1. Role of soil moisture for climate extremes and trends in Europe Eric Jäger and Sonia Seneviratne CECILIA final meeting, 25.November 2009

2. Motivation Several major extreme events over Europe in recent years (e.g. heat wave 2003, Schär et al. (2004), Nature) -10 -5 0 +5 +10K R. Stöckli 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

3. Motivation Several major extreme events over Europe in recent years (e.g. heat wave 2003, Schär et al. (2004), Nature) Likely to become more frequent in future (IPCC 2007; e.g. Meehl and Tebaldi (2004), Science; Christensen and Christensen (2003), Nature) -10 -5 0 +5 +10K IPCC 2007 R. Stöckli 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

4. Motivation Several major extreme events over Europe in recent years (e.g. heat wave 2003, Schär et al. (2004), Nature) Likely to become more frequent in future (IPCC 2007; e.g. Meehl and Tebaldi (2004), Science; Christensen and Christensen (2003), Nature) Land-atmosphere interactions are a substantial contribution (Seneviratne et al. (2006), Nature) -10 -5 0 +5 +10K IPCC 2007 R. Stöckli 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

5. Model experiments Regional climate model CLM, 50km, driven by ECMWF re- analysis and operational analysis (1958-2006) (see Jaeger et al. [2008; 2009] for a validation of the model) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

6. Model experiments Interactive SM: CTL: control simulation Prescribed SM: SSV: lowpass filtered SM from CTL (cutoff ~10d) ISV: lowpass filtered SM from CTL (cutoff ~100d) IAV: SM climatology from CTL PWP: SM const. at plant wilting point FCAP: SM const. at field capacity Regional climate model CLM, 50km, driven by ECMWF re- analysis and operational analysis (1958-2006) (see Jaeger et al. [2008; 2009] for a validation of the model) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

7. Model experiments Interactive SM: CTL: control simulation Prescribed SM: SSV: lowpass filtered SM from CTL (cutoff ~10d) ISV: lowpass filtered SM from CTL (cutoff ~100d) IAV: SM climatology from CTL PWP: SM const. at plant wilting point FCAP: SM const. at field capacity Regional climate model CLM, 50km, driven by ECMWF re- analysis and operational analysis (1958-2006) (see Jaeger et al. [2008; 2009] for a validation of the model) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS Jaeger and Seneviratne., Clim. Dynam. (submitted)

8. Assessment of Extremes  Probability density functions of T max  Extreme indices (as defined in CECILIA)  Extreme Value Theory (block maxima & peak over threshold, both stationary and non-stationary) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

9. Assessment of Extremes  Probability density functions of T max  Extreme indices (as defined in CECILIA)  Extreme Value Theory (block maxima & peak over threshold, both stationary and non-stationary) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

10. Extremes: PDFs of T max 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS Eastern Europe

11. Extremes: PDFs of T max Soil moisture effects high T max values stronger than low ones 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS Eastern Europe

12. Extremes: PDFs of T max Soil moisture has a dampening effect on temperature 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS Eastern Europe

13. Extremes: HWDI HWDI = heat wave duration index: ‘mean number of consecutive days (at least two) with values above the long-term 90 th -percentile’ 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

14. Extremes: HWDI T time 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

15. Extremes: HWDI T time Long-term 90 th -percentile 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

16. Extremes: HWDI T time potential heat waves Long-term 90 th -percentile 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

17. Extremes: HWDI T time 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

18. T 1. ΔTΔT CTL Extremes: HWDI time 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

19. Δ T=0 Δ persistence T 2. CTL Extremes: HWDI time 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

20. Extremes: HWDI SSV-CTL ISV-CTL IAV-CTL PWP-CTL FCAP-CTL hwdi +9 -9 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

21. Extremes: HWDI SSV-CTL ISV-CTL IAV-CTL PWP-CTL FCAP-CTL hwdi +9 -9 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS Due to changes in the PDF of T max or due to changes in persistence?

22. T Extremes: HWDI time ΔTΔT CT L EXP 10% 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

23. Extremes: HWDI SSV-CTL ISV-CTL IAV-CTL PWP-CTL FCAP-CTL hwdi* hwdi +9 -9 +2.7 -2.7 Lorenz et al., GRL (submitted) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

24. Extremes: HWDI SSV-CTL ISV-CTL IAV-CTL PWP-CTL FCAP-CTL hwdi* hwdi +9 -9 +2.7 -2.7 Continuous reduction in HWDI, likely due to reduction in persistence associated with a loss of SM memory Lorenz et al., GRL (submitted) 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

25. Trends in T max T max 1981-2006: CTL 1959-1980:CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

26. Trends in T max : mechanisms? T max 1981-2006: CTL 1959-1980:CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS global-dimming global-brightening due to aerosol trends

27. Trends in T max : mechanisms? T max 1981-2006: CTL 1959-1980:CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS global-dimming global-brightening due to aerosol trends BUT: in CLM and ECMWF model aerosols are constant only an indirect effect of aerosols via data assimilation

28. Trends in T max : mechanisms? T max clouds SM 1981-2006: CTL 1959-1980:CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

29. no trend Trends in T max : link to SM T max clouds SM 1981-2006: IAV 1981-2006: CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

30. no trend Trends in T max : link to SM Aerosols are kept const. in CLM, hence (mostly) only cloud trends are the cause for T max trends, whereas SM act as an amplifier T max clouds SM 1981-2006: IAV 1981-2006: CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

31. Trends in T max (extremes) extremes mean 1981-2006: IAV 1981-2006: CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

32. Trends in T max (extremes) extremes mean Stronger than those in mean 1981-2006: IAV 1981-2006: CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

33. Trends in T max (extremes) extremes mean Stronger influence of SM 1981-2006: IAV 1981-2006: CTL 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

34. Conclusions SM asymmetrically effects the PDFs of Tmax, and dampens temperature variability 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

35. Conclusions SM asymmetrically effects the PDFs of Tmax, and dampens temperature variability Reduced SM memory causes a reduction in persistence in T max extremes hwdi* 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS

36. Conclusions 1959-1980 1981-2006 SM asymmetrically effects the PDFs of Tmax, and dampens temperature variability Reduced SM memory causes a reduction in persistence in T max extremes Trends in T max are likely due to trends in clouds, whereas SM acts as an amplifier hwdi* 1. BACKGROUND 2. RESULTS 3. CONCLUSIONS