Influences of the tropical Indian and Atlantic Oceans on the

of 20

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
PDF
20 pages
0 downs
7 views
Share
Description
1 2 Influences of the tropical Indian and Atlantic Oceans on the predictability of ENSO 3 4 Claudia Frauen 1, 2 and Dietmar Dommenget IFM-GEOMAR, Leibniz Institute of Marine Sciences (at Kiel
Transcript
1 2 Influences of the tropical Indian and Atlantic Oceans on the predictability of ENSO 3 4 Claudia Frauen 1, 2 and Dietmar Dommenget IFM-GEOMAR, Leibniz Institute of Marine Sciences (at Kiel University) 2 School of Mathematical Sciences, Monash University, Clayton, Australia C. Frauen, IFM-GEOMAR, Leibniz Institute of Marine Sciences (at Kiel University), D Kiel, Germany. Current affiliation: School of Mathematical Sciences, Monash University, Melbourne, Victoria, Australia. D. Dommenget, School of Mathematical Sciences, Monash University, Melbourne, Victoria, Australia. Abstract The El Niño Southern Oscillation (ENSO) is the leading mode of interannual tropical variability and the only predictable mode on interannual time scales. Recent studies suggest that the tropical Indian and Atlantic Oceans influence the dynamics and predictability of ENSO. Here we examine these effects in a hybrid coupled model consisting of a full complexity atmospheric general circulation model (GCM) coupled to a linear 2-dimensional ENSO recharge oscillator ocean model. We find a strong dynamical effect of the tropical Indian Ocean onto ENSO and in perfect model forecast simulations the initial conditions of the tropical Indian and Atlantic Oceans have a significant impact on the predictability. Interestingly, the initial conditions of the tropical Atlantic Ocean have a stronger impact on the predictability skill than those of the tropical Indian Ocean, although the coupling to the tropical Atlantic has almost no effect on the dynamical parameters of ENSO. 35 2 Introduction For global seasonal and long-term climate predictions the predictability of ENSO is of particular importance. The origin of the ENSO mode lies in the interactions of the tropical atmosphere and the tropical Pacific Ocean, but the teleconnections of ENSO reach far beyond the tropical Pacific region. Areas influenced by ENSO are especially the tropical Indian and the tropical Atlantic Oceans and the adjacent continents (see for example Latif and Barnett, 1995; Enfield and Mayer, 1997). Recently, different studies suggest the tropical Indian and the tropical Atlantic Ocean may not only react to ENSO but have an influence on ENSO themselves. Especially a possible feedback of the tropical Indian Ocean on ENSO was subject of many studies (Liu, 2002; Yu et al., 2002; Wu and Kirtman, 2004; Annamalai et al., 2005; Kug and Kang, 2006; Dommenget et al., 2006; Jansen et al., 2009). Yu et al. (2002), Wu and Kirtman (2004), and Dommenget et al. (2006) analysed coupled ocean atmosphere GCMs where the influence of the Indian Ocean was suppressed. However, they find contradictory results regarding the Indian Oceans influence on ENSO variability and periodicity. While the analysed time series in Yu et al. (2002) and Wu and Kirtman (2004) are relatively short Dommenget et al. (2006) analysed 500-year-long coupled GCM simulations and found an increase in ENSO variability and a shift of the ENSO period towards longer periods when the Indian Ocean variability is suppressed. This is also in agreement with a conceptual model study of Jansen et al. (2009). Wu and Kirtman (2004) and Kug and Kang (2006) also find, based on observations, that an influence of the SST anomalies in the Indian Ocean on ENSO exists. However, these studies are based on relative short time series. Another problem is the difficulty in quantifying the influence of the Indian Ocean on ENSO, because the Indian Ocean is primarily influenced by ENSO, and the feedback of the Indian Ocean onto ENSO is only a secondary effect. Some studies also propose an influence of the Atlantic Ocean on ENSO. Dommenget et al. (2006) found in their coupled GCM study a shift in ENSO period towards longer periods and an increase in 3 ENSO variability if the SST variability in the tropical Atlantic Ocean is suppressed. Jansen et al. (2009) found in their conceptual model study only very little changes in the frequency and variability of ENSO by including a feedback from the Atlantic Ocean. However, an inclusion of the Atlantic Ocean can improve the ENSO forecast skill of their simple conceptual model based on parameters from observations. Recently, Rodriguez-Fonseca et al. (2009) and Ding et al. (2011) suggested an influence of the Atlantic zonal mode on ENSO. Thus, many studies indicate possible influences of the tropical Indian and Atlantic Oceans on ENSO, but the exact mechanisms of these teleconnections are not yet fully understood and need further investigation. In this study we use a hybrid coupled model consisting of a full complexity atmospheric GCM coupled to the minimum complex ENSO recharge oscillator ocean model in the tropical Pacific to analyse the influences of the tropical Indian and Atlantic Oceans on ENSO and ENSO predictability The Model Simulations The model used for this study is the hybrid coupled model RECHOZ as described in Frauen and Dommenget (2010). The model consists of the atmospheric GCM ECHAM5 (Roeckner et al., 2003) coupled to the ENSO recharge oscillator ocean model based on Burgers et al. (2005) in the tropical Pacific and a simple single column mixed layer ocean model elsewhere (Dommenget and Latif, 2008). Despite the simplistic and, by construction, linear representation of the ENSO ocean dynamics in the RECHOZ model, it is able to correctly simulate the main statistical features of ENSO, such as the power spectrum, skewness and the seasonal phase locking. The advantage of this kind of model is that the SST in various regions can easily be controlled and the influences of different feedbacks on ENSO can be quantified in terms of changes in the recharge oscillator parameters. In particular, we can alter or eliminate initial SST or thermocline depth anomalies without causing dynamical shocks to the ocean locally or at remote regions. For comparison also a simple Monte Carlo reference model (REOSC-MC) is constructed in which the 4 84 85 recharge oscillator equations are integrated in time with white noise forcing instead of the ECHAM5 forcing The effects of decoupling the tropical Indian and Atlantic Oceans To study the interactions of the tropical Indian and Atlantic Oceans with ENSO three 500 years long sensitivity experiments were performed with the RECHOZ model in addition to a 500 years long control run (CTRL). In each of the sensitivity experiments the SST in a specific region is prescribed by monthly varying climatologies obtained from the control run. First the SST in the tropical Atlantic (30 S 30 N) is prescribed (NOAT), then the SST in the tropical Indian Ocean (30 S northern boundary) (NOIND), and finally the SST in both tropical oceans is prescribed (NOAI). The resulting power spectra of monthly mean NINO3 SST anomalies for the different experiments illustrate that decoupling the tropical Indian and/or Atlantic Ocean has a significant impact on the ENSO statistics (Fig.1a). Decoupling the tropical Atlantic Ocean leads to a 25% increase in NINO3 SST anomaly standard deviation while decoupling the tropical Indian Ocean not only leads to a 63% increase in NINO3 standard deviation but also shifts the period of ENSO towards longer periods. When both tropical oceans are decoupled the increase in the NINO3 standard deviation is of same magnitude but the shift in the period is even stronger. The results concerning the Indian Ocean are in agreement with Kug and Kang (2006), Dommenget et al. (2006), and Jansen et al. (2009). Only the increase in the NINO3 SST variability is stronger than in the previous studies. For the Atlantic Ocean the results are also in agreement with Jansen et al. (2009) while Dommenget et al. (2006) also find a shift in the period of ENSO by decoupling the tropical Atlantic. In the simple recharge oscillator ocean model the ENSO dynamics are completely defined by four prescribed dynamical parameters and two forcing terms (see Frauen and Dommenget (2010)). Since the atmospheric forcings from the ECHAM model are not just linear response and white noise the effective (resulting) parameters can be different from the original prescribed parameters. To quantify 5 the impact of decoupling the tropical Indian and/or Atlantic Oceans on ENSO, the changes in the effective recharge oscillator parameters can be analysed. Therefore, the model parameters are estimated by a linear regression method of the simulations output data as done for observational data in Burgers et al. (2005) and Jansen et al. (2009). No significant changes in the parameters are found for the NOAT experiment (see Table 1). For the NOIND and the NOAI experiments the changes in the parameters are of similar magnitude. Thus, in the following, only the changes in the NOIND experiment will be analysed. The strongest changes are found in the damping of the SST (a 11 ) and the coupling of the SST to the thermocline (a 12 ). Further the standard deviations of the residuals, which correspond to the noise forcing terms ξ 1 and ξ 2 are increased by 13% each. The implications of these changes in the parameters can be illustrated by including these parameters in the REOSC-MC model. Different years long experiments with the REOSC-MC model were performed in which the parameters resulting from the CTRL and NOIND experiments and combinations of these were used. First of all we note that the shift in the power spectrum from CTRL to NOIND is about the same as in the RECHOZ model, indicating that the REOSC-MC model dynamically behaves in a similar way (see Fig.1b). Further we find that the changes in the coupling of the thermocline to the SST (a 21 ) and the damping the thermocline (a 22 ) have no influence on the frequency and amplitude of ENSO. The changes in the standard deviations of the noise forcing terms ξ 1 and ξ 2, where ξ 2 represents the zonal wind stress forcing and ξ 1 a combination of the zonal wind stress and the net heat flux forcing, lead to increased variability on all time scales. The decrease of the damping of the SST (a 11 ) leads to increased variability on ENSO time scales. Only the change in the coupling of the SST to the thermocline depth (a 12 ) leads to decrease of variability on time scales up to four years and shifts the peak of the spectrum towards longer periods. 6 The effects of the initial SST of the tropical Indian and Atlantic Oceans in ENSO Forecast experiments To study the influences of the tropical Indian and the tropical Atlantic Oceans on ENSO predictability two sets of perfect model forecast experiments were performed. First, based on the CTRL run, every 5 years 12 months long forecast experiments were started on January 1 st and July 1 st. For each of the 200 forecast start dates four ensemble members with slightly perturbed initial conditions were calculated. In addition to the control experiments (F-CTRL) again three sensitivity experiments were performed in which the initial conditions in the tropical Indian Ocean (F-NOIND), the tropical Atlantic Ocean (F- NOAT) and both tropical oceans (F-NOAI) were set to climatological values. The anomaly correlation skill of the forecasts decreases significantly for all three sensitivity experiments compared to the F- CTRL experiment. The strongest reduction in the forecast skill is found for the F-NOAT experiment which is outside the 90% confidence interval of the F-CTRL experiment for the months 6 and 7 after the restart while the reduction of the forecast skill is smaller for the F-NOIND experiment. These results are in good agreement with Jansen et al. (2009) although in our experiments the effects are much stronger. To study possible mechanisms leading to an enhanced forecast skill if the Indian and/or Atlantic Oceans initial states are included another set of experiments was performed. Therefore, 8 events were chosen from the CTRL time series in which a strong positive SST anomaly in the tropical Pacific was followed by positive anomalies in the tropical Indian and Atlantic Oceans. The mean SST anomalies for these 8 events can be seen in figure 2 (c). For each of the 8 events 6 different forecast experiments were performed with four ensemble members each. In addition to a control experiment (E-CTRL) and experiments with the initial conditions in the tropical Indian Ocean (E-NOIND), the tropical Atlantic Ocean (E-NOAT) and both tropical oceans (E-NOAI) set to climatological values two experiments were performed in which the initial conditions in the tropical Pacific (E-NOPA) and all tropical oceans 7 (E-NOTO) were set to climatology. Firth of all, we find that in all forecast that exclude the initial SST anomalies of the Indian and/or Atlantic Oceans the positive NINO3 SST anomaly persists longer than in the control simulation, which is consistent with the findings of Kang and Kug (2006), Dommenget et al. (2006), and Jansen et al. (2009), that positive (negative) SST anomalies in both tropical oceans force a negative (positive) SST trend in the tropical Pacific (see Fig. 2c). This is also supported by the E- NOPA experiment, where the forcing from the tropical Indian and/or Atlantic Oceans leads to negative NINO3 SST anomalies. The E-NOTO experiment further illustrates that none of the initial condition outside the tropical oceans leads to any significant NINO3 SST anomalies. To understand the reduction of the forecast skill when the Indian and/or Atlantic Oceans are decoupled one can have a look at the evolution of the equatorial Pacific zonal wind stress anomalies in the different experiments (Fig.3). In the E-CTRL experiment one finds a positive zonal wind stress anomaly around the date line which lasts until month 3. From the west develops a negative zonal wind stress anomaly which extends across the date line in month 4. In the experiment without the Atlantic Ocean the positive wind stress anomaly is enhanced and lasts longer. The same is true for the experiment without the Indian Ocean. In this experiment also the negative anomaly in the west in the first months vanishes Conclusion We used the hybrid coupled model RECHOZ to study the influences of the tropical Indian and the tropical Atlantic Oceans on ENSO and ENSO predictability. First, a set of experiments was performed in which the SSTs in the tropical Indian and/or Atlantic Oceans were prescribed with monthly varying climatologic values. Second, a set of perfect model forecast experiments was performed in which the initial conditions in the different basins were set to climatology. Decoupling the tropical Atlantic Ocean only has little influence on the amplitude and frequency of ENSO. Without the Atlantic Ocean the NINO3 SST anomaly variance is slightly increased but no 8 significant changes in the parameters of the recharge oscillator are found. For the Indian Ocean however a strong influence on the amplitude and frequency of ENSO are found. Decoupling the Indian Ocean leads to an increase in NINO3 SST anomaly variability and shifts the period of ENSO towards longer time scales. This can also be seen in the parameters of the recharge oscillator. The damping of the SST is halved and the coupling between the SST and the thermocline is reduced. The results for the Indian Ocean are in good agreement with Kug and Kang (2006), Dommenget et al. (2006), and Jansen et al. (2009). Also for the Atlantic Ocean the results are in good agreement with Jansen et al. (2009) while Dommenget et al. (2006) find also a shift in the ENSO period when decoupling the tropical Atlantic. The tropical Indian as well as the tropical Atlantic Ocean has an influence on the predictability of ENSO. Removing the initial SST anomalies in these basins leads to a reduced forecast skill for NINO3 SST anomalies due to changes in the atmospheric circulation and thus changes in the central Pacific zonal wind stress anomalies. A warm (cold) SST anomaly in the Indian or Atlantic Ocean induces negative (positive) zonal wind stress anomalies over the central Pacific which then leads to cold (warm) SST anomalies in the Pacific. However, although the Indian Ocean has strong influences on the amplitude and frequency of ENSO, its influence on the predictability of ENSO is smaller than the influence of the Atlantic Ocean. Overall it seems that the main aspects of the tropical Indian and Atlantic Oceans influence on ENSO can be understood relatively well in the framework of the simple linear ocean dynamics of the RECHOZ model Acknowledgements We like to thank Mojib Latif and Noel Keenlyside for helpful remarks and discussions. This work was supported by the Deutsche Forschungsgemeinschaft (DFG) through project LA 871/5-2 and by Australian Research Council s Center of Excellence in Climate System Sciences References Annamalai, H., S. P. Xie, J. P. McCreary, and R. Murtugudde (2005), Impact of the Indian Ocean Sea Surface Temperature on Developing El Niño. J. Climate, 18, Burgers, G., F.-F. Jin, and G. J. van Oldenborgh (2005), The simplest ENSO recharge oscillator. Geophys. Res. Lett., 32, L13706, doi: /2005GL Ding, H., N. Keenlyside, and M. Latif (2011), Impact of the equatorial Atlantic on the El Niño Southern Oscillation. Clim. Dyn., doi: /s y. Dommenget, D., V. Semenov, and M. Latif (2006), Impacts of the tropical Indian and Atlantic Oceans on ENSO. Geophys. Res. Lett., 33(L11701), doi: /2006GL Dommenget, D., and M. Latif (2008), Generation of hyper climate modes. Geophys. Res. Lett., 35, L02706, doi: /2007GL Enfield, D. B. and D. A. Mayer (1997), Tropical Atlantic sea surface temperature variability and its relation to El Niño-Southern Oscillation. J. Geophys. Res., 102(C1), , doi: /96JC Frauen, C. and D. Dommenget (2010), El Niño and La Niña amplitude asymmetry caused by atmospheric feedbacks. Geophys. Res. Lett., 37, L18801, doi: /2010GL Jansen, M. F., D. Dommenget, and N. Keenlyside (2009), Tropical atmosphere-ocean interactions in a conceptual framework. J. Climate, 22, Jin, F.-F. (1997), An equatorial recharge paradigm for ENSO: Part 1: Conceptual model. J. Atmos. Sci., 54, Latif, M. and T. P. Barnett (1995), Interactions of the Tropical Oceans. J. Climate, 8, Liu, Z. (2002), A Simple Model Study of ENSO Suppression by External Periodical Forcing. J. Climate, 15, Kug, J.-S. and I.-S. Kang (2006), Interactive Feedback between ENSO and the Indian Ocean. J. Climate, 19, Philander, S. G. H. (1983), El Niño Southern Oscillation phenomena. Nature, 302, Philander, S. G. H. (1985), El Niño and La Niña. J. Atmos. Sci., 42, Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan (2003), Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108(D14), 4407, doi: /2002JD Rodriguez-Fonseca, B., I. Polo, J. Garcia-Serrano, T. Losada, E. Mohino, C. R. Mechoso, and F. Kucharski (2009), Are Atlantic Niños enhancing Pacific ENSO events in recent decades? Geophys. Res. Lett., 36(L20705), doi: /2009GL Roeckner, E., et al. (2003), The atmospheric general circulation model ECHAM5. Part I: Model description. Rep. 349, Max-Planck-Inst. for Meteorol., Hamburg, Germany. Smith, N. (1995), An improved system for tropical ocean sub-surface temperature analyses. J. Atmos. Oceanic Technol., 12, Wu, R. and B. P. Kirtman (2004), Understanding the Impacts of the Indian Ocean on ENSO Variability in a Coupled GCM. J. Climate, 17, Yu, J.-Y., C. R. Mechoso, J. C. McWilliams, and A. Arakawa (2002), Impacts of the Indian Ocean on the ENSO cycle. Geophys. Res. Lett., 29(1204), doi: /2001GL List of Figures 1 (a) Spectra of monthly mean NINO3 SST anomalies for the CTRL run (black), the NOIND experiment (red), the NOAT experiment (blue), and the NOAI experiment (green). The grey shading indicates the 90% confidence interval of the CTRL run spectrum. The black vertical lines indicate periods of 1, 2, 3 and 4 years. (b) Spectra of monthly mean NINO3 SST anomalies for the REOSC-MC experiments with the parameters obtained from the CTRL run (bla
Related Search
Advertisements
Advertisements
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks