Overview and experiences of the Swiss soil monitoring...
Transcript of Overview and experiences of the Swiss soil monitoring...
1
Federal Office for the Environment FOEN
Federal Department of Economic Affairs FDEA
Agroscope Reckenholz-Tänikon Research Station ART
Nationale Bodenbeobachtung Observatoire national des sols Osservatorio nazionale dei suoli Swiss Soil Monitoring Network
Overview and experiences of the Swiss soil monitoring network
over 25 years- Focus on forest soils -
3rd Workshop of the NE American Soil Monitoring Network
USGS New York Water Science Centre, Troy, NY
11-12 March, 2009
www.nabo.admin.ch
André [email protected]
2
2Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Acknowledgements
Swiss soil monitoring networkCurrent collaborators:
André Desaules (25)
Peter Schwab (8)
Stefan Ammann (7)
Kirsten Rehbein (6)
Armin Keller (5)
Reto Meuli (1)
(7) = years of activity
Former collaborators:
Silvia Geering (18)
Ruedi Dahinden (14)
Hans Vogel (5)
Edith Meier (11)
Konrad Studer (10)
Thomas Keller (6)
Hans Brunner (6)
Nicolas Rossier (2)
The acknowledgements reveal that soil monitoring is laborious and many
persons and changes are involved. The major challenge in long-term soil
monitoring is thus to cope with changes of relevant boundary conditions as
persons but also methodologies and site conditions.
3
3Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Outline
• The Swiss soil monitoring network (NABO)
• Results and findings:
- Inorganic pollutants: three sampling campaigns- Organic pollutants: status study
• Lessons learnt
• Basic challenges of soil monitoring
• References
4
4Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Objectives and mandate
Monitoring and assessment of the countrywide status and long-term trend of impacts on soils in Switzerland.
Federal law of 7 October 1983 relating to the protection of the environment (art. 44)
Ordinance of 1 July 1998 relating to impacts on the soil (art. 3)
5
5Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Switzerland (1) : Satelite photo
The major landscapes and settlement areas from N to S:
Jura (tabular and folded Jura): approx. 2‘000 to 5‘200 ft
Plateau (lower and upper plateau): 950 to 3‘000 ft
Alps (northern prealps, central alps and southern alps): 4‘000 to 14‘000 ft
Southern Ticino: 800 to 2‘700 ft
Settlement concentrations (dark pink areas at lower elevations): Basel,
Geneva, Lausanne, Bern, Zurich, Lugano,
6
6Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Switzerland (2) : Land use
Forest (31 %)Arable land
Vineyard, orchard, horticulure
Pasture and prairie
15 830 sq mi (41 000 km2)
7
7Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Switzerland (3) : Geology
Arrow: direction of tectonic movement (S to N)
Pink: cristalline basement outcrops (Central alps, Black forest/Vosges)
Yellow and brownish: nappes of cristalline and southern alpine sediments
Light green and blue: northern alpine and jurassic lime stones and marls
Bright brown: alpine sediment basins
8
8Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Switzerland (4): Basic data
• Surface area: 15 830 sq mi (41 000 km2): unproductive 25%, forested 31%, agriculture 37%, settled 7%)
• Elevation: min. 640 ft to max. 15 210 ft- Jura: 2 000 to 5 200 ft- Plateau: 950 to 3 000 ft- Alps: 4 000 to 14 000 ft- Southern Ticino: 800 to 2 700 ft
• Annual temperature: approx. 27 to 52oF (mean 42oF)
• Annual precipitation: 24 to 71 in (mean 43 in)
• Population: 7.6 millions (480 pers/sq mi)
9
9Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
NABO reference network characteristics
• Up to 105 long-term observation sites of all relevant
land-use types
• The current periodicity of sampling campaigns is 5 years: 1985/89, 1990/94, 1995/99, [2000/04, 2005/09]
• Soil pollutants:Inorganics: Cd, Zn, Pb, Cu, Hg, Ni, Cr, Co (2M HNO3)
and F
Organics: PAH, PCB, [PCDD/F]
The work for the Swiss soil monitoring network started in August 1984 and is
still continuing after almost 25 years and several administrative
restructurations. Results will be presented of the first three sampling
campaigns. Currently the 5th sampling campaign is running.
10
10Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
NABO-reference network and modular concept
NABO Quality
NABO Status NABO Flux NABO Trend
(25)
(36)
(27)
(11)
(6)
(48)
The NABO reference network comprises currently 105 long-term monitoring
sites of all characteristic land use types:
Forests: 27 (deciduous 12, coniferous 15)
Conservation sites: 4
Grassland: 25
Arable: 36
Special crops: 11
Urban parks: 2
At 48 agricultural sites fluxes of pollutants are recorded additionally.
The conceptual framework consists of 4 modules:
•NABO-Quality is the foundation of reliability of the results and their
interpretation.
-NABO-Status records soil pollution levels and assesses them over space.
-NABO-Flux records pollutant fluxes as indicators for soil pollution
forecasts (indirect monitoring).
-NABO-Trend measures and assesses changes of soil pollution over time
(direct monitoring) and is the focus of this presentation.
11
11Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
51 Wartauspecial crops:
vegetables
Sampling design
soil depth: 0-20 cm
Special crop NABO-site no 51 Wartau (inner alpine Rhine valley 1 500 ft)
The sampling design and support consists of 4 parallel stratified composite
samples of 25 increments each from a sampling site of 10 by 10 m.
The reference sampling depth is 20 cm from the surface.
12
12Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Forest NABO-site no. 47 Davos
1999
Site: central alps 5 430 ft (less than 1 mi from the Davos World Economic
Forum)
Sampling auger: diameter 1.18 in, organic-mineral soil core 7.87 (20 cm)
13
13Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
NABO-site no. 58 Mels
‚Vivian‘ stormFeb. 1990
1988 (1)
1993 (2)
Site: Northern prealps 3 000 ft
Windthrow: The only one out of 27 forest sites which was affected.
14
14Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
NABO-site no. 42 Galmwald
‚Lothar‘storm
Dec. 1999
1987 (1) 2003 (4)
Site: Plateau 1 900 ft
Windthrow: 3 out of 27 forest sites were affected.
15
15Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Inorganic pollutants:Three sampling campaigns 1985-1999
16
16Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Frequent concentration levels andguide values (depth 0-20 cm, n = 102)
Element Richtwert*
[mg/kg]
Häufige Gehalte
(10.-90. Perzentil)
[mg/kg]
Richtwert-
überschreitungen
[n]
Cd 0.8 0.11-0.49 5
Zn 200
150
35-89 0
0
Pb 50 16-38 6
Cu 50
40
6-35 6
10
Hg 0.8
0.5
0.06-0.19 0
0
Ni 50 6-40 5
Cr 50 13-38 1
Co 25 3-10 1
F 400
700
234-715 57
13
*VSBo 1986 , VBBo 1998
Guide value* Frequent conc. Guide value
(10-90th perc.) exceedance
(Desaules et al.1993)
The frequent concentration levels of the 1st campaign (1985/89) are below
the respective guide values - except for F.
The revision of the guide values in 1998 (normal characters) partly changed
the number of exceeded guide values – especially for F.
The guide values are legally based soil quality benchmarks, considered as
precautionary general upper limits of negligible risks.
17
17Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Data Quality (1): Site precision (n = 4)
0% 2% 4% 6% 8% 10% 12% 14% 16%
Cd
Zn
Pb
Cu
Hg
Ni
Cr
Co
F
Measurement campaigns 1 to 3 from top to bottom
CV
(Desaules et al.2006)
Site precision def.: repeatability conditions for sampling, sample preparation
and chemical analysis of 4 replicate composite soil samples
Note: Site precision varies with elements and with time!
18
18Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Data Quality (2): Analytical stability
Deviation of 1st campaign%
2nd campaign
3nd campaign
Abw eichung in %
-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60
Cd
Zn
Pb
Cu
Hg
Ni
Cr
Co
F
Deviation from the 1st campaign in %(Desaules et al. 2006)
Analytical stability (chemical preparation and analysis of the same sample
after 5 years) is just one component of measurement stability it excludes
sampling and physical sample preparation stability
Note: There is always an inherent analytical variation in analytical stability
which is generally not constant over time (see esp. Hg and Co).
19
19Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Data Quality (3): Analytical bias
� � � � �� � � � � � � � � � �� � � � � � �-60 -40 -20 0 20 40 60
Cd
Zn
Pb
Cu
Hg
Ni
Cr
Co
F
%
Deviations of „consensus values“ from round robin analyses (Desaules et al. 2006)
There is nothing like a constant analytical bias (systematic error) in chemical
analysis as often believed and it may change over time.
The 3 presented slides of the quality parameters (site precision,
analytical stability and analytical bias) give evidence of temporal
fluctuations and the necessity of a continuous quantitative control of the
data quality.
20
20Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Results of the 1st campaign 1985/89: Pb in top- and subsoil
(Desaules et al. 1993)
The topsoil levels of Pb exceed with few exceptions the subsoil levels, which
usually indicates soil pollution. The figures with bars should prevent spatial
extrapolation, but it appears that in the NE of Switzerland and S of the Alps
exceeding guide values seem to be more frequent.
Note: for reasons of relative and ecotoxic comparability the further
presented concentrations and their changes are normalized in percent of
the respective guide values set as 100 %.
21
21Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Pb and Cu (0-20 cm) by land use
Bleigehalte
0
50
100
150
200
250
300
MW
To
tal 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
102
103
MW
A 5
55
96
101
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
100
104
105
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkultur int. Grasland ext. Grasland Schutz-
standort
Laubwald Nadelwald Stadt-
park
% R
W
Kupfergehalte
0
50
100
150
200
250
300
MW
Tota
l 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
10
2
10
3
MW
A 5
55
96
10
1
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
10
0
10
4
10
5
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkul tur int. Gras land ext. Gras land Schutz-
s tandort
Laubwald Nadelwald Stadt-
park
% R
W
1. Erhebung
(RW = 50 mg/kg)
(RW = 40 mg/kg)
Pb (guide value 50 mg/kg)
Cu (guide value 40 mg/kg)
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urban
land sites parks
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urban
land sites parks
The concentration levels of Pb and Cu for all sites reveal that the variablity
within land use type (white bars) generally exceeds variability between land
use types (grey bars) and puts the common classification by land use into
question. Striking exceptions are the high Cu levels of the 4 vineyard sites
(223-860 mg/kg) and to some extent the Pb concentrations in the two urban
parks. Forest sites show generally greater variations but similar Pb
concentrations and a tendency of lower Cu concentrations
22
22Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Findings from the 1st campaign 1985/89
• All soils of Switzerland – even in remote areas – are to some extent polluted but the level is generally low.
• The rank order of general soil pollution relative to the guide values is Pb > Cu > Cd for the 9 investigated elements.
• The guide values are often exceeded for F > Ni > Cr but also for Cd due to natural causes as parent material concentrations.
• Except for Cu in vineyards land use is not a reliable indicator for soil pollution.
23
23Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Results of the 2nd campaign 1990/94:Changes of Pb and Cu after 5 years
Bleigehalte
0
50
100
150
200
250
300
MW
Tota
l 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
10
2
10
3
MW
A 5
55
96
10
1
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
10
0
10
4
10
5
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkultur int. Gras land ext. Grasland Schutz-
standort
Laubwald Nadelwald Stadt-
park
% R
W
Kupfergehalte
0
50
100
150
200
250
300
MW
To
tal 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
10
2
10
3
MW
A 5
55
96
10
1
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
10
0
10
4
10
5
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkultur int. Gras land ext. Grasland Schutz-
standort
Laubwald Nadelwald Stadt-
park
% R
W
(RW = 50 mg/kg)
2. Erhebung
Pb (guide value 50 mg/kg)
Cu (guide value 40 mg/kg)
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urbanland sites parks
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urbanland sites parks
after 5 years
The concentration changes of Pb and Cu after 5 years (red symbols) show,
that positive as well as negative changes occur which are sometimes
significant. However, the mean changes (grey bars) are hardly significant.
No clear differences of changes are apperent on forest sites.
24
24Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Changes for all elements after 5 years
a all sites 0-20 cm (n = 100)b no till sites 0-20 cm (n = 54)c like b but 0-5 cm
(0.8) = guide value in mg/kg
� � � � � � � � � � � � � (Desaules et al. 2000)
a) For all sites 0-20 cm: The changes are mostly below 5% of the respective
guide values and quite symmetric.
b) For no till sites 0-20 cm: No greater changes due to no tillage could be
observed.
c) For No till sites 0-5 cm: No greater positive changes due to less dilution
effect of inputs could be observed either. Hypothesis: Sampling at
shallow soil depth is not robust.
25
25Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Changes at selected sites
� after 5 years, soil depht 0-5 cm � after 5 years, soil depht 0-20 cm � after 10 years, soil depht 0-20 cm
black symbols: significant variations
(error probebility α=2.9%) ----------------- limit of relevance
(± 5% of the guide value)
Mixed farming
Coniferous forest
(Desaules et al. 2000)
Quite different patterns of changes for all 9 investigated elements can be
observed for the same types of land use. This is an indication that further
factors influence soil concentration changes.
26
26Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Findings from the 2nd campaign 1990/94
• The changes were as well positive as negative.
• After 5 years the majority of the sites (87 %)
showed already at least one significant change for one of the 9 investigated elements.
• The changes are therefore not only a result of
inputs. There must be other causes.
27
27Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Results of the 3rd campaign 1995/99:Changes of Pb and Cu after 5 and 10 years
Bleigehalte
0
50
100
150
200
250
300
MW
Tota
l 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
102
103
MW
A 5
55
96
101
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
100
104
105
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkultur int. Gras land ext. Grasland Schutz-
s tandort
Laubwald Nadelwald Stadt-
park
% R
W
Kupfergehalte
0
50
100
150
200
250
300
MW
Tota
l 3 9
11
13
14
15
17
21
23
25
26
28
29
31
36
38
39
44
46
48
53
54
56
63
65
68
77
78
79
80
86
87
95
102
103
MW
A 5
55
96
101
MW
R 4
20
64
MW
O 51
59
67
94
MW
G 1
30
33
35
37
60
69
74
MW
i.G 6
10
16
32
34
41
49
50
52
57
70
71
72
98
100
104
105
MW
e.G 12
75
84
89
MW
S 8
19
24
27
42
43
62
66
82
85
92
93
MW
L 2 7
18
22
40
45
47
58
73
76
81
83
88
90
91
99
MW
N 61
97
MW
P
Ackerbau Spezialkultur int. Gras land ext. Gras land Schutz-
standort
Laubwald Nadelwald Stadt-
park
% R
W
(RW = 50 mg/kg)
(RW = 40 mg/kg)
2. Erhebung 3. Erhebung
Pb (guide value 50 mg/kg)
Cu (guide value 40 mg/kg)
after 5 after 10 years
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urban
land sites parks
arable land special crops int. grass- extensive grassland cons. dec. forest conif. forest urbanland sites parks
The changes after 10 years (blue symbols) are not always greater than after 5
years (red symbols) as expected. To analyse this, the bar figures are no more
appropriate.
28
28Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Changes of Cu concentrations after 5 and 10 years by land use
Veränderung Cu-Gehalte (RW = 40 mg/kg)
n= 97 96 15 15 8 8 34 33 7 7 2 2 4 4 12 12 15 15
-50
-40
-30
-20
-10
0
10
20
30
40
50
% R
W
5 10 5 10 5 10 5 10 5 10 5 10 5 10 5 10 5 10
Alle extensives Grasland
intensives Grasland Ackerbau
Spezial- kultur
Stadt- park
Schutz- standort
Laub- Wald
Nadel- Wald
Cu guide value (RW) = 40 mg/kg
yearsall extensive intensive arable special urban parks conservation deciduous coniferous
grassland grassland land crops sites forest forest
(Desaules et al. 2006)
•At first sight the important zigzag changes with a great variability for Cu
under special crops (4 vineyards) are striking. The reasons must be closer
investigated (sampling, deep tillage, stage migration)
•Note further the general mean tendency of increase under grassland and
deciduous forest. For the intensive grassland the reason is mainly attributed to
Cu additives in the fodder for porcs.
•For all sites the Cu changes almost neutralise themselves. Sense to express
mean changes?
No clear differences of Cu changes are apperent on forest sites
29
29Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Changes at an arable site after 5 and 10 years
95 Coldrerio TI: arable mixed farming - cattleCd Zn Pb Cu Hg Ni Cr Co F
0
20
40
60
80
100
120
% R
W
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120(25)(0.8) (150) (700)(50) (0.5) (50) (50)(40)
5 %
(0.8) = guide value (RW) in mg/kg significant change(Desaules et al. 2006)
The overview of changes for all 9 investigated elements at one site gives an
example of their different levels relative to the guide values (e.g. Cu and Hg)
and their different evolutions of change.
Zigzag evolutions are common as it is characteristic for environmental time
series (e.g. global temperature).
Only 5 changes were significant and/or relevant (>5% of guide value) which
are indicated by red circles.
30
30Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Changes at a forest site after 5 and 10 yearsaffected by windthrow in 1990
Concentrations in % of guide value (RW); samplings :(1) 1988, (2) 1993, (3) 1998
58 Mels / Rütiwald SG, Wald - Nadelwald dominant
Cd Zn Pb Cu Hg Ni Cr Co F
0
20
40
60
80
100
120
% R
W
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
1990
Spruce dominant initially
(0.8) (150) (50) (40) (0.5) (50) (25) (700)(50)
(Desaules et al. 2006)
Surprisingly no effect of windthrow was observed.
31
31Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Findings from the 3rd campaign 1995/99
• Zigzag evolutions of changes are common.
• Therefore trend statements after 3 campaigns in 10 years are not yet possible.
• The dynamics of changes can vary with time.
• The causes of measured changes are complex (real and artefacts).
• Varying site numbers (n) influence mean changes (Will Rogers phenomenon or stage migration).
32
32Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Input models are inadequate for soilconcentration changes
concentration
timet1 t2
input
View to the frequently measured zigzag evolutions, (black box) input models
are inadequate for soil monitoring.
33
33Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Organic pollutants:Status study
34
34Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Data Quality:Estimation of the analytical and measurement
uncertaintyStoff Konzentrationen
1 Vergleichs-
Analysen-
unsicherheiten
(VAU)1
Vergleichs-Messunsicherheit
(mg/kg)
(2 RSD)
sehr ähnliche
Bodenbedingungen
(2 x VAU)
sehr unterschiedliche
Bodenbedingungen
(5 x VAU)
PAH16 Minimum
10. Perzentil
Median
Richtwert2
Maximum
0.032
0.072
0.163
1
8.46
±54 %
±48 %
±44 %
±41 %
±37 %
±108 %
±96 %
±88 %
±82 %
±74 %
±270 %
±240 %
±220 %
±205 %
±185 %
BaP Minimum
10. Perzentil
Median
Richtwert2
Maximum
0.0005
0.003
0.013
0.02
1.13
±100 %
±77 %
±45 %
±37 %
±33 %
±200 %
±154 %
±90 %
±74 %
±66 %
±500 %
±385 %
±225 %
±185 %
±165 %
PCB7 Minimum
10. Perzentil
Median
Maximum
0.005
0.008
0.0016
0.012
±100 %
±94 %
±48 %
±46 %
±200 %
±188 %
±96 %
±92 %
±500 %
±470 %
±240 %
±230 %
Analyticaluncertainty
(AU)
Compound ConcentrationsMeasurement uncertainty
Soil conditions:
similar different(2 AU) (5 AU)
The extended analytical uncertainty (p = approx. 95%) is much concentration
dependent and reaches partly 100% at the minimum concentrations.
Since no data for the whole measurement process including sampling are
available, its estimate is based here on the analytical uncertainty and soil
conditions as an estimate of the sample materialisation error.
35
35Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
PAH-concentrations (top 20 cm)
(Desaules et al. 2008)
PAH16 concentrations only twice exceed the respectivew guide value of 1
mg/kg. There is a tendency of higher levels under forests and settlement areas
36
36Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
BaP-concentrations (top 20 cm)
(Desaules et al. 2008)
The same holds for benzo(a)pyrene
37
37Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
PCB-concentrations (top 20 cm)
(Desaules et al. 2008)
The measured PCB7-concentrations are all much below the quality
benchmark of 0.02 mg/kg. Again a slight tendency to higher values under
forest are observed which may, however, be attributed to the lower bulk
density of forest topsoils.
38
38Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
PAH- and PCB- concentrations by land use
(a) PAH16
all
(n=9
9)
ext. g
rassla
nd (
n=1
7)
int. g
rassla
nd (
n=
8)
ara
ble
land (
n=3
3)
vitic
ulture
(n=
4)
hort
iculture
(n=
4)
orc
hard
(n=
3)
conserv
atio
n s
ites (
n=
4)
conife
rous fore
st
(n=1
6)
decid
uous fore
st
(n=1
0)
0
200
400
600
800
! " # # #1200
2600
2800
g/k
g
(b) BaP
all
(n=9
9)
ext. g
rassla
nd (
n=1
7)
int. g
rassla
nd (
n=
8)
ara
ble
land (
n=3
3)
vitic
ulture
(n=
4)
hort
iculture
(n=
4)
orc
hard
(n=
3)
conserv
atio
n s
ites (
n=
4)
conife
rous fore
st
(n=1
6)
decid
uous fore
st
(n=1
0)
0
50
100
150
$ % & ' '250
300
(C) PCB7
Median
25%-75%
10%-90%
Outliers
Extremes
all
(n=10
5)
ext. g
rassla
nd (
n=1
7)
int. g
rassla
nd (
n=
8)
ara
ble
land (
n=3
5)
vitic
ulture
(n=
4)
hort
iculture
(n=
4)
orc
hard
(n=
3)
urb
an p
ark
(n=
2)
conserv
atio
n s
ites (
n=
4)
conife
rous fore
st
(n=1
6)
decid
uous fore
st
(n=1
2)
0
DL 2
4
QL 6
8
10
12
14
(Desaules et al. 2008)
Compared to all sites and the bulk of other land use types the forest sites
show no distinct differences. Only the PCB7 concentrations seem a little
higher. And relative to deciduous forest, the concentrations are generally
higher under coniferous forest.
39
39Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
PAH- und PCB-profiles
(a) PAH16 (b) PCB7
Median
25%-75%
10%-90% NA
P
AC
Y
AC
E
FLU
PH
E
AN
T
FL
T
PY
R
Ba
A
CH
R
Bb
F
Bk
F
Ba
P
IPY
DB
A
BP
E
PC
B 2
8
PC
B 5
2
PC
B 1
01
PC
B 1
18
PC
B 1
38
PC
B 1
53
PC
B 1
80
0
5
10
15
20
25
30
35
40
45
[%]
(Desaules et al. 2008)
Over all sites the prevailing compounds are phenanthrene and PCB 135 with
little variations between sites.
40
40Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Correlations: PAH16/BaP a)
0
200
400
600
800
1000
1200
0 20 40 60 80 100
BaP [µg/kg]
PA
H16 [
µg/k
g]
r = 0.88** (n = 99)
(Desaules et al. 2008)
BaP is a fairly good indicator for PAH16.
41
41Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Correlations: PCB7/PAH16
all
int. grassland
coniferous forest
conservation sites
ext. grassland
viticulture
arable land
h)
0
2
4
6
8
10
12
0 200 400 600 800 1000 1200
PAH16 [µg/kg]
PC
B7 [
µg
/kg
]
(n=99) r=0.46**
(n=8)
(n=16) r=0.85**
(n=4) r=0.97*
(n=16)
(n=4)
(n=33)
(Desaules et al. 2008)
The correlation among PCB7 and PAH16 is promising on coniferous and
perhaps conservations sites.
42
42Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Correlations: PAH/altitude
all
int. grassland
coniferous forest
conservation sites
ext. grassland
viticulture
arable land
(n=99) r=-0.31**
(n=8)
(n=16) r=-0.68**
(n=4)
(n=17) r=-0.62**
(n=4)
(n=33)
0
200
400
600
800
1000
1200
0 500 1000 1500 2000 2500
altitude [m]
PA
H16 [
µg/k
g]
(Desaules et al. 2008)
The positive correlation of PAH16 with increasing altitude found in literature
could not be confirmed.
43
43Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Correlations: PCB/altitudeg)
0
2
4
6
8
10
12
0 500 1000 1500 2000 2500
altitude [m]
PC
B7 [
µg/k
g]
r = -0.15 (n = 102)
(Desaules et al. 2008)
The situation is still worse for PCB7
44
44Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Methodological comparability of the NABO-study with other studies
(Computed from Desaules et al. 2008)
PAH-studies (1989-2008)Studies Congeners Sample depth
(cm)
Laboratory
1 NABO 16 20 ART
2 (16) ?
3 7 0-6/6-12
4 16 10
5 16 5
6 16 10 ART
7 14 8
8 14 3-10
9 27 5
10 (16) ?
11 16 3
12 20 10
13 21 A-horizon
14 12 5/20
15 18 5
16 17 15.24
17 16 10
18 15 10
19 6 10
20 14 5
Comparable: 8/20 1/20 2/20
PCB-studies (1994-2008)Studies Congeners Sample depth
(cm)
Laboratory
1 NABO 7 20 ART
2 (7) ?
3 51 20
4 7 10
5 17 A-horizon
6 33 8
7 6 Ap-horizon
8 6 5
9 7 3-10
10 7 3
11 6 humus layer
12 6 5
13 14 10
14 7 10 EMPAComparable: 6/14 2/14 1/14
There is a flagrant lack of harmonization among published studies, thus
comparisons of concentration values are mostly ill founded.
45
45Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Findings: organic compounds
1. The PAH- and PCB concentrations of the NABO-network
are low compared to the precautionary guide values.
2. Relatively high concentrations are partly found in soils
under special cultures, urban areas, and surprisingly
conservation sites.
3. The low degree of methodological harmonization hampers
the comparability of different studies.
4. View to considerable measurement uncertainties and lowcorrelations interpretations are difficult.
5. Only a correct and meaningful methodologicalharmonization of the whole measurement chain from
sampling to chemical analysis allows improvements of
comparability and interpretation.
46
46Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Lessons learntand
challenges of soil monitoring
47
47Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
1) Site specific interpretation
0
A
B
So far we can summarise the consequences of our findings by 6 lessons:
Lesson 1: Methodology, evaluation and interpretation of soil monitoring are
so far still too strongly influenced by agronomic field trials.
Each soil monitoring site is an individual case, the sites cannot be treated like
a field trial with the same or similar boundary conditions. Pooling of sites
with different boundary conditions is a considerable source of artefacts.
Therefore the „relevant“ boundary conditions of the individual soil
montioring sites must be much more carefully recorded and considered.
48
48Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
2) Increasing measuring periodicity
year
Mean values
Trend after 15 years
Trend after 55 years
95% confidence interval
Lesson 2: Trends can be identified and certified only after sufficiently
intense and long measurement series. Measurements within the noise cannot
be interpreted. With increasing number of measurements and accurracy noise
can be reduced and trends earlier detected. This is the foundation of pleading
for increasing measurement periodicity in soil monitoring as well.
49
49
Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
U sum = USampling + Upretreatment + Uanalytical
(total?) - grid effect - splitting - repeatability w.s.
- sample effect - crushing - repeatability b.s.
- sieving - moisture effect
- accuracy
- soil condition(short time variability)
- drying - aliquote effect
- seasons effect(long time variability)
- handselection
- sample stability
0
2
4
6
8
10
12[%]
3) Optimising measurement uncertaintyCV
Lesson 3: One measure to reduce the noise of trends is to increase
measurement accuracy by opitimising measurement uncertainty.
The given example of an uncertainty budget for Zn of a grassland soil under
repeatability conditions shows optimising potential in sampling, splitting and
analytical repeatability between series.
50
50Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
4) Increase process understanding
Beobachtung
Verstehen
Modellierung
Prozess-wirkungen
Erfassung und
Interpretation von
Veränderungen
Risikobeurteilung in Bezug auf die
Veränderungen
Experimente Dauer-beobachtung
Prozess-wirkungs-Modelle
Regionale Simulationen
(Kräuchi 1995, modified)
observation
understanding
modelling
experiments monitoring
process-effectmodels
regionalsimulation
process effects
assessment andinterpretation ofchanges
risk evaluation ofchanges
Lesson 4: Measured changes of concentrations can only be correctly
assessed and evaluated if their processes (causes and effects) are sufficiently
understood. Soil monitoring is therefore a basic element in ecosystems
research and understanding.
51
51Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
5) Control changes of boundary conditions
5
(mg/kg)
55
(mg/kg)
96
(mg/kg)
101
(mg/kg)
106
(mg/kg)
t1 433 314 244 654 520 nd
t2 301 313 223 456 nd 213
Veränderungen -132 -1 -21 -198
Probenahme-Effekt
Tiefpflüg-Effekt
Gruppierungs-Effekt (Will-Rogers-Phänomen)
RebbaustandorteMittelwert
n = 4
(mg/kg)
Erhebungs-
zeitpunkt
-307
Example: Temporal changes of measured Cu concentrations at four vineyard sites.
samplingcampaign
changes
vineyard sitesmean
Sampling effect
Tillage effect
Grouping effect (Will Rogers phenomenon)
Lesson 5: The temporal change of relevant boundary conditions as sampling
depth, pedoturbation, analytical conditions, number of sites etc. may affect
the measured soil concentrations or mean values.
The influence of changing number of monitoring sites on the mean value is
known as Will-Rogers phenomenon or stage migration.
Time series in environmental monitoring should therfore be consistent and
rely on the same sites.
52
52Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
EconomicalRisk(losses / costs)
Missed Alarm(false negative
signal)
False Alarm(false positive signal)
Ecological Risk(envrionmentaldamage /
remediation costs)
long-termrisk assessment
6) Quantify and increase reliability of soil monitoring
Lesson 6: View to the enormous possible damages and costs due to false
monitoring signals, the best possible reliable soil monitoring is a very cost
saving public investment in the long-term. The required degree of reliability is
determined by the politically accepted risks.
Environmental monitoring networks which are not able to adequatly assess
their degree of reliability are threfore not fit for taking political decisions.
.
53
53Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
Basic challenges of soil monitoring
• Assurance of long-termcontinuity and consistencyunder changing boundary conditions as
site conditions, methodologies, etc.
• Assessment of reliability by adequate quantification of long-term bias, variance and confidence intervals.
• All changes of potentially relevant boundary conditions must be documented by appropriate metadata.
The basic challenges are as well requirements for a ‚good soil monitoring
practice‘ GSMP
54
54Overview and experiences of the Swiss soil monitoring network over 25 years
André Desaules | © Agroscope Reckenholz-Tänikon Research Station ART
References
This presentation is based on a presentation held at
EUROSOIL 2008 Conference, 25-29 August in Vienna,
Austria available in English under:
www.nabo.admin.ch >Presentations: NABO-Trend: T009
For references see: www.nabo.admin.ch >Bibliography