1. Combining hydraulic test data for building a site-scale model MACH 1.3 Modélisation des Aquifères Calcaires Hétérogènes Site Expérimental Hydrogéologique de l’Université de Poitiers (SEH) 6 october 2008, Normandoux meeting Jean-Raynald de Dreuzy, Alexandre Boisson Géosciences Rennes

2. Hydraulic test interpretations DataInterpretationArticleK (m/s) S s (1/m) time series Spectral analysis[ Bernard and Delay, 2008] 10 -5 Slug tests Barker[ Audouin and Bodin, 2008] [4.4 10 ‑ 5,7.7 10 ‑ 4 ] 2 10 ‑ 4 10 ‑ 12 -5 10 ‑ 7 Slug tests Homogeneous[ Audouin and Bodin, 2007] [1.6 10 ‑ 5,1.3 10 ‑ 4 ]10 ‑ 17 -10 ‑ 12 Well tests Barker[ Bernard et al., 2006] [5.5 10 ‑ 5,3.6 10 ‑ 3 ]3 10 ‑ 10 Well tests Dual porous medium homogeneous [ Kaczmaryk and Delay, 2007] 10 -5 S sf =[3 10 ‑ 7 ‑ 2 10 ‑ 6 ] S sm =[4 10 ‑ 5 ‑ 2.4 10 ‑ 4 ]  Broad range of values (diffusivity D=K/S ~10-100 m 2 /s)  Matrix value S sm =2 10 -5 m -1  Fracture value S sf =2 10 -6 m -1  Hydraulic properties  Homogeneous S s  Heterogeneous K

3. 5 heterogeneous transmissivity models of increasing complexity “Outer-site” transmissivityInner-site transmissivity

4. Homogeneous & Butler’s model from pumping tests ( M21 ) Butler, 1998 Theis H B

5. Zonation models from drawdown distribution (pumping in M21 ) 0,3m 0m0m test dates:10 to 14/02/2005 2min4min10min ? BI B4Z

6. Slug Pattern from cross-hole slug tests Audouin, O., and J. Bodin (2008), Cross-borehole slug test analysis in a fractured limestone aquifer, Journal of Hydrology, 348(3-4), 510-523. BS

7. M12-M15: “F IELD RESULTS ” Pumping tests in M15 and M12 Doublet = Composition of pumping tests

8. M12-M15: “F IELD RESULTS ” |Expected drawdowns| <20 cms 2 transient regimes: - fast ~3 minutes - slow > day to be confirmed Drawdown inversion

9. M12-M15: H & B Small differences between models Sensitivity concentrated to the interwell zone Right order of magnitude for the fats transient regime Right order of magnitude for the darwdowns Very good for M4 and MP6

10. Larger |drawdowns| because of the impervious zone M12-M15: BI & B4Z Drawdown inversion Larger variability facilitates model rating

11. M12-M15: BS

12. M12-M15: C OMPARISON Comparison of “large time” values

13. M12-M15: L ARGE TIME VALUES sensitive wells H&BBI, B3Z, BS

14. M12-M15: M ODEL Q UALITY Good performance without calibration

15. M12-M15: D ISCUSSION  Why is BS the closest model?  Use of slug tests  Integration of local information  Why are H, B, BI and B4Z not better?  All information taken from a single well (M21)  Models could be better for M21  Partial conclusions without M21 drawdown?  Is a 3D model necessary for flow?

16. M12-M15: BS  Head maps  Velocity maps

17. M12-M15: D ISCUSSION t=1e1s

18. t=1e2s

19. t=1e3s

20. t=1e4s

21. t=1e5s

22. t=1e1s

23. t=1e2s

24. t=1e3s

25. t=1e4s

26. t=1e5s

27. M6-M22: D RAWDOWNS

28. M6-M22: S ENSITIVITY

29. t=1e1s

30. t=1e2s

31. t=1e3s

32. t=1e4s

33. t=1e5s

34. t=1e1s

35. t=1e2s

36. t=1e3s

37. t=1e4s

38. t=1e5s

39. Conclusions  On the transmissivity structure  1 st : Impervious zone  2 nd : Transmissivity pattern  On the interest in the available hydraulic data  Well tests: Characterize almost exclusively the outer zone  Slug tests: Most useful data for the inner site hydraulic characterization  On the planned field tests  Importance of the very short times (<1 min)  Pumping and reinjection flux chronicles  Is there a slow transient regime (>30h)?  What can be identified with the different hydraulic test types?  A priori sensitivity maps according to the hydraulic test configuration

40. Well tests

41. Doublet test

42. Relative drawdown

43. Perspectives  There could be better than the slug tests  Doublet drawdowns derived from single well tests  Improvement of the slug test pattern (BS)  Automatic calibration on doublet heads  Interest of field realization of doublets  Same recording time  Identification of non-Darcean flow conditions (O. Bour)

44. M12-M15: M ODEL Q UALITY