gh f

requ

ency

of

chan

ge i

n co

ncen

trat

ions

thr

ough

out

the

regi

on,

with

litt

le a

ppar

ent

asso

ciat

ion

to

unde

rlyi

ng b

edro

ck ty

pes

(Fig

s. 4

and

6d)

. T

his

smal

l-sc

ale

spat

ial d

istr

ibut

ion

is a

lso

disp

laye

d in

Tab

le 3

, whe

re P

2O5

is c

ontr

olle

d by

a s

hort

-ran

ge s

patia

l

func

tion

(a =

17

km)

with

ove

r ha

lf o

f its

tota

l var

ianc

e (C

) ac

coun

ted

for

by a

nug

get e

ffec

t (C

0).

A la

rge

nugg

et v

aria

nce

and

a sh

ort-

rang

e sp

atia

l fun

ctio

n

sugg

est

that

sm

all

scal

e pr

oces

ses

not

asso

ciat

ed w

ith u

nder

lyin

g ge

olog

y ar

e in

flue

ncin

g a

mod

elle

d pa

ram

eter

(D

ober

man

n et

al.,

199

5; E

inax

and

Sol

dt,

1999

). E

leva

ted

P 2O

5 in

sha

llow

soi

ls is

like

ly to

be

the

resu

lt of

agr

icul

tura

l pra

ctic

es s

uch

as p

hosp

hate

fer

tiliz

er a

pplic

atio

ns, a

n ob

serv

atio

n w

hich

has

als

o

been

not

ed b

y Jo

rdan

et

al.

(200

1).

Las

tly,

limes

tone

s, m

udst

ones

and

san

dsto

nes

unde

rlie

soi

ls h

ostin

g hi

gh m

ean

conc

entr

atio

ns o

f si

licon

dio

xide

(Si

O2;

Figs

. 4 a

nd 6

c), c

hara

cter

istic

of

the

pare

nt b

edro

ck m

ater

ial t

hat w

ill c

ontr

ibut

e to

hig

h sa

nd a

nd c

lay

min

eral

con

tent

in s

oil (

Spar

ks, 1

995)

.

LO

I, p

H a

nd s

ulph

ur t

riox

ide

(SO

3) a

re s

tron

gly

corr

elat

ed s

patia

lly a

cros

s N

orth

ern

Irel

and,

with

dis

trib

utio

n pa

ttern

s al

igni

ng w

ith a

cidi

c up

land

pea

t so

il

type

s w

hich

ove

rlie

mou

ntai

nous

for

mat

ions

com

pris

ed o

f ac

id v

olca

nic

and

psam

mite

& s

emi-

pelit

e ro

ck t

ypes

(Fi

gs. 6

e an

d 7;

Jor

dan

et a

l., 2

001)

. W

hils

t

aver

age

pH a

nd L

OI

rem

ain

rela

tivel

y co

nsta

nt a

cros

s be

droc

k gr

oups

in

the

stud

y ar

ea,

thes

e pa

ram

eter

s sh

ow m

ore

stat

istic

al i

rreg

ular

ity i

n th

e ac

id

volc

anic

and

psa

mm

ite &

sem

i-pe

lite

grou

ps (

Fig.

5).

In

add

ition

to

the

high

est

obse

rved

pro

port

ions

of

LO

I be

ing

loca

ted

in u

plan

d pe

aty

area

s,

inte

rmed

iate

leve

ls o

f L

OI

coul

d al

so b

e as

soci

ated

with

the

pres

ence

of

clay

min

eral

s in

oth

er p

arts

of

the

stud

y ar

ea (

Fig.

7),

as

the

LO

I m

etho

d ca

n re

sult

in

stru

ctur

al w

ater

loss

fro

m c

lay

in a

dditi

on to

pro

vidi

ng a

n es

timat

ion

of s

oil c

arbo

n co

nten

t (K

onen

et a

l., 2

002)

. T

he p

rese

nce

of s

ulph

ur c

ompo

unds

in p

eat

soils

may

be

indi

cativ

e of

mic

robi

al p

roce

sses

occ

urri

ng u

nder

red

ucin

g, w

ater

-log

ged

cond

ition

s or

cou

ld b

e as

soci

ated

with

atm

osph

eric

pol

lutio

n so

urce

s

(Cam

pbel

l, 19

96a;

Cam

pbel

l, 19

96b;

Neb

el a

nd W

righ

t, 20

00).

3.2

Ni,

V a

nd C

r B

ioac

cess

ibili

ty

� Mea

n an

d m

axim

um b

ioac

cess

ible

fra

ctio

ns (

BA

F; %

) fo

r N

i, V

and

Cr

wer

e hi

ghes

t in

the

gast

ric

phas

e. I

n lin

e w

ith b

est p

ract

ice

for

ensu

ring

hum

an h

ealth

risk

s ar

e no

t un

dere

stim

ated

, gas

tric

dat

a ar

e th

eref

ore

pres

ente

d to

pro

vide

a c

onse

rvat

ive

estim

ate

of t

he m

axim

um a

mou

nt o

f tr

ace

elem

ents

tha

t w

ould

be

pote

ntia

lly b

ioav

aila

ble.

N

ine

gene

ric

bedr

ock

clas

sifi

catio

ns w

ere

capt

ured

by

the

bioa

cces

sibi

lity

stud

y se

t (F

ig.

8).

BA

F ar

e pr

esen

ted

in T

able

4

���

acco

rdin

g to

gen

eric

bed

rock

typ

e w

ith g

abbr

o ex

clud

ed d

ue t

o th

e sm

all

sam

ple

size

in

this

cla

ssif

icat

ion.

B

AF

prov

ide

a pr

efer

red

basi

s fo

r co

mpa

riso

n

acro

ss b

edro

ck g

roup

s du

e to

nor

mal

isat

ion

by to

tal e

lem

ent c

once

ntra

tions

.

On

aver

age,

app

roxi

mat

ely

12%

of

tota

l N

i in

sha

llow

soi

ls a

cros

s N

orth

ern

Irel

and

was

sol

ubili

sed

by g

astr

ic e

xtra

ctio

n, e

quiv

alen

t to

a m

ean

bioa

cces

sibl

e

conc

entr

atio

n of

3.8

mg

kg-1

. T

he h

ighe

st q

uant

ity o

f bi

oacc

essi

ble

Ni

was

mea

sure

d at

30.

7 m

g kg

-1 a

nd o

ccur

red

in t

he l

ithic

are

nite

bed

rock

gro

up;

how

ever

, the

hig

hest

pro

port

ion

of b

ioac

cess

ible

Ni w

as o

bser

ved

in th

e ps

amm

ite &

sem

i-pe

lite

grou

p (m

ax =

46.

3%; T

able

4).

Soi

ls o

verl

ying

sed

imen

tary

limes

tone

s, s

ands

tone

s an

d m

udst

ones

had

hig

her

mea

n pr

opor

tions

of

bioa

cces

sibl

e N

i ov

eral

l co

mpa

red

to s

oils

ove

rlyi

ng o

ther

roc

k ty

pes.

A

lthou

gh t

he

high

est t

otal

Ni c

once

ntra

tions

are

obs

erve

d in

soi

ls o

verl

ying

bas

alt (

Figs

. 2 a

nd 3

), th

e lit

hic

aren

ite a

nd m

udst

one

bedr

ock

grou

ps h

oste

d up

to tw

ice

(max

=

30.7

mg

kg-1

) an

d on

e an

d a

half

tim

es (

max

= 2

4.2

mg

kg-1

) as

muc

h bi

oacc

essi

ble

Ni,

resp

ectiv

ely,

tha

n w

as m

easu

red

in t

he b

asal

t be

droc

k gr

oup

(max

=

15.7

mg

kg-1

).

Med

ian

Ni

BA

Fs w

ere

low

est

in t

he c

ongl

omer

ate

and

lithi

c ar

enite

gro

ups,

how

ever

, w

hile

the

hig

hest

med

ian

occu

rred

in

soils

ove

rlyi

ng

sand

ston

e.

App

roxi

mat

ely

8% o

f to

tal

V w

as f

ound

to

be b

ioac

cess

ible

on

aver

age

acro

ss 1

45 s

ampl

e lo

catio

ns,

with

mea

sure

d av

erag

e an

d m

axim

um b

ioac

cess

ible

conc

entr

atio

ns o

f 7.

4 an

d 52

.1 m

g kg

-1, r

espe

ctiv

ely.

Max

imum

pro

port

iona

l bio

acce

ssib

ility

occ

urre

d in

the

sand

ston

e gr

oup

(23.

1%),

alth

ough

this

was

onl

y

mar

gina

lly g

reat

er t

han

the

max

imum

BA

F of

22.

5% m

easu

red

in t

he b

asal

t be

droc

k gr

oup

(Tab

le 4

).

Sim

ilar

valu

es w

ere

reco

rded

in

the

mud

ston

e gr

oup,

with

up

to 2

2% o

f V

ren

dere

d so

lubl

e du

ring

gas

tric

ext

ract

ion.

O

vera

ll, t

he h

ighe

st m

ean

and

med

ian

leve

ls o

f ga

stri

c V

bio

acce

ssib

ility

wer

e re

cord

ed i

n

soils

ove

rlyi

ng b

asal

t bed

rock

, bot

h in

term

s of

pro

port

iona

l and

abs

olut

e m

easu

rem

ents

. T

he m

edia

n bi

oacc

essi

ble

V c

once

ntra

tion

in s

oils

ove

rlyi

ng b

asal

t

bedr

ock

was

11.

5 m

g kg

-1 (B

AF

= 1

0.3%

). M

edia

n V

BA

Fs r

ange

d fr

om 5

.0%

to 8

.6%

acr

oss

the

rem

aini

ng b

edro

ck g

roup

s (T

able

4).

Cr

disp

laye

d a

max

imum

BA

F of

5.4

% o

f its

tota

l con

cent

ratio

n, w

ith a

mea

n an

d m

edia

n B

AF

of 1

.2%

and

1.0

%, r

espe

ctiv

ely,

acr

oss

the

stud

y se

t (T

able

4).

Des

pite

low

pro

port

ions

of

Cr

bein

g bi

oacc

essi

ble,

9 -

10

mg

kg-1

of

gast

ric

bioa

cces

sibl

e C

r w

as e

xtra

cted

fro

m tw

o sa

mpl

es, w

ith m

easu

red

bioa

cces

sibi

lity

rang

ing

betw

een

5 -

6 m

g kg

-1 a

t a

furt

her

thre

e sa

mpl

e lo

catio

ns.

Eac

h of

the

se f

ive

soils

obs

erve

d to

hav

e re

lativ

ely

high

lev

els

of C

r bi

oacc

essi

bilit

y

occu

rred

in

sam

ples

ove

rlyi

ng b

asal

t be

droc

k.

The

max

imum

Cr

BA

F al

so o

ccur

red

in t

he b

asal

t be

droc

k gr

oup,

alth

ough

the

mea

n B

AF

was

hig

hest

in

the

gran

ite b

edro

ck g

roup

(1.

7%; F

ig. 8

). A

vera

ge to

tal C

r bi

oacc

essi

bilit

y ac

ross

the

stud

y ar

ea w

as m

easu

red

at 1

.5 m

g kg

-1.

Med

ian

bioa

cces

sibi

lity

acro

ss a

ll

���

bedr

ock

grou

ps r

ange

d fr

om 0

.9%

- 1

.4%

(T

able

4),

with

the

hig

hest

med

ian

BA

F ob

serv

ed i

n so

ils o

verl

ying

bas

alt.

It

shou

ld b

e no

ted

that

pot

entia

l he

alth

risk

fro

m C

r in

soi

l is

high

ly d

epen

dent

upo

n sp

ecia

tion

whi

ch w

as n

ot id

entif

ied

as p

art o

f th

is s

tudy

(N

atha

nail

et a

l., 2

009)

.

3.3

Iden

tifie

d Pr

inci

pal C

ompo

nent

s

� Prin

cipa

l co

mpo

nent

ana

lysi

s of

the

ful

l ge

oche

mis

try

data

set

with

thi

rtee

n se

lect

ed v

aria

bles

ide

ntif

ied

thre

e co

mpo

nent

s ac

coun

ting

for

a cu

mul

ativ

e

vari

ance

of

79%

acr

oss

Nor

ther

n Ir

elan

d.

Ana

lysi

s of

the

UB

M d

ata

set

foun

d fo

ur c

ompo

nent

s al

so a

ccou

ntin

g fo

r 79

% o

f va

rian

ce (

Tab

le 5

).

Fact

or

load

ings

for

eac

h co

mpo

nent

are

pre

sent

ed in

Tab

le 6

. PC

1, li

nked

to v

aria

bles

ass

ocia

ted

spat

ially

with

bas

alt,

acco

unte

d fo

r th

e gr

eate

st a

mou

nt o

f va

rian

ce

(45%

) in

the

data

. A

fur

ther

24%

of

geoc

hem

ical

var

ianc

e fr

om P

C2

is e

xpla

ined

by

vari

able

s as

soci

ated

with

sed

imen

tary

for

mat

ions

suc

h as

Ca,

Fe

and

Mg

(Tab

le 6

). S

uch

elem

ents

are

abu

ndan

t in

min

eral

s of

low

har

dnes

s su

ch a

s do

lom

ite, c

alci

te a

nd c

hlor

ite (

Tar

buck

and

Lut

gens

, 199

9) a

lthou

gh th

e pr

esen

ce

of t

hese

spe

cifi

c m

iner

als

in t

he s

tudy

are

a is

not

con

firm

ed.

In

the

UB

M s

tudy

set

, C

r is

the

onl

y tr

ace

elem

ent

who

se b

ioac

cess

ible

con

cent

ratio

ns w

ere

wei

ghte

d to

war

ds t

he s

ame

geol

ogic

com

pone

nt a

s its

tot

al c

once

ntra

tions

, hi

ghlig

htin

g its

def

initi

ve g

eoge

nic

cont

rols

(T

able

6).

B

ioac

cess

ible

Ni

and

V

conc

entr

atio

ns,

how

ever

, w

ere

mor

e cl

osel

y as

soci

ated

with

fac

tors

spa

tially

rel

ated

to

acid

pea

t so

il ch

emis

try

such

as

SO3

and

LO

I an

d al

so w

ith f

acto

rs

sugg

estiv

e of

less

rec

alci

tran

t min

eral

s.

Eig

enva

lue

load

ing

plot

s m

ay b

e us

ed t

o ill

ustr

ate

both

ind

ivid

ual

sam

ple

site

loa

ding

s an

d fa

ctor

loa

ding

s ag

ains

t id

entif

ied

prin

cipa

l co

mpo

nent

s.

Figu

res

9a a

nd 9

b ill

ustr

ate

fact

or lo

adin

gs o

nly

for

purp

oses

of

clar

ity d

ue to

the

high

num

ber

of s

oil s

ampl

e si

tes

anal

ysed

by

PCA

. W

here

obt

use

angl

es a

re p

rese

nt

betw

een

vect

ors

in t

he p

lot,

a ne

gativ

e co

rrel

atio

n am

ong

two

vari

able

s is

exp

ecte

d.

Con

vers

ely,

acu

te a

ngle

s su

gges

t a

posi

tive

rel

atio

nshi

p be

twee

n tw

o

fact

ors,

with

sm

alle

r de

gree

s of

sep

arat

ion

indi

catin

g a

stro

nger

cor

rela

tion

(Rei

man

n et

al.,

200

9).

Tot

al N

i, V

and

Cr

and

Fe a

nd M

n ox

ides

for

m a

clu

ster

of

vect

ors

in th

e bo

ttom

rig

ht q

uadr

ant o

f th

e pl

ot in

Fig

. 9a,

dem

onst

rativ

e of

thei

r ge

oche

mic

al c

o-

occu

rren

ces

in b

asal

t. T

he p

rese

nce

of A

l ox

ides

is

infe

rred

by

less

tha

n 90

deg

ree

angl

es b

etw

een

the

eige

nvec

tors

of

Al 2

O3,

MgO

and

CaO

, a s

olid

pha

se

whi

ch w

as a

lso

iden

tifie

d by

Cox

et a

l. (2

013)

. A

l mai

ntai

ns m

ore

than

one

geo

chem

ical

ass

ocia

tion,

how

ever

, dem

onst

rate

d by

its

corr

elat

ion

also

with

SiO

2

���

in t

he l

oadi

ng p

lot,

affi

rmin

g th

e in

flue

nce

of c

lay

min

eral

s in

the

stu

dy a

rea

(The

ng, 1

974;

Spa

rks,

199

5; C

ox e

t al

., 20

13).

L

OI

and

SO3

are

also

str

ongl

y

corr

elat

ed, s

ugge

stin

g a

soil-

geoc

hem

ical

inf

luen

ce f

rom

pea

t so

il ty

pes.

T

he p

rese

nce

of s

edim

enta

ry m

iner

als

is i

llust

rate

d by

the

ass

ocia

tions

bet

wee

n C

a

and

Mg

oxid

es in

Fig

. 9a

(Tar

buck

and

Lut

gens

, 199

9).

Furt

herm

ore,

oxi

des

of th

ese

two

elem

ents

are

in c

lose

r pr

oxim

ity to

Ni i

n th

e lo

adin

g pl

ot r

elat

ive

to

Cr

or V

, sug

gest

ing

that

Ni r

esid

es in

dif

fere

nt h

ost m

iner

al p

hase

s th

an C

r or

V.

Insp

ectio

n of

Fig

. 9b

inc

lusi

ve o

f U

BM

ext

ract

ion

data

sho

ws

that

bio

acce

ssib

le c

once

ntra

tions

of

Ni,

V a

nd C

r ar

e st

ill c

lose

ly a

ssoc

iate

d w

ith t

heir

tot

al

conc

entr

atio

ns, a

lthou

gh th

is f

igur

e ill

ustr

ates

PC

1 an

d PC

2 on

ly.

Of

part

icul

ar n

ote

is th

e gr

eate

r pr

oxim

ity o

f bi

oacc

essi

ble

Ni c

once

ntra

tions

to C

a an

d M

g

oxid

es in

the

load

ing

plot

than

tota

l Ni c

once

ntra

tions

, illu

stra

ting

diff

erin

g ge

oche

mic

al a

ssoc

iatio

ns b

etw

een

tota

l and

bio

acce

ssib

le N

i.

Com

pone

nt s

ampl

e sc

ores

are

spa

tially

illu

stra

ted

in F

igs.

10a

-c.

In

line

with

the

abo

ve P

CA

int

erpr

etat

ions

, sa

mpl

e sc

ores

are

hea

vily

wei

ghte

d to

PC

1 in

the

basa

lt re

gion

s, w

ith m

inor

var

iabi

lity

also

acc

ount

ed f

or b

y un

derl

ying

cla

y an

d sa

ndst

one

rock

type

s. S

patia

l pat

tern

s ill

ustr

ated

by

PC1

indi

cate

bed

rock

geol

ogy

prov

ides

the

prim

ary

basi

s fo

r ge

oche

mic

al v

aria

bilit

y. P

C2

conf

orm

s to

a s

patia

l pat

tern

mor

e cl

osel

y al

igne

d w

ith r

egio

nal v

aria

bilit

y in

soi

l typ

es,

illus

trat

ing

the

addi

tiona

l im

port

ance

of

soil

chem

istr

y.

Alth

ough

Tab

le 5

sho

ws

the

thir

d m

inor

com

pone

nt a

ccou

nts

for

only

10

and

9% o

f va

rian

ce i

n th

e

full

geoc

hem

istr

y an

d U

BM

stu

dy s

ets,

res

pect

ivel

y, s

patia

l PC

A e

ffec

tivel

y ill

ustr

ates

the

im

pact

of

soil-

geoc

hem

ical

con

trol

s fr

om P

C3

rela

ting

to s

oil

acid

ity (

Tab

le 6

).

The

cla

rity

of

spat

ial

patte

rns

of P

C3

in F

ig.

10c

addi

tiona

lly s

ugge

sts

that

exp

lori

ng g

eoch

emis

try

in t

he a

bsen

ce o

f st

rong

sou

rces

of

vari

ance

suc

h as

the

basa

lts m

ay r

evea

l add

ition

al s

oil-

geoc

hem

ical

com

pone

nts

othe

rwis

e m

aske

d by

the

basa

lts.

4.0

Dis

cuss

ion

� The

min

imum

Ni

BA

F oc

curr

ed i

n th

e ba

salt

bedr

ock

grou

p.

Tot

al N

i, M

gO, A

l 2O

3, P

2O5,

MnO

and

Fe 2

O3,

pre

viou

sly

obse

rved

by

Palm

er e

t al

. (20

13)

to

have

sig

nifi

cant

neg

ativ

e co

rrel

atio

ns w

ith N

i bi

oacc

essi

bilit

y, w

ere

also

fou

nd t

o be

pre

sent

at

high

con

cent

ratio

ns i

n so

ils o

verl

ying

bas

alt

com

pare

d to

the

othe

r be

droc

k gr

oups

(Fi

gs. 2

– 6

).

A s

igni

fica

nt p

ropo

rtio

n of

var

iabi

lity

in t

he a

bove

var

iabl

es i

s ex

plai

ned

by P

C1,

ass

ocia

ted

with

bas

alt

bedr

ock

in t

he

stud

y ar

ea,

supp

ortin

g th

e co

nclu

sion

tha

t N

i in

soi

ls o

verl

ying

bas

alt

bedr

ock

will

hav

e re

lativ

ely

low

bio

acce

ssib

le f

ract

ions

irr

espe

ctiv

e of

tot

al N

i so

il

���

conc

entr

atio

ns.

Thi

s is

like

ly d

ue to

the

pres

ence

of

min

eral

s of

hig

h re

calc

itran

ce s

uch

as o

livin

e, q

uart

z, f

elds

par

or p

yrox

ene

(Lyl

e, 1

980)

. T

he m

axim

um

Ni

BA

F w

as o

bser

ved

in s

oils

fro

m t

he p

sam

mite

& s

emi

pelit

e gr

oup.

A

ssoc

iatio

ns w

ith t

his

bedr

ock

grou

p su

gges

t a

peat

spa

tial

infl

uenc

e ra

ther

tha

n

cont

rol f

rom

und

erly

ing

bedr

ock

itsel

f, a

s th

is b

edro

ck g

roup

is c

over

ed la

rgel

y by

upl

and

peat

soi

l typ

es in

Nor

ther

n Ir

elan

d.

As

show

n by

Fig

s. 2

– 6

, soi

ls

over

lyin

g ps

amm

ite &

sem

i pel

ite h

osts

low

ave

rage

tota

l Ni,

low

ave

rage

con

cent

ratio

ns o

f M

g, A

l, Si

, Fe,

and

Mn

oxid

es p

revi

ousl

y ob

serv

ed to

neg

ativ

ely

infl

uenc

e N

i bio

acce

ssib

ility

, and

rel

ativ

ely

high

mea

sure

d SO

3 an

d L

OI.

In

the

UB

M d

ata

set t

he m

ajor

ity o

f va

rian

ce in

SO

3 an

d L

OI,

whi

ch h

ad p

revi

ousl

y

obse

rved

str

ong

posi

tive

corr

elat

ions

with

Ni

bioa

cces

sibi

lity,

was

acc

ount

ed f

or b

y PC

2.

Peat

soi

ls h

ave

high

org

anic

mat

ter

cont

ent

and

exhi

bit

acid

ic s

oil

pH, b

oth

of w

hich

are

exp

ecte

d to

incr

ease

Ni m

obili

ty, r

esul

ting

in h

ighe

r or

al b

ioac

cess

ibili

ty.

Ni p

osse

sses

a h

igh

affi

nity

for

soi

l org

anic

mat

ter

whi

ch c

an

incr

ease

Ni

soil

mob

ility

thr

ough

the

for

mat

ion

of s

olub

le o

rgan

ic c

hela

tes

(Allo

way

, 20

05).

In

add

ition

, N

i m

obili

ty i

ncre

ases

with

dec

reas

ing

pH

(Env

iron

men

t Age

ncy,

200

9; P

oggi

o et

al.,

200

9).

An

inve

stig

atio

n of

Ni b

ioac

cess

ibili

ty in

bas

alts

in N

orth

ern

Irel

and

usin

g ch

emom

etri

c ex

trac

tion

data

det

erm

ined

that

a la

rge

prop

ortio

n of

bio

acce

ssib

le N

i

in s

oil

sam

ples

ove

rlyi

ng b

asal

t w

as h

oste

d by

sol

uble

car

bona

te p

hase

s, a

lum

iniu

m o

xide

s an

d cl

ay.

A s

mal

l pr

opor

tion

of b

ioac

cess

ible

Ni

was

ass

ocia

ted

with

iro

n ox

ide,

alth

ough

the

non

-bio

acce

ssib

le p

ropo

rtio

n w

as h

oste

d by

the

rem

aini

ng i

ron

oxid

e m

iner

als

(Cox

et

al.,

2013

).

Whi

le t

hese

fin

ding

s w

ere

appl

icab

le i

n so

ils o

verl

ying

bas

alt,

the

gene

ral

regi

onal

tre

nd o

bser

ved

acro

ss N

orth

ern

Irel

and

appe

ars

to d

iffe

r sl

ight

ly f

rom

thi

s, w

ith v

aria

bles

ass

ocia

ted

with

cla

y, A

l an

d Fe

oxi

des

appe

arin

g to

hav

e a

nega

tive

infl

uenc

e ov

er N

i bi

oacc

essi

bilit

y (P

alm

er e

t al

., 20

13).

C

alci

um a

risi

ng f

rom

ins

olub

le m

iner

als

pres

ent

in b

asal

ts m

ay b

e as

soci

ated

with

red

uced

Ni

bioa

cces

sibi

lity

but

Ca

in s

oils

ove

rlyi

ng s

edim

enta

ry l

imes

tone

s w

ill a

ccou

nt f

or h

ighe

r N

i

bioa

cces

sibi

lity.

The

min

imum

V B

AF

occu

rred

in

the

psam

mite

& s

emi-

pelit

e gr

oup,

whi

le l

ow m

edia

n V

BA

Fs w

ere

obse

rved

in

soils

inc

lude

d in

the

lim

esto

ne a

nd l

ithic

aren

ite b

edro

ck g

roup

s.

Hig

her

mea

n an

d m

edia

n V

bio

acce

ssib

ility

(%

) w

as o

bser

ved

in s

oils

fro

m t

he b

asal

t be

droc

k sa

mpl

e gr

oup

com

pare

d to

oth

er

bedr

ock

type

s, w

here

tota

l soi

l V w

as a

lso

the

high

est.

Pal

mer

et a

l. (2

013)

iden

tifie

d si

mila

r co

rrel

atio

ns b

etw

een

maj

or o

xide

s, S

O3,

LO

I an

d bi

oacc

essi

ble

V a

s w

ere

obse

rved

for

bio

acce

ssib

le N

i; ho

wev

er,

resu

lts c

once

rnin

g m

ore

deta

iled

fact

ors

cont

rolli

ng V

bio

acce

ssib

ility

are

les

s co

nclu

sive

fro

m t

his

part

icul

ar s

tudy

. B

ased

on

prev

ious

cor

rela

tion

anal

yses

, fa

ctor

s re

latin

g to

V b

ioac

cess

ibili

ty w

ere

expe

cted

to

be s

imila

r to

tho

se i

dent

ifie

d fo

r N

i.

Alth

ough

bot

h el

emen

ts h

ave

a hi

gh a

ffin

ity f

or o

rgan

ic m

atte

r in

soi

l, V

is

unde

rsto

od t

o be

mor

e m

obile

in

alka

line

cond

ition

s, u

nlik

e N

i (U

SEPA

, 20

05;

���

Env

iron

men

t A

genc

y, 2

009)

. A

s sh

own

by F

ig.

3, t

otal

V i

n so

ils i

s le

ss c

onst

rain

ed b

y ba

salts

acr

oss

the

stud

y ar

ea w

hen

com

pare

d to

Ni

or C

r, w

ith

mod

erat

e V

con

cent

ratio

ns o

bser

ved

in s

oils

thr

ough

out

the

wes

tern

hal

f of

the

stu

dy a

rea.

Pu

blis

hed

info

rmat

ion

rega

rdin

g V

bio

acce

ssib

ility

is

limite

d,

alth

ough

fac

tors

lin

ked

to p

eat

wer

e pr

evio

usly

pos

itive

ly c

orre

late

d w

ith V

BA

Fs w

hile

Al

and

Fe o

xide

s w

ere

nega

tivel

y co

rrel

ated

to

V B

AFs

(Pa

lmer

et

al.,

2013

).

The

min

imum

Cr

BA

F w

as o

bser

ved

in t

he p

sam

mite

& s

emi-

pelit

e be

droc

k gr

oup.

T

his

grou

p ho

sts

low

ave

rage

con

cent

ratio

ns o

f to

tal

Cr,

MgO

, A

l 2O

3,

CaO

, M

nO,

and

Fe2O

3 w

hile

hav

ing

high

er a

vera

ge l

evel

s of

SO

3 an

d L

OI

com

pare

d to

the

oth

er b

edro

ck g

roup

s.

The

max

imum

Cr

BA

F oc

curr

ed i

n th

e

basa

lt be

droc

k gr

oup,

whi

ch g

ener

ally

exh

ibits

the

opp

osite

tre

nds

in m

ajor

oxi

de a

bund

ance

and

LO

I w

hen

com

pare

d to

the

psa

mm

ite

& s

emi-

pelit

e gr

oup.

Cox

et a

l. (2

013)

fou

nd p

ortio

ns o

f to

tal C

r w

ere

host

ed b

y ir

on o

xide

s in

bas

alt a

nd id

entif

ied

stro

ng p

ositi

ve c

orre

latio

ns b

etw

een

bioa

cces

sibl

e C

r an

d to

tal

Cr,

MgO

, A

l 2O

3, P

2O5,

CaO

, M

nO,

and

Fe2O

3, f

acto

rs w

hich

wer

e al

so a

ssoc

iate

d w

ith t

he b

asal

t co

mpo

nent

as

iden

tifie

d by

PC

1.

Hill

et

al.

(200

1) f

ound

that

Cr

in t

he A

ntri

m B

asal

ts w

as a

ssoc

iate

d w

ith s

tabl

e ir

on o

xide

s an

d as

a r

esul

t w

as e

nric

hed

duri

ng w

eath

erin

g, s

ugge

stin

g th

at C

r is

not

mob

ile a

nd

ther

efor

e un

likel

y to

be

bioa

cces

sibl

e. S

imila

r tr

ends

in C

r ha

ve b

een

note

d pr

evio

usly

with

res

pect

to it

s af

fini

ty to

Fe

and

Al o

xide

s an

d cl

ay m

iner

als

in th

e

soil

envi

ronm

ent (

Rai

et a

l., 1

989;

Fre

i and

Pol

at, 2

013)

. T

his

is in

con

tras

t to

Ni w

hose

min

eral

ass

ocia

tions

are

in a

mor

e bi

oacc

essi

ble

form

, an

obse

rvat

ion

whi

ch m

ay n

ot b

e lim

ited

to s

oils

ove

rlyi

ng b

asal

ts b

ut w

hich

may

als

o ap

ply

to N

i in

soils

acr

oss

Nor

ther

n Ir

elan

d ba

sed

on P

CA

fin

ding

s.

5.0

Con

clus

ion

Stro

ng g

eoge

nic

cont

rols

in

Nor

ther

n Ir

elan

d in

flue

nce

the

spat

ial

dist

ribu

tion

and

conc

entr

atio

ns o

f m

ajor

and

tra

ce e

lem

ents

in

soils

. H

owev

er,

the

oral

bioa

cces

sibi

lity

of N

i, V

and

Cr

is i

nflu

ence

d by

com

plex

soi

l-ge

oche

mis

try

inte

ract

ions

tha

t ar

e no

t fu

lly a

ccou

nted

for

by

unde

rlyi

ng g

eolo

gy o

r to

tal

trac

e

elem

ent

conc

entr

atio

ns.

Hig

her

Ni

bioa

cces

sibi

lity

was

obs

erve

d in

soi

ls o

ver

psam

mite

& s

emi-

pelit

e an

d lim

esto

ne r

ock

type

s, c

oncl

uded

to

be l

inke

d

spec

ific

ally

to

the

pres

ence

of

acid

ic s

oil

cond

ition

s, o

rgan

ic m

atte

r, a

nd c

arbo

nife

rous

sed

imen

tary

geo

logy

. H

igh

frac

tions

of

bioa

cces

sibl

e V

wer

e

obse

rved

in

soils

ove

rlyi

ng b

asal

t an

d sa

ndst

one

bedr

ock,

alth

ough

dis

tinct

con

trol

s fr

om u

nder

lyin

g be

droc

k ov

er V

bio

acce

ssib

ility

wer

e no

t cl

earl

y

���

disc

erni

ble.

B

oth

tota

l an

d bi

oacc

essi

ble

Cr

conc

entr

atio

ns s

how

ed t

he s

tron

gest

ass

ocia

tions

to

geol

ogic

fac

tors

, ex

plai

ning

in

part

its

low

bio

acce

ssib

le

frac

tions

but

hig

h bi

oacc

essi

ble

conc

entr

atio

ns.

Ack

now

ledg

emen

ts a

nd D

iscl

aim

er

2013

UB

M t

estin

g w

as s

uppo

rted

by

the

EU

IN

TE

RR

EG

IV

A-f

unde

d T

ellu

s B

orde

r pr

ojec

t. T

he v

iew

s an

d op

inio

ns e

xpre

ssed

in

this

rep

ort

do n

ot

nece

ssar

ily r

efle

ct t

hose

of

the

Eur

opea

n C

omm

issi

on o

r th

e SE

UPB

.��T

ellu

s B

orde

r is

a €

5 m

illio

n cr

oss-

bord

er p

roje

ct t

o m

ap t

he e

nvir

onm

ent

and

natu

ral

reso

urce

s in

the

bor

der

regi

on o

f Ir

elan

d an

d co

ntin

ue t

he a

naly

sis

of d

ata

in t

he b

orde

r co

untie

s of

Nor

ther

n Ir

elan

d.

It i

s a

join

t in

itiat

ive

betw

een

the

Geo

logi

cal

Surv

ey o

f N

orth

ern

Irel

and,

the

Geo

logi

cal

Surv

ey o

f Ir

elan

d, Q

ueen

’s U

nive

rsity

, B

elfa

st a

nd D

unda

lk I

nstit

ute

of T

echn

olog

y.

2009

UB

M

test

ing

was

fun

ded

by th

e B

GS

Uni

vers

ity F

undi

ng I

nitia

tive.

The

Nor

ther

n Ir

elan

d T

ellu

s pr

ojec

t w

as f

unde

d by

the

Nor

ther

n Ir

elan

d D

epar

tmen

t of

Ent

erpr

ise,

Tra

de a

nd I

nves

tmen

t an

d by

the

Rur

al D

evel

opm

ent

Prog

ram

me

thro

ugh

the

Nor

ther

n Ir

elan

d Pr

ogra

mm

e fo

r B

uild

ing

Sust

aina

ble

Pros

peri

ty.

Spec

ial t

hank

s to

sta

ff a

t the

Ana

lytic

al G

eoch

emis

try

Faci

lity

at th

e B

GS

Kin

gsle

y D

unha

m C

entr

e, e

spec

ially

Mr.

Tom

Bar

low

and

Mr.

Elli

ott H

amilt

on f

or

thei

r te

chni

cal a

nd a

naly

tical

sup

port

. T

hank

s al

so to

Mar

k C

ave

and

Jo W

ragg

for

thei

r va

luab

le a

ssis

tanc

e w

ith 2

013

UB

M d

ata

QA

.

Ref

eren

ces

� Abo

llino

O.,

Mal

adri

no M

., G

iaco

min

o A

., M

enta

sti

E.

(201

1) T

he r

ole

of c

hem

omet

rics

in

sing

le a

nd s

eque

ntia

l ex

trac

tion

assa

ys:

A r

evie

w P

art

I.

Ext

ract

ion

proc

edur

es, u

ni-

and

biva

riat

e te

chni

ques

and

mul

tivar

iate

var

iabl

e re

duct

ion

tech

niqu

es f

or p

atte

rn r

ecog

nitio

n. A

nal.

Chi

m. A

cta

688,

104

-121

. A

llow

ay B

. J.

(20

05)

Bio

avai

labi

lity

of e

lem

ents

in

soil.

In

Ess

enti

als

of M

edic

al G

eolo

gy,

Els

evie

r A

cade

mic

Pre

ss,

Lon

don.

S

elin

us,

Allo

way

, C

ente

no,

Fink

elm

an, F

uge,

Lin

dh, S

med

ley

(eds

.).

BA

RG

E/I

NE

RIS

(2

011)

U

BM

pr

oced

ure

for

the

mea

sure

men

t of

in

orga

nic

cont

amin

ant

bioa

cces

sibi

lity

from

so

lid

mat

rice

s.

A

vaila

ble

at

http

://w

ww

.bgs

.ac.

uk/B

AR

GE

/ubm

.htm

l. B

arsb

y A

., M

cKin

ley

J.M

., O

fter

ding

er U

., Y

oung

M.,

Cav

e M

.R.,

Wra

gg J

. (20

12)

Bio

acce

ssib

ility

of

trac

e el

emen

ts in

soi

ls in

Nor

ther

n Ir

elan

d. S

ci. T

otal

E

nvir

on. 4

33, 3

98-4

17.

��

Cab

oche

J.

(200

9)

Val

idat

ion

d'un

tes

t de

mes

ure

de b

ioac

cess

ibili

té.

App

licat

ion

à qu

atre

élé

men

ts t

race

s m

étal

lique

dan

s le

s so

ls:

As,

Cd,

Pb

et S

b.

Scie

nce

Agr

onom

ique

. PhD

. L'In

stitu

t Nat

iona

l Pol

ytec

hniq

ue d

e L

orra

ine,

Nan

cy, p

p. 3

48.

C

ampb

ell

N.

A.

(199

6a)

Prok

aryo

tes

and

the

orig

ins

of m

etab

olic

div

ersi

ty.

In

Bio

logy

4th e

d.

Ben

jam

in/C

umm

ings

Pub

lishi

ng C

ompa

ny,

New

Yor

k. p

p.

498-

517.

C

ampb

ell,

N. A

. (19

96b)

Wat

er a

nd th

e fi

tnes

s of

the

envi

ronm

ent.

In

Bio

logy

4th e

d. B

enja

min

/Cum

min

gs P

ublis

hing

Com

pany

, New

Yor

k. p

p. 4

1-52

. C

ande

ias

C.,

Ferr

eira

da

Silv

a E

., Sa

lgue

iro

A.

R.,

Pere

ira

H.

G.,

Rei

s A

.P.,

Patin

ha C

., M

atos

J.X

., A

vila

P.H

. (2

011)

The

use

of

mul

tivar

iate

sta

tistic

al

anal

ysis

of

geoc

hem

ical

dat

a fo

r as

sess

ing

the

spat

ial

dist

ribu

tion

of s

oil

cont

amin

atio

n by

pot

entia

lly t

oxic

ele

men

ts i

n th

e A

ljust

rel

min

ing

area

. E

nvir

on.

Ear

th S

ci. 6

2, 1

461-

1479

. C

ox S

. F.,

Che

lliah

M.,

McK

inle

y J.

M.,

Palm

er S

., O

fter

ding

er U

., C

ave

M. R

., W

ragg

J.,

You

ng M

. (20

13)

The

impo

rtan

ce o

f so

lid-p

hase

dis

trib

utio

n on

the

oral

bio

acce

ssib

ility

of

Ni a

nd C

r in

soi

ls o

verl

ying

Pal

aeog

ene

basa

lt la

vas,

Nor

ther

n Ir

elan

d. E

nvir

on. G

eoch

em. H

lth

35(5

), 5

53-5

67.

The

Cha

rter

ed I

nstit

ute

of E

nvir

onm

enta

l H

ealth

(C

IEH

) (2

009)

Pro

fess

iona

l pr

actic

e no

te:

revi

ewin

g hu

man

hea

lth r

isk

asse

ssm

ent

repo

rts

invo

king

co

ntam

inan

t ora

l bio

avai

labi

lity

mea

sure

men

ts o

r es

timat

es.

Ava

ilabl

e at

ww

w.c

ieh.

org/

polic

y/re

view

ing_

hum

an_h

ealth

_ris

k.ht

ml.

Upd

ated

201

3.

Den

ys S

., C

aboc

he J

., T

ack

K.,

Ryc

hen

G.,

Wra

gg J

., C

ave

M.,

Jond

revi

lle C

., Fe

idt C

. (20

12)

In v

ivo

valid

atio

n of

the

unif

ied

BA

RG

E m

etho

d to

ass

ess

the

bioa

cces

sibi

lity

of a

rsen

ic, a

ntim

ony,

cad

miu

m a

nd le

ad in

soi

ls.

Env

iron

. Sci

. Tec

h. 4

6, 6

252-

6260

. D

epar

tmen

t fo

r E

nvir

onm

ent,

Food

and

Rur

al A

ffai

rs (

DE

FRA

) (2

012)

Env

iron

men

tal

Pro

tect

ion

Act

199

0: P

art

2A.

Con

tam

inat

ed L

and

Stat

utor

y G

uida

nce.

HM

Gov

ernm

ent,

Her

Maj

esty

’s S

tatio

nery

Off

ice.

D

ober

man

n A

., G

ooav

erts

P.,

Geo

rge

T. (

1995

) So

urce

s of

soi

l var

iatio

n in

an

acid

Ulti

sol o

f th

e Ph

ilipp

ines

. G

eode

rma

68, 1

73-1

91.

Ein

ax J

.W.,

Sold

t U

. (1

999)

Geo

stat

istic

al a

nd m

ultiv

aria

te s

tatis

tical

met

hods

for

the

ass

essm

ent

of p

ollu

ted

soils

—m

erits

and

lim

itatio

ns.

Che

mom

etr.

In

tell

. Lab

. 46,

79-

91.

Env

iron

men

t Age

ncy

(200

9) S

oil g

uide

line

val

ues

for

nick

el in

soi

l. S

cien

ce R

epor

t SC

0500

21/ N

icke

l SG

V.

Env

iron

men

tal

Stan

dard

s R

esea

rch

Inst

itute

(E

SRI)

(20

10)

Arc

GIS

, A

rcM

ap a

nd A

rcC

atal

ogue

10.

0 ge

ogra

phic

al i

nfor

mat

ion

syst

ems

soft

war

e an

d he

lp

file

s.

��

Frei

R.,

Pola

t A

. (20

13)

Chr

omiu

m i

soto

pe f

ract

iona

tion

duri

ng o

xida

tive

wea

ther

ing—

impl

icat

ions

fro

m t

he s

tudy

of

a Pa

leop

rote

rozo

ic p

aleo

sol,

Schr

eibe

r B

each

, Ont

ario

, Can

ada.

Pre

cam

bria

n R

es. 2

24, 4

34-4

53.

Gia

com

ino

A,

Abo

llino

O.,

Mal

andr

ino

M.,

Men

tast

i E

. (2

011)

The

rol

e of

che

mom

etri

cs i

n si

ngle

and

seq

uent

ial

extr

actio

n as

says

: A

rev

iew

. P

art

II.

Clu

ster

ana

lysi

s, m

ultip

le li

near

reg

ress

ion,

mix

ture

res

olut

ion,

exp

erim

enta

l des

ign

and

othe

r te

chni

ques

. A

nal.

Chi

m. A

cta

688,

122

-139

. G

ooav

erts

P. (

1992

) Fa

ctor

ial k

rigi

ng a

naly

sis:

a u

sefu

l too

l for

exp

lori

ng th

e st

ruct

ure

of m

ultiv

aria

te s

patia

l soi

l inf

orm

atio

n. J

. Soi

l Sci

. 43,

597

-619

. H

ill I

. G.,

Wor

den

R. H

., M

eigh

an I

. G. (

2001

) Fo

rmat

ion

of in

terb

asal

tic la

teri

te h

oriz

ons

in N

E I

rela

nd b

y ea

rly

Ter

tiary

wea

ther

ing

proc

esse

s. P

. Geo

logi

st

Ass

oc. 1

12 (4

), 3

39–3

48.

Jord

an C

., H

iggi

ns A

., H

amill

K.,

Cru

icks

hank

J.

G.

(200

1) T

he S

oil

Geo

chem

ical

Atl

as o

f N

orth

ern

Irel

and.

D

epar

tmen

t of

Agr

icul

ture

and

Rur

al

Dev

elop

men

t, N

orth

ern

Irel

and.

Jo

rdan

C.,

Zha

ng C

., H

iggi

ns A

. (2

007)

Usi

ng G

IS a

nd s

tatis

tics

to s

tudy

inf

luen

ces

of g

eolo

gy o

n pr

obab

ility

fea

ture

s of

sur

face

soi

l ge

oche

mis

try

in

Nor

ther

n Ir

elan

d. J

. Geo

chem

. Exp

lor.

93,

135

-152

. K

onen

M. E

., Ja

cobs

P. M

, Bur

ras

C. L

., T

alag

a B

.J.,

Mas

on J

. A. (

2002

) E

quat

ions

for

pre

dict

ing

soil

orga

nic

carb

on u

sing

loss

-on-

igni

tion

for

Nor

th C

entr

al

U.S

. Soi

ls.

Soil

Sci

. Soc

. Am

. J. 6

6(6)

, 187

8-18

81.

Lyl

e P.

(19

80).

A p

etro

logi

cal a

nd g

eoch

emic

al s

tudy

of

the

tert

iary

bas

altic

roc

ks o

f no

rthe

ast I

rela

nd.

J. E

arth

Sci

. 2, 1

37-1

52.

Meu

nier

L.,

Wal

ker

S. R

, W

ragg

J.,

Pars

ons

M.

B.,

Koc

h I.

, Ja

mie

son

H.

E.,

Rei

mer

K.

J. (

2010

) E

ffec

ts o

f so

il co

mpo

sitio

n an

d m

iner

alog

y on

the

bi

oacc

essi

bilit

y of

ars

enic

fro

m ta

iling

s an

d so

il in

gol

d m

ine

dist

rict

s of

Nov

a Sc

otia

. E

nvir

on. S

ci. T

ech.

44,

266

7-26

74.

Mitc

hell

W. I

. (20

04)

The

Geo

logy

of N

orth

ern

Irel

and:

Our

Nat

ural

Fou

ndat

ion.

Geo

logi

cal S

urve

y of

Nor

ther

n Ir

elan

d, A

ntri

m.

Nat

hana

il P.

, McC

affr

ey C

., A

shm

ore

M.,

Che

ng Y

., G

illet

t A

., O

gden

R.,

Scot

t D

. (20

09)

The

LQ

M/C

IEH

Gen

eric

Ass

essm

ent

Cri

teri

a fo

r H

uman

Hea

lth

Ris

k A

sses

smen

t (2nd

ed.

). L

and

Qua

lity

Pres

s, N

ottin

gham

. N

atur

al E

nvir

onm

ent

Res

earc

h C

ounc

il (N

ER

C)

Bio

acce

ssib

ility

tes

ting

of c

onta

min

ated

lan

d fo

r th

reat

s to

hum

an h

ealth

. E

cono

mic

val

uatio

n re

port

pr

epar

ed f

or N

ER

C b

y D

TZ

Con

sulta

ncy.

Ava

ilabl

e at

ww

w.n

erc.

ac.u

k/bu

sine

ss/c

ases

tudi

es/r

epor

ts.a

sp.

Page

upd

ated

28

May

201

2.

Neb

el B

. J.,

Wri

ght

R. T

. (20

00)

Atm

osph

eric

Pol

lutio

n.

In E

nvir

onm

enta

l Sc

ienc

e: t

he w

ay t

he w

orld

wor

ks (

7th e

d.).

Pr

entic

e-H

all,

Inc.

, New

Jer

sey.

pp

. 52

5-55

2.

���

Oks

anen

J.,

Bla

nche

t F.

G.,

Kin

dt R

., L

egen

dre

P.,

Min

chin

P.

R.,

O'H

ara

R.

B.,

Sim

pson

G.

L.,

Soly

mos

P.,

Stev

ens

M.,

Wag

ner

H.

(201

3) v

egan

: C

omm

unity

Eco

logy

Pac

kage

. R p

acka

ge v

ersi

on 2

.0-1

0. h

ttp://

CR

AN

.R-p

roje

ct.o

rg/p

acka

ge=

vega

n.

Oom

en A

. G

., R

ompl

eber

g C

., B

ruil

M.,

Dob

be C

., Pe

rebo

om D

., Si

ps A

(20

03)

Dev

elop

men

t of

an

in v

itro

dige

stio

n m

odel

for

est

imat

ing

the

bioa

cces

sibi

lity

of s

oil c

onta

min

ants

. A

rch.

Env

iron

. Con

. Tox

. 44,

281

-287

. Pa

lmer

S.,

Oft

erdi

nger

U.,

McK

inle

y J.

M.,

Cox

S.,

Bar

sby

A. (

2013

) C

orre

latio

n an

alys

is a

s a

tool

to in

vest

igat

e th

e bi

oacc

essi

bilit

y of

nic

kel,

vana

dium

and

zi

nc in

Nor

ther

n Ir

elan

d so

ils.

Env

iron

Geo

chem

. Hlt

h. 3

5(5)

, 569

-584

. Po

ggio

L.,

Vrščaj

B.,

Schu

lin R

., H

eppe

rle

E.,

Mar

san

F. A

. (2

009)

Met

als

pollu

tion

and

hum

an b

ioac

cess

ibili

ty o

f to

psoi

ls i

n G

rugl

iasc

o (I

taly

).

Env

iron

. P

oll.

157,

680

-689

.

R C

ore

Tea

m (

2013

) R

: A la

ngua

ge a

nd e

nvir

onm

ent f

or s

tatis

tical

com

putin

g. R

Fou

ndat

ion

for

Stat

istic

al C

ompu

ting,

Vie

nna,

Aus

tria

. UR

L h

ttp://

ww

w.R

-pr

ojec

t.org

/. R

ai D

. L.,

Ear

y L

. E.,

Zac

hara

J. M

. (19

89)

Env

iron

men

tal c

hem

istr

y of

chr

omiu

m.

Sci.

Tot

al E

nvir

on. 8

6, 1

5-23

. R

eim

ann

C.,

Filz

mos

er P

., G

arre

tt R

., D

utte

r R

. (2

009)

Sta

tist

ical

Dat

a A

naly

sis

Exp

lain

ed:

App

lied

Env

iron

men

tal

Stat

isti

cs w

ith

R.

Wile

y-B

lack

wel

l, C

hich

este

r.

Rub

y M

. V. D

avis

A.,

Scho

of R

., E

berl

e S.

, Sel

lsto

ne C

. M. (

1996

) E

stim

atio

n of

Lea

d an

d A

rsen

ic B

ioav

aila

bilit

y U

sing

a P

hysi

olog

ical

ly B

ased

Ext

ract

ion

Tes

t. E

nvir

on. S

ci. T

echn

ol. 3

0(2)

, 422

-430

. R

uby

M. V

., Sc

hoof

R.,

Bra

ttin

W.,

Gol

dade

M.,

Post

G.,

Har

nois

M.,

Mos

by D

. E.,

Cas

teel

S. W

., B

erti

W.,

Car

pent

er M

., E

dwar

ds D

., C

ragi

n D

., C

happ

ell

W. (

1999

) A

dvan

ces

in E

valu

atin

g th

e O

ral

Bio

avai

labi

lity

of I

norg

anic

s in

Soi

l fo

r U

se i

n H

uman

Hea

lth R

isk

Ass

essm

ent.

Env

iron

. Sci

. Tec

hnol

. 33

(21)

, 36

97-3

706.

Sa

lehi

M.

H.;

Ben

i O

. H

., H

arch

egan

i H

. B

., B

oruj

eni

I. E

., M

otag

hian

H.

R.

(201

1) R

efin

ing

soil

orga

nic

mat

ter

dete

rmin

atio

n by

los

s-on

-ign

ition

. P

edos

pher

e 21

(4),

473

-482

. Sm

yth,

D. (

2007

) M

etho

ds u

sed

in th

e T

ellu

s ge

oche

mic

al m

appi

ng o

f Nor

ther

n Ir

elan

d. B

ritis

h G

eolo

gica

l Sur

vey

Ope

n R

epor

t OR

/07/

022,

200

7.

Spar

ks D

. L. (

1995

) E

nvir

onm

enta

l Soi

l Che

mis

try.

Aca

dem

ic P

ress

, Inc

., N

ew Y

ork.

���

Stew

art M

. A.,

Jard

ine

P. M

., B

rand

t C. C

., B

arne

tt M

. O.,

Fend

orf

S. E

., M

cKay

L. D

, Meh

lhor

n T

. L.,

Paul

K. (

2003

) E

ffec

ts o

f co

ntam

inan

t con

cent

ratio

n,

agin

g, a

nd s

oil p

rope

rtie

s on

the

bioa

cces

sibi

lity

of C

r(II

I) a

nd C

r(V

I) in

soi

l. S

oil S

ed. C

onta

min

atio

n. 1

2(1)

, 1-2

1.

Suba

cz J

., B

arne

tt M

., Ja

rdin

e P.

, Ste

war

t M. (

2007

) D

ecre

asin

g ar

seni

c bi

oacc

essi

bilit

y/bi

oava

ilabi

lity

in s

oils

with

iron

am

endm

ents

. J.

Env

iron

. Sci

. Hea

l. A

42,

131

7-13

29.

Tar

buck

E.J

., L

utge

ns F

.K. (

1999

). E

arth

: A

n In

trod

ucti

on to

Phy

sica

l Geo

logy

. Pr

entic

e H

all,

Upp

er S

addl

e R

iver

, New

Jer

sey.

T

empl

M.,

Filz

mos

er P

., H

ron

K. (

2009

). r

obC

ompo

sitio

ns: R

obus

t Est

imat

ion

for

Com

posi

tiona

l Dat

a. R

Pac

kage

ver

sion

1.3

.3.

The

ng, B

. (19

74)

The

Che

mis

try

of C

lay

Org

anic

Rea

ctio

ns.

Hal

sted

Pre

ss, N

ew Y

ork.

U

nite

d St

ates

E

nvir

onm

enta

l Pr

otec

tion

Age

ncy

(USE

PA)

(200

5)

Eco

logi

cal

Soil

Scre

enin

g le

vels

fo

r V

anad

ium

.

OSW

ER

D

irec

tive

9285

.7-7

5,

Was

hing

ton,

D.C

. V

an D

e W

eile

T.,

Oom

en A

., W

ragg

J.,

Cav

e M

., M

inek

us M

., H

ack

A.,

Cor

nelis

C.,

Rom

pleb

erg

C.,

De

Zw

art L

., K

linck

B.,

Van

Wijn

en J

., V

erst

raet

e W

., Si

ps A

. (2

007)

Com

pari

son

of f

ive

in v

itro

dig

estio

n m

odel

s to

in

vivo

exp

erim

enta

l re

sults

: L

ead

bioa

cces

sibi

lity

in t

he h

uman

gas

troi

ntes

tinal

tra

ct.

J.

Env

iron

. Sci

. Hea

l. A

42,

120

3-12

11.

Wed

epho

l, K

. H. (

1978

) H

andb

ook

of G

eoch

emis

try.

Ber

lin: S

prin

g-V

erla

g.

Wils

on, H

. E. (

1972

) R

egio

nal G

eolo

gy o

f Nor

ther

n Ir

elan

d. G

eolo

gica

l Sur

vey

of N

orth

ern

Irel

and;

Her

Maj

esty

’s S

tatio

nery

Off

ice,

Bel

fast

. W

ragg

, J. (

2009

) B

GS

guid

ance

mat

eria

l 102

, iro

nsto

ne s

oil,

cert

ific

ate

of a

naly

sis.

Bri

tish

Geo

logi

cal S

urve

y.

Wra

gg, J

., C

ave

M. (

2003

) In

-vitr

o M

etho

ds f

or t

he M

easu

rem

ent

of t

he O

ral

Bio

acce

ssib

ilit

y of

Sel

ecte

d M

etal

s an

d M

etal

loid

s in

Soi

ls:

A C

riti

cal

Rev

iew

. B

GS

R&

D T

echn

ical

Rep

ort P

5-06

2/T

R/0

1.

Wra

gg J

., C

ave

M.,

Bas

ta N

., B

rand

on E

., C

aste

el S

., D

enys

S.,

Gro

n C

., O

omen

A.,

Rei

mer

K.,

Tac

k K

., V

an d

e W

iele

T. (

2011

) A

n in

ter-

labo

rato

ry tr

ial o

f th

e un

ifie

d B

AR

GE

bio

acce

ssib

ility

met

hod

for

arse

nic,

cad

miu

m a

nd le

ad in

soi

l. S

ci. T

otal

Env

iron

. 409

, 401

6-40

30.

Wra

gg J

., C

ave

M.,

Tay

lor

H.,

Bas

ta N

., B

rand

on E

., C

aste

el S

., G

ron

C.,

Oom

en A

., V

an d

e W

iele

T.

(200

9) I

nter

-lab

orat

ory

tria

l of

a u

nifi

ed

bioa

cces

sibi

lity

proc

edur

e. B

ritis

h G

eolo

gica

l Sur

vey

Che

mic

al &

Bio

logi

cal H

azar

ds P

rogr

amm

e, O

pen

Rep

ort O

R/0

7/02

7.

Zha

ng C

., Fa

y D

., M

cGra

th D

., G

renn

an E

., C

arto

n O

. (20

07)

Stat

istic

al a

naly

ses

of g

eoch

emic

al v

aria

bles

of

soils

in I

rela

nd.

Geo

derm

a, 1

46, 3

78-3

90.

���

Figu

res

Fi

g. 1

Sim

plif

ied

geol

ogic

map

of

Nor

ther

n Ir

elan

d an

d U

BM

sam

ple

loca

tions

, n=

145.

Sim

plif

ied

clas

sifi

catio

ns b

ased

on

Smyt

h (2

007)

. N

orth

ern

Irel

and

imag

e us

ed w

ith p

erm

issi

on f

rom

GSN

I, C

row

n C

opyr

ight

.

���

��

Fig.

2 R

ange

of

Ni,

V a

nd C

r to

tal s

oil c

once

ntra

tions

(m

g kg

-1)

acro

ss e

leve

n ge

neri

c be

droc

k gr

oups

pre

sent

in N

orth

ern

Irel

and.

All

thre

e el

emen

ts h

ave

��

high

er c

once

ntra

tions

in b

asal

ts, a

lthou

gh V

dis

play

s a

grea

ter

degr

ee o

f va

riab

ility

in it

s so

il co

ncen

trat

ions

acr

oss

the

stud

y ar

ea.

��

���

��

Fig.

3 T

otal

Ni,

V a

nd C

r so

il co

ncen

trat

ions

acr

oss

Nor

ther

n Ir

elan

d in

terp

olat

ed b

y or

dina

ry k

rigi

ng.

Inte

rmed

iate

con

cent

ratio

ns o

f N

i, V

and

Cr

can

be

��

obse

rved

in s

ands

tone

s an

d lit

hic

aren

ite in

the

sout

heas

t whi

le r

elat

ivel

y lo

w c

once

ntra

tions

are

illu

stra

ted

in u

plan

d lo

catio

ns r

elat

ing

to m

ount

aino

us g

rani

te

�

and

psam

mite

& s

emi-

pelit

e, p

artic

ular

ly f

or C

r an

d V

.

�

���

��

Fig.

4 B

ox p

lots

for

six

maj

or e

lem

ent o

xide

s ex

pres

sed

as p

erce

nt to

tal o

xyge

n ac

ross

ele

ven

bedr

ock

grou

ps.

Irre

gula

rity

in th

e st

atis

tical

dis

trib

utio

ns o

f ��

som

e va

riab

les

in s

oils

ove

rlyi

ng a

cid

volc

anic

s or

psa

mm

ite &

sem

i-pe

lite

may

sug

gest

the

pres

ence

of

sour

ces

of s

oil-

geoc

hem

ical

var

iabi

lity

not

���

dist

ingu

ishe

d by

the

pres

ent g

roup

ing

of d

ata

by g

ener

ic b

edro

ck ty

pe.

���

���

���

���

Fig.

5 B

ox p

lots

of

MnO

and

Fe 2

O3

as p

erce

nt to

tal o

xyge

n, p

H a

nd p

erce

nt lo

ss o

n ig

nitio

n (L

OI)

with

y-a

xis

limite

d fo

r M

nO to

per

mit

illus

trat

ion

of

���

stat

istic

al d

istr

ibut

ion.

Low

er a

vera

ge p

H in

psa

mm

ite &

sem

i-pe

lite

and

acid

vol

cani

c gr

oups

is th

e re

sult

of u

plan

d pe

at s

oils

ove

r th

ese

bedr

ock

type

s.

���

Irre

gula

rity

in L

OI

dist

ribu

tions

als

o in

thes

e tw

o be

droc

k gr

oups

may

sug

gest

the

LO

I m

etho

d is

det

ectin

g m

ultip

le s

ourc

es o

f co

mbu

stio

n an

d w

ater

loss

��

from

bot

h pe

at a

nd c

lay

soil

min

eral

s (S

aleh

i et a

l., 2

011)

.��

Fig. 6 Kriged major oxide distributions across Northern Ireland. MgO, CaO, and Fe2O3 are distinctly controlled by basalts in the northeast and most oxides with the exception of SO3 display relative deficiency in mountainous upland peat soils. Al2O3 and SiO2 show high concentrations in sandstone, lithic arenite, mudstones and limestones, whilst P2O5 is the only oxide whose distribution shows little apparent relationship to underlying geology.

���

Fig. 7 Kriged soil pH and % loss on ignition across Northern Ireland. LOI gives an approximation of soil carbon content as illustrated by high values corresponding with spatial patterns of upland peat distributions, also related to acidic soil pH conditions as shown left.

Fig 8. Boxplot distributions for gastric bioaccessible fractions of Ni, V and Cr according to nine underlying bedrock groups captured by combined UBM study set.

Figgeofact

Figgeobasass �

g. 9 ochetors

g. 10ochesalt ocia

Eigemis wh

0 Memibedated

genistryhile

Mapical drocd pr

nvaluy dae ob

pedvar

ck, Predo

ue fata stuse

d sorianPC2omin

factset (e an

il sance f2 is nan

tor l(a) ngle

ampfromlink

ntly

loadand

es il

ple m thkedwit

dingd in llust

sitehirted to th a

g ploUBtrate

e scoeen soilpH

�

ots BM e a

oresgeo

l facH inf

for studneg

s froochctorflue

pridy sgativ

om emir vaence

ncipset (ve r

PCical

ariane.

pal (b).relat

A. Fl varnce

com Ation

Firsriaband

mpocutenshi

st thblesd re

onene veip.

hrees. Pemai

nts ecto

e coPC1 inin

idenor an

ompis a

ng v

ntifngle

oneassovaria

fied es im

entsociaanc

in fmpl

accatede lin

fullly a

cou witnke

Noa cor

nt fth v

ed to

orthrrel

for 8variao PC

ern latio

87%ancC3 i

Ireon b

% ofce fris

elandbetw

f totrom

d ween

tal m

n

���

Tables Table 1. pH tolerances for prepared digestive fluids. Solution pH Tolerance Solution pH Tolerance

Saliva 6.5 +/- 0.5 Duodenal 7.4 +/- 0.2

Gastric 1.0 +/- 0.1 Bile 8.0 +/- 0.2

Table 2. Summary statistics for trace elements (mg kg-1), major element oxides (%), pH and LOI, n = 6,862 (Palmer et al., 2013).

Ni V Cr MgO Al2O3 SiO2 P2O5 SO3 CaO MnO Fe2O3 pH LOI

Min 1.40 5.90 4.10 0.50 3.50 13.8 0.05 0.00 0.30 0.00 0.30 2.83 1.01

Max 334 402 1229 5.80 17.2 87.9 1.70 2.00 16.3 15.0 42.2 7.68 98.0

Mean 46.2 99.6 131 1.45 10.6 49.6 0.26 0.18 1.15 0.08 4.65 4.74 23.4

Median 29.1 85.0 94.1 1.30 11.4 54.1 0.25 0.10 0.85 0.06 4.19 4.82 12.9

StdDev 48.7 65.0 121 0.66 2.98 15.0 0.11 0.24 0.78 0.26 2.85 0.74 25.2

Table 3. Geostatistical summary of trace elements, oxides, pH and LOI. Cx = maximum variance accounted for by spatial function, C0 = nugget variance not explicable within defined range of spatial function, a = maximum range of spatial dependence between sample locations in meters, Total Cx = C0 + C1 + C2. Large C0: Total Cx ratios suggest micro-scale processes may influence spatial distributions (Einax and Soldt, 1999). Variable C0 C1 C2 a1 a2 Total Cx C0/Total Cx (%)

Ni 262.1 756.3 1885 9038 76,595 2903 9.0

V 697.8 1257 3014 13,816 74,194 4969 14.0

Cr 2223 11,604 3064 86,790 7038 16,891 13.2

MgO 0.02 0.12 0.02 1029 1872 0.16 12.5

Al2O3 3.46 3.40 1.86 45,354 6720 8.72 39.7

SiO2 30.57 81.98 132.4 80,301 9927 245.0 12.5

P2O5 0.006 0.004 -- 17,233 -- 0.01 60.0

SO3 0.04 0.02 -- 36,438 -- 0.06 67.7

CaO 0.32 0.51 -- 177,902 -- 0.83 39.5

MnO 0.06 0.03 -- 31,404 -- 0.09 67.7

Fe2O3 2.71 6.86 -- 96,708 -- 9.57 28.3

pH 0.27 0.09 0.29 176,782 21,738 0.65 41.5

LOI 155.9 105.1 379.6 33,006 7886 640.6 24.3

���

Tab

le 4

. Ni,

V a

nd C

r ga

stri

c bi

oacc

essi

bilit

y (B

AF

= %

of

tota

l con

cent

ratio

n; m

g kg

-1 =

abs

olut

e m

easu

red

bioa

cces

sibl

e co

ncen

trat

ion)

acr

oss

stud

y ar

ea

��

and

by e

ight

gen

eric

bed

rock

type

s ac

cord

ing

to c

ombi

ned

2013

and

200

9 U

BM

ext

ract

ion

data

(B

arsb

y et

al.,

201

2).

Bol

d te

xt in

dica

tes

occu

rren

ce o

f ��

max

imum

and

min

imum

val

ues

from

ful

l stu

dy s

et.

Gab

bro

excl

uded

due

to s

mal

l sam

ple

size

. ��

Fu

ll St

udy

Set

Bas

alt n

= 2

9 Sa

ndst

one

n =

16

Psam

mite

& S

emi-

Pelit

e n

= 2

2 C

ongl

omer

ate

n =

6

BA

F M

ean

Med

M

ax

Min

M

ean

M

ed

Max

M

in

Mea

n M

ed

Max

M

in

Mea

n M

ed

Max

M

in

Mea

n M

ed

Max

M

in

Ni

12.4

15

.2

46.3

1.

4 12

.1

8.2

43.8

1.

4 15

.2

15.2

24

.4

5.6

12.8

8.

2 46

.3

4.7

7.6

7.5

9.9

5.5

V

8.3

7.1

23.1

1.

4 11

.3

10.3

22

.5

3.8

9.8

8.6

23.1

5.

5 7.

3 6.

3 15

.5

1.4

7.1

6.6

10.6

4.

7 C

r 1.

2 1.

0 5.

4 0.

03

1.5

1.4

5.4

0.4

0.85

1.

0 1.

3 0.

05

1.1

1.1

2.7

0.03

1.

3 1.

3 1.

8 0.

8

Gra

nite

n =

7

Lim

esto

ne n

= 1

9 M

udst

one

n =

23

Lith

ic A

reni

te n

= 1

6 Fu

ll St

udy

Set m

g kg

-1

BA

F

Mea

n M

ed

Max

M

in

Mea

n

Med

M

ax

Min

M

ean

Med

M

ax

Min

M

ean

Med

M

ax

Min

M

ean

Med

M

ax

Min

N

i 13

.2

9.3

28.3

3.

0 14

.4

12.9

35

.8

3.6

13.3

10

.3

33.6

4.

1 9.

3 6.

6 42

.2

3.0

3.8

2.4

30.7

0.

3 V

8.

8 7.

8 13

.1

6.7

6.0

5.5

14.1

1.

9 8.

4 6.

7 22

.0

2.5

6.5

5.0

16.7

2.

6 7.

4 5.

1 52

.1

0.6

Cr

1.7

1.7

2.6

0.7

1.1

1.0

2.1

0.6

0.9

0.9

1.6

0.2

1.3

1.2

2.8

0.7

1.5

1.0

10.0

0.00

4

��

��

�

�

��

��

���

���

���

Table 5. Eigenvalues greater than 1 explaining variance for component loadings. Component Value % Variance Cumulative % Value % Variance Cumulative %

All Northern Ireland, n = 6862 UBM Study Set, n = 145

1 5.87 45 45 6.06 34 34 2 3.13 24 69 5.10 28 623 1.34 10 79 1.55 9 714 -- -- -- 1.46 8 79

Table 6. Eigenvalue factor loadings for identified principal components in Table 5. MnO not identified as significant variable within UBM study set. Bold values indicate factor is most strongly weighted to indicated component, underlined values show where a large amount of remaining factor variance is accounted for by a secondary component.

Variable Component

1 2 3 1 2 3 4

All NI, n=6862 UBM Study Set, n=145 V 0.742 -0.534 0.114 0.807 -0.342 0.224 -0.116

Ni 0.631 -0.688 0.138 0.750 -0.478 0.019 -0.360

Cr 0.770 -0.399 0.136 0.857 -0.224 -0.087 -0.175

MgO -0.511 -0.680 0.010 -0.299 -0.685 0.004 -0.426

Al2O3 -0.843 -0.200 -0.340 -0.689 -0.497 -0.386 -0.064

SiO2 -0.747 0.174 -0.521 -0.646 -0.201 -0.662 0.121

P2O5 -0.620 -0.436 -0.362 -0.388 -0.617 -0.430 0.045

SO3 -0.875 -0.395 0.155 -0.649 -0.712 0.143 0.093

CaO -0.385 -0.713 0.090 -0.097 -0.780 0.129 -0.018

MnO 0.565 -0.422 -0.512 -- -- -- --

Fe2O3 0.401 -0.670 -0.249 0.443 -0.435 -0.049 -0.330

pH 0.631 0.038 -0.523 0.555 0.036 -0.531 -0.092

LOI -0.787 -0.438 0.358 -0.556 -0.644 0.436 0.089

G V -- -- -- 0.348 -0.634 0.280 0.512

GI V --� -- -- 0.397 -0.404 -0.127 0.765 G Ni --� -- -- 0.240 -0.866 0.043 -0.016

GI Ni --� -- -- 0.433 -0.719 -0.193 -0.077

G Cr --� -- -- 0.855 0.100 -0.162 0.120

GI Cr --� -- -- 0.691 -0.001 -0.219 0.283

���

��������

• Two soil-geochemical factors were identified accounting for 69% of total variance

• Low chromium bioaccessible fractions were linked to basalt bedrock factors

• Bioaccessible nickel was associated with peat and calcareous soil components

• The median vanadium bioaccessible fraction was 7%; chromium 1% and nickel 15%