Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Department of Environmental Sciences and Land Use Planning - UCL
Effective approaches for Assessing thePredicted Environmental COncentrations ofPesticides
APECOPA project supporting the harmonised registrationof pesticides in Europe
M. Vanclooster, J.D. Pineros-Garcet, J. Boesten, F. Van den Berg, M. Leistra, J.Smelt, R. Hendriks, N. Jarvis, S. Roulier, M. Larsbo, P. Burauel, H. Vereecken,
A. Wolters,V. Linnemann, K. Smith, E. Fernandez, M. Trevisan, E. Capri, L.Padovani, M. Klein, A. Tiktak, A. Van der Linden, D. Denie, G. Bidoglio, F.
Baouroui, A. Jones, A. Armstrong, C. Mouvet
EU-FP5- Quality of Life programme
Department of Environmental Sciences and Land Use Planning - UCL
Who?
• Université catholique de Louvain (M. Vanclooster et al.)• Alterra (J. Boesten et al.)• Swedish university of agricultural sciences (N. Jarvis et al.)• Forschungszentrum Jülich GmbH (P. Burauel et al.)• Instituto de Recurso Naturales y Agrobiologica de Sevilla
(E. Fernandez et al.)• Universita Cattolica de Sacro Cuore (M. Trevisan et al.)• Fraunhofer Institut für Umweltchemie und Okotoxikologie
(M. Klein et al.)• Rijksinstituut voor volksgezondheid en milieu (A. Tiktak et al.)• Joint Research Centre (G. Bidoglio et al.)
Department of Environmental Sciences and Land Use Planning - UCL
Research context
Directive 91/414/EU: ‘ Uniform principles ’ for registrationof crop protection products
“...Use validated simulation models for PEC to groundwater, surfacewater and air…”
EU/DG-SANCO: FOCUS (Forum for the co-ordination ofthe pesticide fate models and their use) working groups
• Develop guidelines on the selection and use of pesticide emissionmodels• Define scenarios for the use of pesticide emission models in theprocess of registration
Department of Environmental Sciences and Land Use Planning - UCL
Objectives
To evaluate the validation status of PEC groundwatermodels
To evaluate the validation status of PEC groundwaterscenarios (i.e. FOCUS scenario’s)
To propose new effective approaches considering• preferential flow• volatilization
Department of Environmental Sciences and Land Use Planning - UCL
Pan-European agri-environmental data
i.e. FOCUS Groundwater Scenario
Fate and transport model for the soil crop continuum• PEARL• MACRO• PELMO
Pan-European PEC
Valid?
Valid?
Department of Environmental Sciences and Land Use Planning - UCL
Validating local scale PEC groundwatermodels
Department of Environmental Sciences and Land Use Planning - UCL
Research method
Department of Environmental Sciences and Land Use Planning - UCL
Local PEC model
Fate and transport model for the soil-crop continuum• PEARL• MACRO• PELMO
Local scale agro-environmental data
Local PEC
Department of Environmental Sciences and Land Use Planning - UCL
Lanna
Jülich
Andelst VredepeelBrimstone
Bologna
Lebrija Coria
Local PEC model test sites
Department of Environmental Sciences and Land Use Planning - UCL
Local PEC model test sites
Lanna Sweden Cold humid Silty clay over clay Larsson and Jarvis,1999
Brimstone United Kingdom Moderate seaclimate
Cracking heavyclay
Bromilow et al.,1998; Harris andCatt, 1999;Armstrong et al.,2000.
Andelst the Netherlands Moderate seaclimate
Medium clay In press.
Vredepeel the Netherlands Moderate seaclimate
Sand Boesten and vander Pas, 1999,2000
Lebrija Spain Mediterranean Silty (drainedmarshland)
Andreu et al.,1996, Rieu et al,1998;
Coria Spain Mediterranean Silty clay loam In press.Bologna Italy Mediterranean Loam Araldi, 1997,
Scarabello, 1999
Department of Environmental Sciences and Land Use Planning - UCL
Evaluation protocol for the local scale PEC model
• Blind validation (neither laboratory or field data)– Evaluate generic parameter identification procedures
(pedo-transfer functions, environmental fate data base,…)
• Calibration using laboratory data– Evaluate local scale parameter identification approach
• Calibration using field data– Evaluate inverse modelling approach
• Blind prediction
Department of Environmental Sciences and Land Use Planning - UCL
Evaluation protocol for the local scale PEC model
• Phased approach– Water, solute, heat and pesticide fate evaluated
separately• Use of standardised evaluation criteria
– Graphical and numerical criteria
Department of Environmental Sciences and Land Use Planning - UCL
Target objective
• Soil moisture in the profile• Soil temperature• Soil water drainage• Groundwater position• Bromide in profile, drainage water and
groundwater• Active substance in profile, drainage water and
groundwater• Products: bentazone, ethoprophos, imidacloprid,
aclonifen, isporoturon, chloridazon
Department of Environmental Sciences and Land Use Planning - UCL
Exemple
Department of Environmental Sciences and Land Use Planning - UCL
10/12/1994
Moisture content (% vol)
0.20 0.25 0.30 0.35 0.40 0.45 0.50
Dep
th (
cm)
0
20
40
60
80
100
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
9/19/1995
Moisture content (% vol)
0.20 0.25 0.30 0.35 0.40 0.45 0.50
Dep
th (c
m)
0
20
40
60
80
100
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
measured and predicted drainage
09-94 01-95 05-95 09-95 01-96
drai
nage
(mm
d-1 )
0
2
4
6
8
10
12
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
11/10/94
Bromure content (g/m3)
0 5 10 15 20 25 30
Dep
th (c
m)
0
20
40
60
80
100
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
9/19/1995
Bromure content (g/m3)
0 1 2 3 4 5
Dep
th (c
m)
0
20
40
60
80
100
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
bromide concentration in drainage
09-94 01-95 05-95 09-95 01-96
conc
entr
atio
n (m
gl-1
)
0
2
4
6
8
10
12
14
16
predictedmeasured
Department of Environmental Sciences and Land Use Planning - UCL
12/15/1994
Bentazone content (mg/m3)
0 100 200 300 400 500 600
Dep
th (
cm)
0
20
40
60
80
100
measuredpredicted
Herbicide
Department of Environmental Sciences and Land Use Planning - UCL
09/19/1995
Bentazone content (mg/m3)
0 20 40 60 80 100 120
Dep
th (c
m)
0
20
40
60
80
100
measuredpredicted
Department of Environmental Sciences and Land Use Planning - UCL
bentazone concentration in drainage
09-94 01-95 05-95 09-95 01-96
conc
entr
atio
n (µ
gl-1)
0
50
100
150
200
250measuredsimulated
Herbcide concentration in drainage water
Department of Environmental Sciences and Land Use Planning - UCL
Some remarks: weak points on the procedure
• No benchmarking on analytical solutions• No stochastic validation approach (No ‘statistical’
validation test is done since uncertaintypropagation is not considered explicitly in thevalidation approach)
Department of Environmental Sciences and Land Use Planning - UCL
New modelling approaches for preferentialflow and volatilization
Department of Environmental Sciences and Land Use Planning - UCL
Local scale PEC model improvement
Preferential flow
Department of Environmental Sciences and Land Use Planning - UCL
Exemple
Department of Environmental Sciences and Land Use Planning - UCL
Time (days)
0.0 100.0 200.0 300.0 400.0
Con
cent
ratio
n in
dra
inflo
w ( m
g dm
-3)
0.0
2.0
4.0
6.0
8.0
Drain set 1Drain set 2Simulated
Department of Environmental Sciences and Land Use Planning - UCL
Time (days)
0.0 100.0 200.0 300.0 400.0
Con
cent
ratio
n in
dra
inflo
w (
mg d
m-3
)
0.0
5.0
10.0
15.0
20.0
Drain set 1Drain set 2Simulated
Department of Environmental Sciences and Land Use Planning - UCL
Department of Environmental Sciences and Land Use Planning - UCL
Volatilization
Department of Environmental Sciences and Land Use Planning - UCL
Validating groundwater scenarios
Department of Environmental Sciences and Land Use Planning - UCL
Why scenario validation?
• The selection/definition of the “scenario”, whichis a data model, is equally important than theselection/definition of the “leaching model”
• Bias in expert judgement cannot be excluded• Prepare authorization procedures after 2012• Prepare harmonized European higher tier
assessment
Department of Environmental Sciences and Land Use Planning - UCL
Pan-European scale PEC modelPan-European agri-environmental data
i.e. FOCUS Groundwater Scenario
Fate and transport model for the soil crop continuum• PEARL• MACRO• PELMO
Pan-European PEC
Department of Environmental Sciences and Land Use Planning - UCL
Definition of FOCUS groundwater scenario"A representative combination of crop, soil, climate and agronomicparameters to be used in modeling; representative means in thiscontext that the selected scenarios should represent physical sitesknown to exist, i.e. the combination of crop, soil, climate andagronomic conditions should be realistic" (FOCUS, 1995).
A combination of parameter values selected in such a way that theleaching calculated with pesticide leaching models using theseparameter values equals the 90th percentile of leaching inside thecorresponding FOCUS area.
Department of Environmental Sciences and Land Use Planning - UCL
Evaluation/validation of a groundwaterscenario?
Validation level 1 : Are FOCUS scenarios combinations ofparameter values selected in such a way that when used inpesticide PEC modeling the calculated leaching equals the 90thpercentile of leaching inside the corresponding FOCUS area, asobtained from measurements?
Validation level 2 : Are FOCUS scenarios combinations ofparameter values selected in such a way that when used inpesticide PEC modelling the calculated leaching equals the90th percentile of leaching inside the corresponding FOCUSarea, as obtained from detailed and spatially distributedmodeling results ?
Department of Environmental Sciences and Land Use Planning - UCL
Validation level 1
Not possible for the time being since pan Europeandata on pesticide concentration in groundwater are notavailable.
Department of Environmental Sciences and Land Use Planning - UCL
Validation level 2: Summary of the protocol
• Step 1: Identify the FOCUS area
• Step 2: Calculate probability of PEC groundwater withFOCUS scenario è(PEC_FOCUS)
• Step 3: Calculate probability of PEC groundwater withdetailed and spatially distributed pan European assessmentmodel è(PEC_PE)
• Step 4: Test: 90-th perc. (PEC_PE) 90-th perc.(PEC_FOCUS)
è Reject/accept
Department of Environmental Sciences and Land Use Planning - UCL
PEC_ RM PEC_ FOCUS
Pdf Pdf
Spatially distributed approach Actual FOCUS approach
Department of Environmental Sciences and Land Use Planning - UCL
Step 1: Definition of the FOCUS area’s
Department of Environmental Sciences and Land Use Planning - UCL
Definition of the FOCUS groundwater areas
FOCUS AREA
mean annualrainfall
mean annualtemperature
Correspondingscenario
1 12.5 Porto
èDefine spatially
Department of Environmental Sciences and Land Use Planning - UCL
FOCUS area with arable land mask
Department of Environmental Sciences and Land Use Planning - UCL
The issue of grassland
Department of Environmental Sciences and Land Use Planning - UCL
Step 2: Definition of PEC_FOCUS: Trivial
Department of Environmental Sciences and Land Use Planning - UCL
Step 3: Definition of PEC_PE: Europearl
Department of Environmental Sciences and Land Use Planning - UCL
Approach in APECOP:
•Step 1 (RIVM contribution: Tiktak and Denie):Perform simulation with PEARL, using Pan Europeandatabases
• PEARL is one of the FOCUS groundwater models;• Conceptual differences between FOCUS
approach and regional scale model minimised;
•Step 2 (UCL contribution: Pineros-Garcet andVanclooster):Substitute the rest of the data with a statistical(meta)model approach:
• Gives 100% coverage• We use an interpolation method, to make sure
that the original data points are retained.
Department of Environmental Sciences and Land Use Planning - UCL
Parameterisation of EuroPEARL
• Step 1: Derivation of the unique combinations,based on climate and soil mapping units;
• Step 2a: Parameterisation of the soil profiles(linkage between soil map and Soil Profile AnalyticalDatabase of Europe, SPADE);
• Step 2b: Parameterisation of weather conditionsfor each individual plot;
• Step 3: Translation into model parameters, usingpedotransfer functions etc.
Department of Environmental Sciences and Land Use Planning - UCL
Step 1: Derivation of Unique Combinations (UC)
Soil Map (FAO)
• Created by JRC
• 1:1.000.000
• ‘STU’ info
FOCUS areas
•LU mask
•Precipitation andtemperature maps
Country
•Based on theSMU soil map
+ +
Department of Environmental Sciences and Land Use Planning - UCL
Step 1: Derivation of Unique Combinations (UC)
Unique combinations =>
• 1410 ‘calculation’ plots
• each UC contains:
- SMU
- Country
- FA
Department of Environmental Sciences and Land Use Planning - UCL
Step 2a: Combine SPADE & SMU’s to PEARL profiles
Current status:
• SW and AUT notreperesented in SPADE
• 65 % of the UC areaparameterised
Department of Environmental Sciences and Land Use Planning - UCL
Future possibilities:
• Expansion to 75% possible
• Linking 2nd dominant STUto UC ?
• Include more climaticparameters in constructionUC’s
• More detailed irrigationdata
Department of Environmental Sciences and Land Use Planning - UCL
Step 2b: Combine daily weather & climate maps
9 Climatezones
(10 x 10 km2)
Daily weather datafor each UC
• Temperature• Precipitation• Potential evapotranspiration
9 time series(FOCUS)
Precipitation &temperature pattern
10 x 10 km2
Precipitation &temperature pattern
for each UC
Scaling
Department of Environmental Sciences and Land Use Planning - UCL
Step 3:specific modelparameters
RunSubstance IDPlot IDManagement IDStart dateEnd date
RunSubstance IDPlot IDManagement IDStart dateEnd date
SubstanceSubstance properties, such as the half-live and the partitioning coefficient
SubstanceSubstance properties, such as the half-live and the partitioning coefficient
PlotPlot IDFOCUS Area IDLand-use type IDSoil profile IDGroundwater depth group IDSeepage flux and amplitudeDrainage characteristics
PlotPlot IDFOCUS Area IDLand-use type IDSoil profile IDGroundwater depth group IDSeepage flux and amplitudeDrainage characteristics
ManagementApplication dateApplication typeApplication dosage
ManagementApplication dateApplication typeApplication dosage
Plot IDRainfallTemperatureETrefIrrigation switch
Plot IDRainfallTemperatureETrefIrrigation switch
Soil profileSoil layer ID
Soil profileSoil layer ID
Soil layerSoil physical unit IDLayer thicknessTextureOrganic matterpH
Soil layerSoil physical unit IDLayer thicknessTextureOrganic matterpH
FOCUS AreaEmergence dateHarvest dateDevelopment stage IDCritical pressure headsfor drought stress andirrigation
FOCUS AreaEmergence dateHarvest dateDevelopment stage IDCritical pressure headsfor drought stress andirrigation
Development stageLAICrop factorRooting depth
Development stageLAICrop factorRooting depth
Soil physicsParameters ofthe Mualem-van GenuchtenfunctionsDispersion length
Soil physicsParameters ofthe Mualem-van GenuchtenfunctionsDispersion length
Spatially distributed variables
Department of Environmental Sciences and Land Use Planning - UCL
Some preliminary results of EURO-PEARL
Department of Environmental Sciences and Land Use Planning - UCL
• statistical model• similar to kriging,• useful and efficient in multidimensional
parameter spaces.
Fill out missing gaps: the EURO-PEARLmeta-model ?
Department of Environmental Sciences and Land Use Planning - UCL
What are the inputs and outputs of themetamodel?
SandSiltClayOMpHAveWatBulk densityTemp_yrPrec_yrEva_yrTemp_springPrec_springEva_springTemp_springPrec_springEva_springGWLevel
mean yearleaching concentration
(atrazine)me
tamode
l
Department of Environmental Sciences and Land Use Planning - UCL
What do we use for calibration?
• 656combinationsof the inputparameters
• 656 Pearl outputs
P1 P2 P3xc1,1 xc1,2 xc1,3
xc = xc2,1 xc2,2 xc2,3...xc116,1 xc116,2 xc116,3
yc = yc1yc2...
yc116
metam
odel
Department of Environmental Sciences and Land Use Planning - UCL
What do we use for validation?
• 72 combinationsof the inputparameters
• 72 Pearl outputs
metam
odel
Department of Environmental Sciences and Land Use Planning - UCL
• From +-5400 unique combinations of soil and climate inthe Netherlands, RIVM did Pearl simulations.
• From those simulations, 728 unique combinations whereselected, having the following characteristics:
- not irrigated- atrazine- non drained- maize
How the calibration data where obtained ?
72 pts (10%, random) for validation | 656 : calibration
Department of Environmental Sciences and Land Use Planning - UCL
Calibration results
pred
icti
ons:
ln (l
each
ing
( ug
/l) +
1)
predictions: ln (leaching ( ug/l) + 1)
Not surprising:it is calibrated to go trough
all calibrations points
Department of Environmental Sciences and Land Use Planning - UCL
Validation scatterplotpr
edic
tion
s: ln
(lea
chin
g (
ug/l)
+ 1
)
predictions: ln (leaching ( ug/l) + 1)
Department of Environmental Sciences and Land Use Planning - UCL
10-4
10-3
10-2
10-1
100
101
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1EF
ME (µg / l)
RMS
MAD (maximal accepted distance, no units, scaled)
Department of Environmental Sciences and Land Use Planning - UCL
When and where will results be available
• Project deliverables available by march 2003
• Project results will be presented at the XII Pesticidechemistry symposium, Piacenza, Italy, May 2003
• Information
www.agro.ucl.ac.be/geru/recherche/projets/apecop-pub