Figure 1
Surface pressure anomalies in DJF for the observed El Niño composite of Fraedrich and Müller (1992) (above) and for the composite simulated by the coupled model (below). Units are expressed
in tenth of hPa
At a global scale, the impacts from ENSO are well documented in the literature since the end of the 80's (for instance, Kiladis and Diaz 1989; Ropelewski and Halpert 1987,1989,1996; Halpert and Ropelewski 1992). All these articles show that ENSO influence is considerable in all the inter-tropical region, especially in a strip which comes from South America, crosses the Pacific Ocean and Indonesia, and ends on the Indian Ocean. Moreover, some extra-tropical regions are affected, above all near the Pacific Ocean (North America, Australia, Japan), but not only there (South Africa, India). However, most of these studies lack significant correlation on Europe and the Mediterranean basin.
This article intends to take stock of the recent research regarding ENSO impact on Europe and the Mediterranean basin. These results are illustrated and interpreted through recent numerical simulations obtained with a general circulation coupled model.
2a. Northern winter
2a.1. Observations
The first results on Europe were obtained by Van Loon and Maden (1981): they analyse the correlation between the surface pressure in December-January-February (DJF) in the island of Cocos (12°S, 96°E) and the surface pressure in DJF in stations of the northern hemisphere. They show that, during the period 1952-1978, and during the El Niño phenomenon, a strip of negative anomaly in surface pressure extends over Europe, approximately from the North of Spain to the North of the Black Sea (correlation minimum < -0.4 centred on the north-west of Spain); and a strip of positive anomaly in pressure extends over the North of Europe, from Iceland to Scandinavia (correlation maximum > 0.4 centred in the North of Iceland). Through the analysis of the correlation between the surface pressure in the island of Cocos and the air surface temperature over the same period, they obtain during the El Niño period a nucleus of cold anomalies centred on Scandinavia (correlation minimum < -0.3) and a strip of warm anomalies extending from Spain to the Black Sea over the Mediterranean and North Africa. These results are significant only for Scandinavia regarding for surface pressure and for Portugal regarding temperature.
These results have been confirmed afterwards by Kiladis and Diaz (1989): with a composite of 27 El Niño events occurred between 1877 and 1986, they have obtained a significant cold anomaly on Scandinavia. They have also obtained two regions of significant positive anomalies in rainfall: the first one on France and Spain and the second one centred on the Black Sea.
A set of articles (Fraedrich 1990, 1993; Fraedrich and Müller 1992; Fraedrich et al 1992, Wilby 1993) have correlated ENSO with the frequencies of type of weather in Europe. They show that regimes of cyclonic weather are more frequent in western and central Europe during the winters of El Niño years, and that this fact is associated with a southwards movement of "storm-tracks": the trajectory of depressions is shifted towards South during El Niño years
2a.2. Simulation of ENSO impact on Europe during Nothern winter
The tool used in this study is the model of general atmospheric circulation Arpège-Climat (Version 2 T31L19) coupled with the model of general oceanic circulation OPAICE (Version 7 including a thermodynamic model of ice). No flow correction at air/sea interface is used. The studied simulation is said reference simulation or 1xCO2. It is 130-year long, and its 108 last years are analysed (year 23 to 130). For further details on the model, the preparation and implementation of the experiments, please refer to Barthelet et al (1998). In the scope of this simulation, 14 events of "El Niño" type have been selected in the composite. These 14 events are those for which the pressure anomalies in the tropical Pacific are the most significant (study performed from the temporal series associated with the first main component of the field of surface pressure anomaly on the tropical Pacific area).
In order to validate the composite event simulated by the model, we use a composite established by Fraedrich and Müller (1992) from 26 El Niño events between 1880 and 1988. Figure 1, which presents surface pressure anomalies in DJF, shows that the model overestimates (in absolute value) of about 1hPa (about -2.5hPa on Ireland versus -1.6hPa) the negative anomalies on western Europe, whereas at the eastern end of the strip of negative anomalies (on the Black Sea), the magnitude of the anomaly is correctly simulated. The model also overestimates the meridian extension of the strip of negative anomalies. Figure 2, which presents rainfall anomalies in DJF, shows that the maximum of positive anomalies is correctly situated in the model on the north-west part of the Iberian Peninsula; however, the model overestimates the anomalies (more than 40mm versus 34mm observed). On the contrary, anomalies on the British Isles are underestimated (about 10mm in Scotland versus 22mm observed). The general outline of the fields of surface pressure and rainfall anomalies is quite well reproduced by the model. Such is not the case for the anomalies in air temperature at 2 metres high: whereas the model simulates warm anomalies on Scandinavia, observations show cold anomalies.
2b. Northern spring
Kiladiz and Diaz (1989) point out that a wet strip of rainfall positive anomalies associated with El Niño extends in MAM over northern Europe, from Great Britain to the Black Sea, and that a dry area of negative anomalies in rainfall extends over the Maghreb and southern Spain. Both results have been confirmed by the later studies of Oldenborgh et al (1999) and Moron and Ward (1998), who remark as well a dry anomaly in the eastern part of Spain, which has been studied more thoroughly by Rodo et al (1997).
Halpert and Ropelewski (1992) find out a significant area of cold anomalies between February and May in the south-west of Europe and North Africa (centred on the Pyrenees). Figure 5 actually shows that the model simulates a strip of cold anomaly on these regions.
2c. Other seasons
During the autumn season (SON) preceding the El Niño event, Kiladiz and Diaz (1989) point out an area of positive anomaly in rainfall on the Maghreb and southern Spain.
During the summer season (JJA) preceding the El Niño event, Kiladiz and Diaz (1989) point out an area of positive anomaly in rainfall in the western part of the Mediterranean (affecting the Mediterranean parts of France, Italy and Spain).
However, these results should be taken cautiously and have to be confirmed by further studies.
3a. Observations
Van Loon and Madden (1981) analyse the correlation between the surface pressure in DJF in Darwin and the surface pressure in DJF in stations of the Northern hemisphere for 4 different periods (1899-1919, 1920-39, 1946-61 and 1962-77) on Europe: only the last two periods clearly show the structures of correlation expected by Fraedrich composites. Ropelewski and Halpert (1987) also point out that the year 1940 is a turning point regarding ENSO impact on Europe: before 1940, they obtain the same number of seasons with dry and wet anomalies; after 1940, they mainly obtain seasons with a wet anomaly (7 out of 8).
More recently, Rodo et al (1997) point out that the association between ENSO and rainfall on the Iberian Peninsula has significantly intensified in the second half of the XXth century. Other studies also show a similar phenomenon in other regions. Janicot et al (1996) show an intensification of the relationship between ENSO and the Sahelian droughts. Lough (1993) points out an intensification of the correlation between the southern oscillation index (SOI= difference in normalised pressure between Darwin and Tahiti) and rainfall in Queensland over the period 1951-1980 compared with 1921-1950.
3b. Simulations with twice the concentration of CO2
A second simulation with twice the concentration of carbon dioxide has been performed with the Arpège/OPAice coupled model. This simulation, which 66 last years are studied in the present article, is called "2xCO2 simulation " in the rest of the text. With the same selection criteria as those used in the reference experiment, 9 events of El Niño type have been selected to form the El Niño composite of the 2x CO2 simulation, which results are presented in figure 4. El Niño impact over Europe is significantly increased in the 2xCO2 simulation. This is visible on surface pressure (anomalies in the strip of negative anomalies extending over Europe are clearly accentuated; the North-South contrast is also reinforced with negative anomalies on a strip extending from Scandinavia to Greenland) as well as on rainfall (the strip of positive anomalies extending from the Iberian Peninsula to the Black Sea is significantly increased, the North/South contrast is more accentuated, with negative dry anomalies on Scandinavia). Together with an amplification in temperature anomalies, a modification in structure is evidenced, with cold anomalies on Scandinavia and warm ones on the Mediterranean basin. It is worth noting that the general outline of the fields of anomalies simulated in the 2xCO2 simulation is finally more similar to the observed one than to the outline of the reference experiment. This fact is also evidenced for the other seasons (spring, autumn and summer) in which the sign and the location of the anomalies observed on Europe are more correctly simulated in the 2xCO2 experiment than in the 1xCO2 experiment.
With the current acknowledgement of ENSO influence on global climate, the low signal on Africa and Europe may be interpreted with much more confidence. Nevertheless, its impact on these regions is low and the correlation between the various parameters are not very significant as a rule.
It is encouraging to note that, especially during northern winter, ENSO influence simulated by the coupled model is quite in accordance with the observations (location, intensity). This behaviour of the model in the reference simulation allows to trust to a certain extent the increased ENSO impact on Europe simulated in the 2xCO2 experiment.
In the tropical Pacific, the anomalies in ocean surface temperatures of the ENSO events in the 2xCO2 simulation are little modified compared with those of the reference simulation. In the model simulations, the intensification of ENSO impact on Europe does not result from a modification in the ENSO events themselves, but from a modification of the atmosphere status which is classically disturbed with a significant surface warming over the Arctic and in the high tropical troposphere, a more important warming above continents than above oceans, etc.
In conclusion, the increased correlation between ENSO and Europe in the second half of the XXth century could not only be due to an intensification of ENSO events as suggested by Moron and Ward (1998), but could be considered as an expression of global climate change as it has just been suggested by Rodo et al (1997).
Barthelet P., S. Bony, P. Braconnot, A. Braun, D. Cariolle, E. Cohen-Solal, J. Dufresne, P. Delecluse, M. Déqué, L. Fairhead, M. Filiberti, M. Forichon, J. Grandpeix, E. Guilyardi, M. Houssais, M. Imbard, H. L. Treut, C. Levy, Z. Li, G. Madec, P. Marquet, O. Marti, S. Planton, L. Terray, O. Thual, and S. Valcke 1998 : "Simulations couplées globales des changements climatiques associés à une augmentation de la teneur atmosphérique en CO2", C.R. Acad. Sci. Paris, 326, 677-684.
Déqué, M., C. Dreveton, A. Braun and D. Cariolle (1994) : "The version of Arpège/IFS: a contribution to the french community climate modelling", Clim. Dyn., 10, 249-266
Fraedrich K. 1990 : "European Grosswetter during the warm and cold extremes of the El Niño/Southern Oscillation", Intern.J.Clim., 10, 21-31.
Fraedrich K. and K. Müller 1992 : "Climate anomalies in Europe associated with ENSO extremes", Intern.J.Clim., 12, 25-31.
Fraedrich K., K. Müller and R. Kuglin 1993 : "Northern hemisphere circulation regimes during the extremes of ENSO", Tellus, 44A, 33-40.
Halpert M.S. and C.F. Ropelewski 1992 : "Surface temperature patterns associated with the Southern Oscillation", J.Clim., 5, 577-593.
Janicot S., V. Moron and B. Fontaine (1996) : "Sahel droughts and ENSO dynamics", Geophys. Res. Lett, 23, 515-518.
Kiladis G.N. and H.F. Diaz 1989 : "Global climatic anomalies associated with extremes in the Southern Oscillation", J.Clim., 2, 1069-1090.
Lough JM (1993) : "Variations of some seasonal rainfall characteristics in Queensland, Australia: 1921-1987", Intern.J.Clim., 13, 391-409.
Moron V. and M.N. Ward 1998 : "ENSO teleconnections with climate variability in the European and African sector", Weather, 53, 287-295.
Oldenborgh G.J., G. Burgers, A.K. Tank 1999 : "On the El Niño teleconnection to Spring precipitation in Europe", available at http://www.knmi.nl/~oldenbor/publ/oldenbor03.html
Ricard J.L. and L. Goulet, 2000 : "El Niño type events and their teleconnections over Europe simulated by the Arpège/OPAICE CGCM. Impact of a CO2 doubling", submitted to Climate Dynamics
Rodo X., E. Baert and F. Comin 1997 : "Variations in seasonal rainfall in Southern Europe during the present century : relationships with the NAO and the ENSO", Clim. Dyn., 13, 275-284.
Ropelewski C.F. and M.S. Halpert 1987 : "Global and regional patterns associated with the ENSO", Mon.Wea.Rev., 115, 1606-1626.
Ropelewski C.F. and M.S. Halpert 1989 : "Precipitation patterns associated with the high index phase of the Southern Oscillation", J.Clim., 2, 268-284.
Ropelewski C.F. and M.S. Halpert 1996 : "Quantifying SO-precipitation relationships", J.Clim., 9, 1043-1059.
Van Loon H. and R.A. Madden 1981 : "The southern oscillation. Part 1. Global associations with pressure and temperature in northern winter", Mon.Wea.Rev., 109, 1150-1162.
Wilby R. 1993 : "Evidence of ENSO in the synoptic climate of the British Isles since 1880", Weather, 48, 234-239