1. Evaluation of a multisurface complexation reactive transport model on field data. Bert-Jan Groenenberg 1, Joris Dijkstra 2, Rob Comans 2,3 1 Alterra Wageningen UR; 2 ECN; 3 Wageningen University

2. Introduction Use of reactive transport models in risk assessment of transport of metals to ground- and surface waters: Site applications Generic risk assessment e.g. derivation of maximum leaching of metals from building materials to groundwater in the Netherlands # Need for validation # Verschoor et al. 2008, 9th International Symposium on EnvironmentalGeotechnology and Global Sustainable Development, Hong Kong

3. Outline Outline of the model Description of the field site Evaluation of the geochemical model Evaluation of the reactive transport model by hindcast simulation of measured profiles

4. Process based Multisurface model Multisurface models describe the various processes by combining separate models for the distinguished reactive surfaces (SOM, Al/Fe-oxides, clay, …) Intrinsic model parameters derived from lab experiments on purified or synthetic model systems (Humic Acid, HFO, Goethite…)

5. Models: (2) Multisurface Model 1) Meeussen., ES&T 2003

6. Multisurface Model (1) Models in Framework ORCHESTRA Binding to SOM and DOM: NICA-Donnan * Binding to clay: Donnan model Binding hydrous oxides of Fe and Al: GTML Mineral equilibria and inorganic ion pair formation MINTEQ database *) Kinniburgh et al.,Colloids Surf. 1999

7. Multisurface Model (2) Models in Framework ORCHESTRA Binding to SOM and DOM: NICA-Donnan Binding to clay: Donnan model Binding hydrous oxides of Fe and Al: GTML Mineral equilibria and inorganic ion pair formation MINTEQ database

8. Multisurface Model (3) Models in Framework ORCHESTRA Binding to SOM and DOM: NICA-Donnan Binding to clay: Donnan model Binding hydrous oxides of Fe and Al: GTLM * Mineral equilibria and inorganic ion pair formation MINTEQ database *) Dzombak and Morel, 1990

9. Multisurface Model (4) Models in Framework ORCHESTRA Binding to SOM and DOM: NICA-Donnan Binding to clay: Donnan model Binding hydrous oxides of Fe and Al: GTML Mineral equilibria and inorganic ion pair formation MINTEQ database * * Allison et al., US-EPA 1991

10. Transport Model Soil column 0-150 cm Discretisized in layers of 10 cm thickness Continuous flow Every timestep: Transport Equilibration between solution and solid phase (chemical model)

11. Field site Waste water infiltration field 1930- 1970/1980 Used to clean sewage water from the city Tilburg before disposal to surface water Contaminated with Cr and other metals. Cr from leather tanning industry Used for cattle (1930-2000) 2000 plantation of forest

12. Measurements Profile 0-150 cm Soil solid phase: Metal contents total (AR) and reactive (0.43 M HNO 3 ) Soil organic matter: total and HA, FA Al and Fe (hydr)-oxides oxalate, dithionite and ascorbic acid Clay content Solution: 0.002 M CaCl 2 extracts and centrifugation: pH, macro ions (Al, Fe, Ca, P, S,....) DOC and FA and HA therein metals

13. Evaluation geochemical model Prediction of total solution concentrations from: Total reactive metal in soil (0.43 M HNO 3 ) SOM Al/Fe-(hydr)oxides Clay content pH DOC / FA and HA

14. Evaluation geochemical model

17. Hindcast simulation Start in 1930 (start use as infiltration fields) background levels estimated Loading with metals 1930-1980: total load = reactive metal in profile – background 1930 1980-2000 use as pastry pH maintained at 6 2000 plantation of forest (pH decreases) 2003 experimental acidification field plot 2009 measurement soil profile

18. Hindcast simulation Period 1Period 2Period 3acidified time50 years20 years10/ 3 years7 years Precipitation excess 2700 mm/y300 mm/y DOC150-20 mg/L pH6654

19. Hindcast simulation

23. Sensitivity

24. Conclusions / further research Geochemical model predicts solution concentrations satisfactorily except for Pb Model is able to “predict” present concentration profiles for most cationic metals Solubilization by acidification is reproduced well by the model

25. Conclusions / further research Sensitivity analysis shows that DOC is a critical input parameter Pb needs further attention Colloidal transport / aging Evaluation can be improved for better defined systems (known inputs, water fluxes etc.)

26. Thank you for your attention! © Wageningen UR