Analysis of Background PAH levels in East Anglia

Assessment of Anthropogenic Background PAH
Concentrations in East Anglia Soils

Aamer Raza
Associate Director, Harrison Group
Environmental
Director, MPA Global Solutions Ltd

MPA – Global Solutions
www.mpaglobalsolutions.com
G‐CAT

What is Background?
Wh t i B k d?
Background concentrations are generally considered to be the level or amount of
a chemical f
h i l found i common areas not associated with a particular contaminant
d in t i t d ith ti l t i t
release.

Parks, roadways, residential areas, non‐industrialized properties are typical
, y, , p p yp
background areas.

The definition from Tiered Approach to Corrective Action Objectives (TACO
(Illinois) is:
(Illinois) is:

“Area Background” means concentrations of regulated substances that are
consistently present in the environment in the vicinity of a site that are the result
of natural conditions or human activities, and not the result solely of releases at
the site.
Information summarized from Urban Area Polycyclic Aromatic Hydrocarbons Study Tiered Approach to Corrective Action
Objectives, Illinois EPA

G‐CAT

Motivating Factors for the Study
M ti ti F t f th St d
ALARA Principle – Remediation to target levels < background is not practicable,
cost efficient, or b
t ffi i t beneficial f
fi i l from a h lth protection standpoint;
health t ti t d i t

Delineating limits of contamination from point sources;

Screening out sites from Preliminary Investigations relative to more detailed
studies;

Understanding local conditions – will be highly variable

Because we can – large amounts of data are available, and can be retrospectively
analysed, as long as amenable to electronic data reading and manipulation (AGS
data format proved to be the key)

G‐CAT

Study Objectives
St d Obj ti
Assess anthropogenic background levels of PAH levels detected in
East Anglia Soils.

•Rural ( id tifi d current or hi t i source)
R l (no identified t historic )
•Urban (no identified current or historic source)
•Industrial
•Farms
•Automotive works

G‐CAT

Design
Retrospective examination of SI data between 2004 and 2010

•Non‐structured selection of sites – SI
N t t d l ti f it SIs conducted by Harrison Group with
d t db H i G ith
archived AGS electronic data

Cl ifi ti f it l d b d D k St d i f ti
•Classification of sites as per land uses based on Desk Study information

•Statistical review of PAH data
Distributional characteristics
Distributional characteristics
Outlier identification
Development of baseline data set
Derivation of statistical parameters
p
Comparison with published values

G‐CAT

Design

Group Sites Samples
Farm 2 13
Industrial 15 104
Urban 7 28
General 16 69
Automotive 10 45
Total 50 259

G‐CAT

Sites in Urban Areas

Group Sites Samples
Farm 2 13
Industrial 15 104
Urban 7 28
General 16 69
Automotive 10 45
Total 50 259

G‐CAT

Statistical Analysis and Screening

Statistical analysis :
Statistical analysis :

•Normal probability plots
T t f lit (Sh i Wilk )
•Tests for normality (Shapiro Wilkes)
•Test for outliers (Grubbs Test)
•Upper Confidence Limits (Normal, Log‐Normal, Non‐Parametric)
•95% Upper Tolerance Limits
•95% Upper Tolerance Limits
•Multi‐Variate Analysis (Fingerprinting)

Spatial Analysis
Spatial Analysis
•Lateral and Vertical extent
•Hot‐spots
•Concentration Gradients

G‐CAT

Summary Statistics ‐ General (no identified current or historic source)
Summary Statistics General Data
y

Log-Normal

Geometric 95th 95% Upper
Constituent Min Max Mean Mean 95% UCL Percentile Tolerance Limit
Acenaphthene 14 760 53 28 56 132 197
Acenaphthylene 5 890 84 20 114 258 492
Anthracene 9 1700 169 35 222 484 948
Benzo (a) anthracene 12 8100 686 117 1141 2200 4656
Benzo (a) pyrene 12 8700 768 97 1564 2518 5781
Benzo (b) fluoranthene 16 11000 930 128 1794 3049 6846
Benzo (ghi) perylene 10 6700 603 87 1164 2073 4662
Benzo (k) fluoranthene 25 7400 534 97 652 1403 2776
Chrysene 10 7000 716 107 1532 2581 5818
Dibenzo (ah) anthracene 8 1500 129 26 150 336 643
Fluoranthene 25 21000 1529 198 2859 4803 10839
Fluorene 12 1300 79 31 87 213 348
Indeno (1.2.3 - cd) pyrene 11 5900 467 63 746 1350 2947
Naphthalene 10 5100 253 75 362 839 1554
Phenanthrene 21 8500 722 140 1147 2337 4795
Pyrene 22 17000 1240 172 2248 3918 8702

G‐CAT

Statistical Distribution ‐ General (no identified current or historic source)
Normal Probability Plot BaP ‐ General Land Use
Full Data Set
Full Data Set
10

8

6
Ln Value

4

2

0
3
‐3 ‐2
2 ‐1
1 0 1 2 3
‐2
Normal Quantile
20
18
16
14
•High end outliers
•High end outliers 12
•Left censored (detection limits) 10 Frequency

•Bi‐modal 8 Fitted Curve
6
4
2
0
0.2 1.4 2.6 3.8 5 6.2 7.4 8.6 9.8

G‐CAT

Statistical Distribution ‐ General (no identified current or historic source)
Normal Probability Plot BaP ‐ General Land Use
y
Minus High End Outliers
9
8
7
6
5
Ln Value

4
3
2
1
0 BaP in General Soils
Normal Quantile
‐3 ‐2 ‐1 0 1 20 2 3
F 18
r 16
e 14
•Closer conformity with Log
•Closer conformity with Log‐ q 12
normal Distribution u 10
8 Frequency
e
•Left censored (detection limits) n 6 Fitted Curve
c 4
y 2
0
0.2 0.8 1.4 2 2.6 3.2 3.8 4.4 5 5.6 6.2 6.8 7.4 8 8.6 9.2 9.8
Ln Conc

G‐CAT

Dealing with Non‐Detects

•Detection Limit (approach used)
Detection Frequency 75%

½ i i i
•½ Detection Limit

•Maximum Likelihood Estimation for Left Censored Data ‐
replacement of non‐detect values with fitted values
•To be assessed

G‐CAT

Pattern Assessment – Variation in Isomer ratios (assessment of discrete sources)
•Principal Components Analysis of Isomer Ratios
•Data divided by concentration – assess if outliers are from unique sources
5

4

3

2

1
PCA2
0
‐4 ‐2
‐1
0 2 4 6 8 10 12 No clear difference by
Concentration level.
‐2
2

‐3 Most samples showed
‐4 very similar profiles
PCA1

<1 1‐10 >10 Suggestive of common
source

G‐CAT

Statistical Distribution ‐ Industrial
Normal Probability Plot BaP ‐ Industrial Land Use
Normal Probability Plot BaP Industrial Land Use
Full Data Set
8
7
6
5
Ln Value

4
3
2
1
0
Normal Quantile
‐3 ‐2.5 ‐2 ‐1.5 ‐1 ‐0.5 0 8
0.5 1
F 7
r 6
•Good fit with Log normal
•Good fit with Log‐normal e
5
q
Distribution u 4
Frequency
•Left censored (detection limits) e
n
3
Fitted Curve
2
c
y 1
0
0.2 1.4 2.6 3.8 5 6.2 7.4 8.6 9.8
Ln Conc

G‐CAT

Data Comparison Across Groups

Log-Normal

Geometric 95th 95% Upper
Constituent Min Max Mean Mean 95% UCL Percentile Tolerance Limit

Urban 12 4700 498 129 1447 1933 5214

General 12 8700 768 97 1564 2518 5781

Industrial 12 220000 6049 508 13414 19717 50176

Automotive 12 16840 1276 222 2913 4414 9950

Farm 26 14000 2008 561 33072 11559 76123

G‐CAT

Comparison with Published Literature

Background Levels of Polycyclic Aromatic Hydrocarbons (PAH) and Selected Metals
Background Levels of Polycyclic Aromatic Hydrocarbons (PAH) and Selected Metals
in New England Urban Soils
L. J. N. Bradley B. H. Magee and S. L. Allen Journal of Soil Contamination, 3(4): (1994)

•Samples of surficial soils from urban locations in three New England cities. Sixty
samples – twenty from each city

Concentrations of Polynuclear Aromatic Hydrocarbons and Inorganic Constituents in
Concentrations of Polynuclear Aromatic Hydrocarbons and Inorganic Constituents in
Ambient Surface Soils – Chicago Illinois, 2001‐2002
Kay et al, 2003, USGS Report 03‐1045

57 Sit i Chi M t lit A O l f h it
•57 Sites in Chicago Metropolitan Area. One sample from each site

PAHs in background soils from Western Europe: Influence of atmospheric
deposition and soil organic matter
p g
Jae Jak Nam et al. Chemosphere 70 (2008) 1596–1602

•53 Soil Samples from Norway and UK

G‐CAT

Comparison with Published Literature

Broadly, our results are similar to the observations made elsewhere:
Broadly our results are similar to the observations made elsewhere:

UK soils – Geometric Mean BaP 46 ug/kg – range 1.8 – 1600; (our data Geo. Mean
52 range 6‐1724, after high end outliers were excluded)

Chicago soils – BaP 95 Percentile 2,100 ug/kg; (our data 2,500 including outliers,
1,700 excluding)

Highly spatially variable – orders of magnitude differences within 300 metres of each
other.

New England Study – BaP range 40 – 13 000 /k
N E l d St d B P 40 13,000 ug/kg, and 95 percentile 1,800 ((our
d 95 til 1 800 ((
data 2,500 including outliers, 1,700 excluding)

G‐CAT

Implication for Remediation

Approximate Screening Value
Number of Fitzgerald et
Samples Analysed Detects Minimum Maximum Mean 95% UCL HPA BMD10 GAC al.
500 1000 5000
5 5 680 8500 4514 8068.7
8068 7
5 2 12 66 26 47.9
4 3 12 1724 649 1534.6
4 4 22 954 366 883.2
4 3 12 880 275 752.9
4 4 51 390 139 335.9
6 2 12 28 15 20.4
4 2 12 69 36 69.3
5 5 52 8700 2177 5698.6
4 2 12 340 101 289.2
6 5 12 2300 638 1422.9
5 3 12 142 45 98.5
4 4 91 390 200 364.4

D. James Fitzgerald,Neville I. Robinson,and Beverly A. Pester. Application of Benzo(a)pyrene and Coal Tar Tumor Dose–Response
Data to a Modified Benchmark Dose Method of Guideline Development

G‐CAT

Conclusions – Points for Discussion

Stationary and mobile anthropogenic sources account for the majority of PAHs, with
great variability regionally and locally.

Remediation should not commence without establishing / assessing a site‐specific
background, if possible.

The impracticality and infeasibility of remediation to guideline levels if naturally
occurring background exceeds the maximum allowable value is self evident. It deems
remediation unnecessary and unproductive in such scenarios.

G‐CAT

ALARA principal is a vital consideration (setting cleanup levels as low as
reasonably achievable). Assessment and consideration of background
would be imperative.

A tier‐based guideline system may be very beneficial, where the central
tendency and maximum values are compared with relevant background
numbers (95% UCL), (95th percentile). This is particularly relevant in the
current times, given the b d i
i i h broader issues of greenhouse gas emissions, and
f h i i d
need for efficient utilisation of economic resources.

G‐CAT

Analysis of Background PAH levels in East Anglia

More Related Content

Featured (20)

Analysis of Background PAH levels in East Anglia