Embed Size (px)
Citation preview
FOCUS work group on degradation kinetics
Overview and parent kinetics
Presentation at the3rd European Modelling Workshop
Catania, Sicily, 17-19 February 2004
2
Introduction
The degradation rates of parent and metabolites are important variables in assessing environmental exposure and movement to water
Assumptions used in calculating degradation rates can significantly affect these assessments
Need for harmonisation
3
Introduction
FOCUS (an organization co-sponsored by the EU and industry) established a work group to provide guidance on calculation of degradation rates of parent and metabolites
laboratory studiesfield studieswater sediment studies
4
FOCUS WORK GROUP ON DEGRADATION KINETICS
Group of 13 scientists chaired by Jos Boestenresearch institutesregulatory agenciesacademiaindustry
Six meetings starting in September 2002
Essentially complete final report (will be submitted to the FOCUS Steering Committee in early March)
5
Work Group Members
MartinDust
Jeremy Dyson
Inge Fomsgaard
Jos Boesten
Karin Aden
Claude Beigel
SabineBeulke
Russell Jones
SylviaKarlsson
Ton van der Linden
Oriol Magrans
Soria
Otto Richter Guy
Soulas
6
Types of Kinetics
0.00
0.03
0.06
0.09
0.12
0.15
0.18
0 50 100 150 200 250
Days after application
Soil
resi
due
(mg/
kg)
0
20
40
60
80
100
Time
Con
cent
ratio
n (%
)
DT50 DT90
Single first-order (SFO)Time for decrease from 100% to 50% same as time for decrease from 50% to 25%DT90 = 3.32 x DT50
Bi-phasicDegradation initially fast, then slowere.g. due to declining microbial activity (artefact!), increasing sorptionGustafson and Holden (FOMC)Hockey-stick (HS)Bi-exponential (DFOP)
7
Types of kinetics
Lag-phaseDegradation initially slow, then fastere.g. due to inappropriate storage before study (artefact!), adaptation of micro-organisms
Modified Hockey-stick modelLogistic model
0
20
40
60
80
100
0 20 40 60 80 100Time
Con
cent
ratio
n
8
Goodness of Fit
Aim to identify statistical measure that matches expert judgement
No single statistical measure was found to be universally valid
Acceptability of fits are judged on the basis of a Chi2 test and visual assessment
9
Chi2 Test (variant)
∑ −=χ 2
22
)O x 100/err()OC( C = calculated value
O = observed value= mean of all observed values
err = measurement error percentageΟ
If χ2 > tabulated value then the model is not appropriateat the chosen level of significance (usually 5%)
Error unknown Calculate error level at which test is passed
Model with smallest error percentage is best-fit model
10
Visual Inspection
Graph of predicted and observed concentrations vs time
Graph of residuals vs time (for kinetic models for exposure assessment, consider only the residues through the DT90)
11
Spreadsheet for parent SFO and FOMC
12
Objectives for Parent Kinetic Analysis
DT50 and DT90 values used to trigger additional studies
Chose best-fit model kinetics on statistical and visual analysis
Endpoints to calculate predicted environmental concentrations in groundwater and surface water
Standard versions of most fate models use SFOPreference for SFO when an adequate fit is obtained Correction procedures if an adequate fit is not obtained
13
Estimating Endpoints Used as Trigger Values
Run SFO and FOMC.If fit for SFO is better, then use SFO
If FOMC is better, run DFOP. Use whichever model gives the best fit (lowest error in Chi2 test)
Assess by visual inspection whether the best fit model provides an acceptable description of the data
14
Parent Endoints for Exposure ModellingRun SFO
SFO statisticallyand visually acceptable?
SFO DT50yes
no
Decline in study below 10%of dose ?
noyes
Tier 1 fate modelling based onDT50 back-calculated from
DT90 for FOMCDT50 = DT90 / 3.32
worse case approach
Tier 1 fate modelling based onslower degradation rateof
Hockey Stick or DFOP model
worst case approach
15
Example A (laboratory study)
16
Example A (laboratory study)
-4
-2
0
2
4
6
8
0 5 10 15 20 25 30 35
Time
Res
idua
l
17
Example A (laboratory study)
ConclusionsChi2 error value for SFO is 4%No systematic error apparent in residual plotsWell behaved data-set, very limited scatter in the measured dataSFO appropriate for use in modelling
Additional InformationNo improvement in Chi2 error or residual pattern with FOMC
18
Example B (laboratory study)
19
Example B (laboratory study)
-15
-10
-5
0
5
10
0 5 10 15 20 25 30
Time
Res
idua
l
20
Example B (laboratory study)
ConclusionsChi2 error value for SFO is 15%No systematic error apparent in residual plots up to DT90 (~5 days), underestimation afterwardsSFO appropriate for use in modelling
Additional InformationChi2 error value for FOMC is 7%No improvement in residual pattern with FOMC up to DT90
21
Example C (laboratory study)
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80 100 120 140 160 180 200
Time [days]
Con
cent
ratio
n
SFO_fitObserved
22
Example C (laboratory study)
-25
-20
-15
-10
-5
0
5
10
15
0 20 40 60 80 100 120 140
Time
Res
idua
l
23
Example C (laboratory study)
Conclusions-SFOChi2 error value for SFO is 22%SFO misses measured initial concentrationResidual plot indicates systematic deviation for later sampling dates in period up to DT90SFO inappropriate for use in modelling
24
Example C (laboratory study)
Conclusions-FOMCChi2 error value for FOMC drops to 8%Better description of initial concentrationNo improvement with regard to random nature of residuals, however overall smaller absolute deviations compared to SFOUse bi-phasic kinetics approaches for modelling
DT90 FOMC/3.32 or higher tier approaches
25
Example D (field study)
0
200
400
600
800
1000
1200
1400
1600
0 50 100 150 200 250 300 350 400
t [days after application]
c(t)
[% o
f app
lied]
measuredusedcalculated
26
Example D (field study)
-300
-200
-100
0
100
200
300
0 50 100 150 200 250 300 350 400
Time
Res
idua
l
27
Example D (field study)
Conclusions-SFOChi2 error value for SFO is 22%Residuals plots indicate no systematic error of the SFO model, rather that the observed pattern is most likely due to scatter of early measurementsSFO appropriate for use in modelling, confirm against bi-phasic kinetics
28
Example D (field study)
Conclusions-FOMCNo improvement in Chi2 error valueNo improvement in residual pattern
29
Other Topics in the Report
General data issuesReplicatesData transformation and weightingOutliersData below LOD and LOQTime zero samples
30
Other Topics in the Report
Metabolites
Water sediment studies
Use of kinetic endpoints in regulatory assessments
when averages are desired, use geometric meansame value when half-lives and degradation rates are averagedbest description of averages of entire degradation curves
31
Other Topics in the Report
Discussion of statistical measures
Normalization of field data
Higher tier approaches if degradation is bi-phasic
Reporting of results
Review of software packages
