1. Understanding physical processes linked to climate variability and change in the South America Monsoon System Carolina Vera 1, C. Junquas 1,2, H. Le Treut 2, L. Li 2 1 CIMA/CONICET-University of Buenos Aires and DCAO/FCEyN,, Buenos Aires, Argentina 2 LMD/IPSL, Paris, France 1

2. 2 Precipitation changes as depicted by the WCRP/CMIP3 multi-model ensemble (IPCC-AR4, 2007)

3. 3 Which are the physical mechanisms explaining an increase of summer precipitation in southeastern South America under GHG increment scenario?

4. 4 DJF climatological mean precipitation (CMAP) SACZ South Atlantic Convergence zone LPB La Plata Basin

5. Leading pattern of year-to-year variability of DJF precipitation anomalies (CMAP) EOF1 - domain A EOF1 - domain B A B 5

6. To explore changes in the leading pattern of precipitation year-to-year variability in South America in the context of a climate change induced by GHG increment. To explore how much of those changes account for the trends projected for austral summer precipitation in southeastern South America OBJECTIVES 6

7. 18 models from (WCRP/CMIP3) dataset for IPCC- AR4 are used. For each model, EOF analysis of DJF precipitation anomalies from XXI century simulations is performed over southeastern South America. Data and Methodology 7

8. Leading pattern (EOF1) of DJF rainfall anomaly variability from the 18 models Period: 2001-2099 Scenario: SRESA1B 8

9. Principal Components (PC1) of EOF1 from two of the models Blue (red) : PC1 larger (smaller) than 1 (-1) standard deviation  Positive(Negative) EOF1-SE. Positive (negative) EOF1-SE are associated with positive (negative) precipitation anomalies in LPB 9

10. Changes in the number of positive and negative EOF1-SE 10

11. Mean number of positive and negative EOF1-SE 11 Increment of wetter than normal DJFs in the la Plata Basin Decrease of dryier than normal DJFs in the la Plata Basin

12. Mean number of positive and negative EOF1-SE 9 models were selected that show: i) a realistic representation of the dipole-like structure associated to EOF1 events in the present (not shown) as in the future ii) an increase of the projected summer rainfall in LPB by the end of the 21 st century. iii) an increase of the frequency of positive EOF1 events and a decrease of negative EOF1 events during the 21st century 12

13. DJF rainfall difference between (2050-2099) and (2001-2049) EOF1-SE years (~30% of year total) no EOF1-SE years 9-model ensemble 13 All years The differences in the three panels are standardized by the total number of years

14. Temporal evolution of the DJF normalized rainfall anomalies in LPB (38°S-26°S,64°-50°W) from 9-model ensemble 14 Blue (red) dots correspond to the rainfall anomalies associated to each of the positive (negative) EOF1 events identified for each of the models (the corresponding linear trends are depicted in dashed lines)

15. Composite differences of mean DJF SST anomalies between positive and negative EOF1 events from the 9-model ensemble mean (A) 2001-2049(B) 2050-2099 15 (B) –(A)

16. Composites of standardized SST anomalies for positive and negative EOF1 events over the central Tropical Pacific Ocean from 9-model ensemble 16 Warmer than normal Central Pacific conditions associated to an increment of wetter than normal summers in the la Plata Basin

17. Composite differences of DJF geopotential height anomalies at 500 hPa between positive and negative EOF1 events from 9-model ensemble 2001-2049 2050-2099 17

18. 18 Composite differences of moisture fluxes (arrows) and their divergence (contour) at 850 hPa between positive and negative EOF1 events from 9-model ensemble 2001-2049 2050-2099

19. 19 How much of the results are related to the ability of current climate models in correctly reproducing the tropical ocean dynamics not only in the Pacific but also in the Atlantic basins? Which is the role of other sources of climate variability (Indian Ocean, SAM) in explaining such trends? Which is the combined effect of GHG increment and land use change on the regional climate change? How will natural climate variability signal combine with that associated to climate change in the next decades? Discussion

20. Annual mean anomalies of precipitation in Posadas (Northeastern Argentina)

21. 21 How much of the results are related to the ability of current climate models in correctly reproducing the tropical ocean dynamics not only in the Pacific but also in the Atlantic basins? Which is the role of other sources of climate variability (Indian Ocean, SAM) in explaining such trends? Which is the combined effect of GHG increment and land use change on the regional climate change? How will natural climate variability signal combine with that associated to climate change in the next decades? Discussion Strategy Analysis of CMIP5 multi-model ensemble simulations New sensitivity studies performing numerical simulations with global and regional models at specific conditions