1. Spatial overlaping areas of several teleconection indices on Spain´s Mediterranean façade according to spring rainfall. J.C. González-Hidalgo (1), J.A. Lopez-Bustins (2), J. Martin-Vide (2), M. Brunetti (3), T. Nanni (3), P. Stepanek (4), M. de Luis (1) (1)Department of Geography, Zaragoza University, Spain (2)Group of Climatology, University of Barcelona, Spain (3)ISAC-CNR, Bologna, Italy (4) Czech Hydrometeorological Institute, Brno, Czech Republic

2. INTRODUCTION Hydrological cycle modification by global warming might affect precipitation. This impact could be more important than the global warming itself. Rainfall variability analyses is one of the main objectives of the IPCC, particularly in those areas which are strongly dependent upon water resources (e.g. Mediterranean basin). Global projections are not useful at local scale. High spatial resolution analysis are absolutely required.

3. HYPOTHESIS In the Mediterranean façade of Spain monthly precipitation is simultaneously affected by different atmospheric patterns. These patterns can be dominant or not depending on the different regions and months. There are overlapping areas. The analysis of these patterns may help us to understand the Mediterranean rainfall variability.

4. STUDY AREA Mediterranean façade of Iberian Peninsula (Spain), 1/3 of Iberian Peninsula, 37% of conterminous Spain, 182,000 km 2 Precipitation is characterized by irregular, seasonal and torrential nature

5. Precipitation database consists of monthly, homogeneous and complete series over 1,100 meteorological stations (period 1951-2000), Overall density 1 station / 150-200 km 2. Information is available to 1500 m o.s.l. DATA BASE

6. Altitude mSeries < 250292 250-500255 500-750227 750-1000189 1000-1500137 1500-17509 >17504 Monthly Precipitation data, Mediterranean of Spain MOPREDA MES

7. and METHOD Pearsons correlation coefficient between monthly precipitation dataset and teleconnection indices: NAOI series from Jones et al. (1997, IJC 17, 1433-1450.) MOI series from CRU web site McI series from Brunetti et al. (2002, IJC 22, 1455-1471) WeMOI series from Martin-Vide and Lopez- Bustins (2006, IJC 26, 1455-1475).

8. Dipoles of NAOI, MOI, McI and WeMOI teleconnection indices Padua Lod Airp Israel Marseille Padua Lod Airp Israel Marseille Gibraltar/ San Fernando Iceland

9. Correlation between indices NAOI correlates with MOI and McI but not with WeMOI. MOI correlates with McI and WeMOI McI correlates poorly with WeMOI (and negatively) WeMOI is independent of NAOI, MOI is the mediterranean expresion of NAOI, and McI is a subregional pattern linked to MOI-NAOI

10. RESULTS

11. In the study area February-May precipitation is under different teleconnection indices influence.

12. Positive correlation Negative correlation Pearson > 0.50, <-0.50 FEBRUARY correlation between rainfall and teleconnection

13. Highest teleconnection index correlation (number of stations) FebruaryMarchAprilMay NAOI 88 MOI 412 McI 48 WeMOI 546 FEBRUARY (not properly spring month) MOI predominates (inland), and WeMOI (east coastland) NAOI signal is weeker than MOI and it is only detected as predominant in a few meteorological station NW

14. FEBRUARY highest teleconnection index correlation (p < 0.05) +/- NAOI +/- MOI +/- McI +/- WeMOI None

15. McI NAO Positive correlation Negative correlation Pearson > 0.50, <-0.50 MARCH correlation between rainfall and teleconnection

16. Highest teleconnection index correlation (number of stations) FebruaryMarchAprilMay NAOI 292 MOI 734 McI 42 WeMOI 30 MARCH MOI expands towards the NE coast. NAOI appears in Pyrenees McI (-) phase and WeMOI (+) phase are the most correlated indices over NW (upper Ebro catchment)

17. MARCH highest teleconnection index correlation (p < 0.05) +/- NAOI +/- MOI +/- McI +/- WeMOI None

18. Positive correlation Negative correlation Pearson > 0.50, <-0.50 McI NAO APRIL correlation between rainfall and teleconnection

19. Highest teleconnection index correlation (number of stations) FebruaryMarchAprilMay NAOI 503 MOI 76 McI 73 WeMOI 201 APRIL NAOI is the most highly correlated index. WeMOI (negative phase) is predominant over the SE, and McI (negative phase) over the NW.

20. APRIL highest teleconnection index correlation (p < 0.05) +/- NAOI +/- MOI +/- McI +/- WeMOI None

21. Positive correlation Negative correlation Pearson > 0.50, <-0.50 MAY correlation between rainfall and teleconnection McI NAO

22. Highest teleconnection index correlation (number of stations) FebruaryMarchAprilMay NAOI 366 MOI 107 McI 135 WeMOI 193 MAY The overall spatial distribution of the atmospheric patterns influence is not homogeneous. The overall spatial distribution of the atmospheric patterns influence is not homogeneous. NAOI is still the most correlated index with precipitation on the Iberian Mediterranean façade. NAOI is still the most correlated index with precipitation on the Iberian Mediterranean façade.

23. MAY highest teleconnection index correlation (p < 0.05) +/- NAOI +/- MOI +/- McI +/- WeMOI None

24. CONCLUSIONS In the study area February-May precipitation is under different teleconnection indices influence.

25. CONCLUSIONS Atlantic fluxes increase their influence on spring rainfall while Mediterranean ones decrease (opposed results for autumn and winter analyses). Highest teleconnection index correlation (number of stations) FebruaryMarchAprilMay NAOI 88292503366 MOI 41273476107 McI 484273135 WeMOI 54630201193

26. CONCLUSIONS Monthly analysis let us to detect spatial and temporal shifts of the highest correlated areas between precipitation and the teleconection indices. The low correlation (although significant) between NAOI and precipitation suggests that MOI explains better than NAOI February-March precipitation over eastern Iberian Peninsula. This approach could help us to understand the nature of eastern Iberia rainfall variability. And… overall ………….

27. ¿?

28. Thank you very much Autumn and winter analysis can be found in Geophysical Research Abstracts, Vol. 9, 02219, 2007. SRef- ID: 1607-7962/gra/EGU2007-A-02219 EMS7/ECAM8 Abstract, Vol 4, EMS2007-A-00190, 2007. 7th EMS Annual Meeting / 8th ECAM.

29. +/- NAOI +/- MOI +/- McI +/- WeMOI None OCTOBER-NOVEMBER-DECEMBER-JANUARY Highest correlation

30. +/- NAOI +/- MOI +/- McI +/- WeMOI None FEBRUARY-MARCH-APRIL-MAY Highest correlation