Considering Variability in Groundwater Exposure to Pesticides in LCA

Transcription

1 Considering Variability in Groundwater Exposure to Pesticides in LCA Georg Geisler, Simon Liechti, Stefanie Hellweg, Konrad Hungerbühler, Swiss Federal Institute of Technology (ETH), Zürich Method of incorporating variability in LCA Case study of pesticide leaching Relevance of pesticide leaching

2 Groundwater Exposure to Pesticides Acute and chronic toxicity => registration Unknown/uncertain effects?! => mixture toxicity atrazine, alachlor, nitrate 1) => Highly variable exposure and uncertain/unknown impacts on local scales => Precautionary approach: groundwater as safeguard subject 1) Jaeger et al. (1999) Tox Indust Health

3 FOCUS Modelling Framework and Leached Fraction Pesticide Registration FOCUS: harmonized modelling framework for pesticide leaching assessment site/management data user guidelines simulation models θ t =- z ψ K +1 -S r+ S z w Pelmo Macro p y,a value A y logic tree leached fraction (f j ) scenarios y p y,c p y,b value B y value C y z p z,a value A z m(leached) f j = m(applied) m(intercept) p z,b value B z

4 Processes influencing Pesticide Leaching: Water Balance Parameter in logic tree (derived from FOCUS) site: soil/weather combination 20 years of weather per site simulated => percentiles irrigation schedule (per site) crop 1 (min. percolate flow) crop 2 (max. percolate flow) influencing factor weather irrigation crop runoff precipitation/ irrigation evapotranspiration site: soil/weather combination Pelmo: matrix flow Macro: matrix and macropore flow soil water storage matrix flow macropore flow percolate flow

5 Logic Tree: Site Parameter FOCUS-sites mean annual precipitation characteristics of the sites mean annual temperature topsoil texture mean occontent (1m depth) site name \ unit mm/a C - % Sevilla < 600 > 12.5 silt loam 0.72 Chateaudun < silty clay loam 0.76 Hamburg sandy loam 0.78 Kremsmünster loam/silt loam 0.90 Okehampton loam 0.90 Piacenza > 12.5 loam 0.55

6 Processes influencing Pesticide Leaching: Pesticide Mass Balance Parameter in logic tree (derived from FOCUS) influencing factor application interception user input (interception tables) crop growth volatilisation Henry-constant cutoff (K h < J/mole) volatility of pesticide runoff Groundwater Ubiquity Score pesticide K oc pesticide DT 50,soil adsorption plant uptake artificial drainage degradation leaching

7 GUS-Space and Representative Pesticides log(dt50,soil) 1.E+03 1.E+02 1.E+01 Groundwater Ubiquity Score (GUS) 1) GUS = 6.4 GUS = 4.8 GUS = 3.5 GUS = GUS-categories 5 GUS = log(dt 50,soil ) * (4 log(k oc )) Pesticides from PETE-database Representative pesticides calculated Non-leachers (GUS<1.8; screening simulations only) GUS = log(dt 50,soil ) * (4 log(k oc )) GUS = E+00 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 log(k oc ) 1) Gustafson D I (1989) Environ Toxicol Chem

8 Processes influencing Pesticide Leaching: Pesticide Mass Balance Parameter in logic tree (derived from FOCUS) influencing factor application interception user input (interception tables) crop growth volatilisation Henry-constant cutoff (K h < J/mole) volatility of pesticide runoff Groundwater Ubiquity Score pesticide K oc pesticide DT 50,soil adsorption plant uptake season of application artificial drainage (exists => insignificant leaching) all artificial drainage artificial drainage leaching degradation

9 Logic Tree of Scenarios of Pesticide Leaching parameter group pesticide substance properties site parameters management parameters leached fraction f j parameter name volatility (K h ) GUSregion artificial drainage GUScategory site irrigation macropore temporal crop season of variability of application flow weather Piacenza pesticide application; interception low does not exist exists Okehampton Hamburg Chateaudun Sevilla yes no yes no f j = m(leached) m(applied) m(intercept) Kremsmünster worstcase average max. percolate flow autumn 0.12 high 1 0 bestcase min. percolate flow summer spring 0.041

10 Contributions of Parameters to the Variability of the Leached Fraction 1.E+20 Quotient Q 1.E+18 1.E+16 1.E+14 1.E+12 1.E+10 1.E+08 1.E+06 1.E+04 1.E+02 Q= f j /f base median 75 th to 25 th percentile minimum resp. maximum 1.E+00 scenario j GUS = 6.4 GUS = 4.8 GUS = 3.5 GUS =3.9 GUS = 2.7 GUS = 5.3 GUS = 3.9 Piacenza(irr.) Kremsmünster Piacenza(n-irr.) Hamburg Okehampton Chateaudun(irr.) Chateaudun(n-irr.) worst-case average Macro-Model max. percolate flow autumn summer base scenario parameter Sevilla GUS site temporal best-case Pelmo min. percolate spring flow macropore Technology Group application crop season of Safety and variability Environmental of weather flow

11 Aggregation of Leaching Scenarios: Site Factors parameter group site parameters Model output parameter name artificial drainage 0.10 site Piacenza irrigation macropore flow temporal variability of weather probability of occurrence of scenario j (p j ) leached fraction f j 0.80 does not exist Okehampton Kremsmünster 0 yes pesticide application; interception 0.20 exists Hamburg 0.18 Chateaudun 1 no 0 1 yes no worstcase average Sevilla 0.25 bestcase Weighted Leached Fraction: fw = Σ f j * p j

12 Aggregation of Leached Fractions: Substance Properties of Pesticides K oc, DT 50,soil cp(gus) variability due to site uncertainty of measurement GUS... Cumulative probability of GUS of the Pesticide: p 3 p 2 p 1 GUS(pesticide) Weighted Leached Fraction per GUS-category (fw c ): fw 1... fw 3... fw 5 GUS-category Total Weighted Leached Fraction: fw s = Σ fw c * p c => weighted average that incorporates variability

13 Case Study: Atrazine in Maize Assumptions: Results: season of application = spring no interception (preemergence application); all sites included (geographical scope = EU-agriculture) irrigation yes crop 1 GUScategory GUS-range logic tree output Atrazine GUS leached fraction per GUS-category (fw c ), % probability of GUS-value of Atrazine (p c ), % 0 GUS < GUS < E GUS < E GUS < GUS < GUS < GUS total weighted leached fraction fw s = Σ fw c * cp c fw s = 1.6 % substance properties of Atrazine: Fenner (2001);

14 Case Study: Atrazine in Maize Human Toxiciy Potential in EDIP Tentative normalisation and weighting of HTP(groundwater) Nitrate and pesticide emissions to groundwater considered Normalisation and weighting factor similar to that of HTP(surface water) Full pesticide life cycle for Atrazine in maize: Functional Unit: 1 ha Maize treated with recommended dose (1 kg/ha) Atrazine production: ecoinvent 1994 (only energy demand) Application by tractor: ecoinvent 2000 (emissions); ecoinvent 1994 (infrastructure); BUWAL300 (Diesel) => Fraction of HTP(groundwater) due to Atrazine leaching in HTP(total life cycle): 0.5 % Valuation Question?

15 Use of Leached Fractions in LCA A) Emission mass flows in LCI impacts of pesticide leaching other impacts in a life cycle characterisation factors for emissions to groundwater unavailable normalisation of HTP(groundwater) possible for EDIP valuation of LCIA not always in line with goals of study impacts on local scales often dominated by energy impacts in LCA 1) B) Separate impact indicators with groundwater as safeguard subject for studies where local, uncertain impacts are valued highly (precautionary approach), separate indicator would be appropriate 1) Beck A, Scheringer M, Hungerbühler K (2000): Int J LCA 5

16 Conclusions/Outlook number of scenarios Leached fractions... Incorporation of variability of local-scale impacts, in a sophisticated average.... use in LCA pesticide application Relevance of groundwater exposure to pesticides will often be low in LCA logic tree total weighted leached fraction Separate indicator may be appropriate, depending on goal of study Logic tree and modelling of leached fractions Model uncertainties (FOCUS-framework): Improvements of logic tree Scenarios of median leaching vulnerability Assessing pesticide flux in 1 m depth Higher resolution of GUS-categories Include more management factors Scenario/Aggregation approach to incorporate variability in LCA Use for other emission pathways of pesticides. Broader use for site-dependent LCA.

17 Acknowledgements Simon Liechti Andreas Huber Gabriel Popow Stefan Erzinger Group S&U, ETH Zürich Syngenta Crop Protection S.A. LIB Strickhof, Lindau FAT, Tänikon Ulrich Fischer Steffanie Hellweg Group S&U, ETH Zürich Konrad Hungerbühler

18

19 Risk Assessment Life Cycle Assessment FOCUS-framework realistic worst-case 80th percentile of soil vulnerability leaching vulnerability not modifyable Life Cycle Assessment median case plus variability 80 th percentile soil in different sites 80th percentile of leaching vulnerability due to weather median weather plus 20 th and 80 th percentile target parameter pesticide concentration in 1 m depth mass flow in 1 m depth leached fraction of dose

20 Site and Management Factors - Interdependencies scenarios with combinations of dependent parameter values calculated for each site irrigation applied macropore flow scenario calculated crop case 4 crop corresponding to crop case season of application mean annual precipitation characteristics of the sites mean annual temperature topsoil texture site name \ unit mm/a C - % Sevilla n/r 3 n only one 1 winter wheat autumn, spring < 600 > 12.5 silt loam 0.72 Chateaudun n y only one 1 winter wheat autumn, spring < silty clay loam Chateaudun y y 1 maize autumn, spring, summer Chateaudun y y 2 sugar beets autumn, spring, summer silty clay loam silty clay loam Hamburg n/a 2 n 2 winter wheat autumn, spring sandy loam 0.78 Hamburg n/a 2 n 1 maize autumn, spring, summer Kremsmünster n/a 2 n 1 maize autumn, spring, summer mea organ carbo conte sandy loam loam/silt loam Kremsmünster n/a 2 n 2 winter rape autumn, spring loam/silt loam Okehampton n/a 2 n 1 maize autumn, spring, summer loam 0.90 Okehampton n/a 2 n 2 winter wheat autumn, spring loam 0.90 Piacenza n n 1 winter wheat autumn, spring > 12.5 loam 0.55 Piacenza n n 2 winter rape autumn, spring > 12.5 loam 0.55 Piacenza y n only one 1 sugar beets autumn, spring, summer > 12.5 loam only one crop case calculated 2 n/a not applicable Georg 3 n/r Geisler: irrigation Variability is not a relevant in Pesticide factor for Leaching the leached in LCA fraction at the Sevilla-site 4 crop case 1 minimum percolate flow; crop case 2 maximum percolate flow

21 Statistics of the Leached Fractions calculated representative pesticide all GUS-value all GUS-region all ranked by median(f j ) GUS-value percentiles of f j median 75.percentile 25.percentile 95.percentile 5.percentile 3.7E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-03 minimum 0E+00 1E-20 2E-20 3E-09 3E-04 8E-03 2E-20 6E-17 8E-19 2E-16 1E-15 2E-11 maximum 8E+01 1E-04 1E-01 1E+01 5E+01 8E+01 2E-01 3E+00 3E+01 3E+00 2E+01 3E percentile/ 5.percentile 2E+14 6E+14 2E+13 3E+05 1E+02 3E+01 4E+16 8E+14 4E+08 2E+08 6E+04 5E+03 no. of scenarios with f j > fraction, % with f j > 30% fraction, % total