El Niño–Southern Oscillation Simulation at 6000 Years before Present with the MRI-CGCM2.3: Effect of Flux Adjustment

Akio Kitoh Meteorological Research Institute, Tsukuba, Japan

Search for other papers by Akio Kitoh in
Current site
Google Scholar
PubMed
Close
,
Tatsuo Motoi Meteorological Research Institute, Tsukuba, Japan

Search for other papers by Tatsuo Motoi in
Current site
Google Scholar
PubMed
Close
, and
Shigenori Murakami Frontier Research Center for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokohama, Japan

Search for other papers by Shigenori Murakami in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Modulation of El Niño–Southern Oscillation at the mid-Holocene [6000 yr before present (6 ka)] is investigated with a coupled ocean–atmosphere general circulation model. The model is integrated for 300 yr with 6-ka and present (0 ka) insolation both with and without flux adjustment, and the effect of flux adjustment on the simulation of El Niño is investigated. The response in the equatorial Pacific Ocean in 6 ka is in favor of weaker El Niño variability resulting from lowered sea surface temperature (SST) and a more diffuse thermocline. Atmospheric sensitivity in 6 ka is larger than that in 0 ka because of increased trade winds, while oceanic sensitivity in 6 ka is weaker than that in 0 ka, resulting from destabilization of the upper ocean, both in the flux- and non-flux-adjusted experiments. However, the use of flux adjustment causes a difference in the total response. El Niño variability in 6 ka does not change much from that in 0 ka with the flux-adjusted case, while the 6-ka El Niño variability is weaker without flux adjustment. Because the observed proxy data suggest weaker El Niño variability in the mid-Holocene, the non-flux-adjusted version gives a more reasonable response despite a larger bias in its basic states, implying that nondistortion of sensitivity to forcing is more important.

Corresponding author address: Akio Kitoh, Climate Research Department, Meteorological Research Institute, 1-1 Nagamine, Tsukuba, Ibaraki 305-0052, Japan. Email: kitoh@mri-jma.go.jp

Abstract

Modulation of El Niño–Southern Oscillation at the mid-Holocene [6000 yr before present (6 ka)] is investigated with a coupled ocean–atmosphere general circulation model. The model is integrated for 300 yr with 6-ka and present (0 ka) insolation both with and without flux adjustment, and the effect of flux adjustment on the simulation of El Niño is investigated. The response in the equatorial Pacific Ocean in 6 ka is in favor of weaker El Niño variability resulting from lowered sea surface temperature (SST) and a more diffuse thermocline. Atmospheric sensitivity in 6 ka is larger than that in 0 ka because of increased trade winds, while oceanic sensitivity in 6 ka is weaker than that in 0 ka, resulting from destabilization of the upper ocean, both in the flux- and non-flux-adjusted experiments. However, the use of flux adjustment causes a difference in the total response. El Niño variability in 6 ka does not change much from that in 0 ka with the flux-adjusted case, while the 6-ka El Niño variability is weaker without flux adjustment. Because the observed proxy data suggest weaker El Niño variability in the mid-Holocene, the non-flux-adjusted version gives a more reasonable response despite a larger bias in its basic states, implying that nondistortion of sensitivity to forcing is more important.

Corresponding author address: Akio Kitoh, Climate Research Department, Meteorological Research Institute, 1-1 Nagamine, Tsukuba, Ibaraki 305-0052, Japan. Email: kitoh@mri-jma.go.jp

1. Introduction

El Niño–Southern Oscillation (ENSO) is among the most energetic source of interannual variability of atmosphere–ocean system in the Tropics, and its influence can reach midlatitudes as well as other ocean basins through teleconnection (e.g., Lau and Nath 1996, 2003). The amplitude of El Niño can change by various factors (e.g., Meehl et al. 2001; Yeh and Kirtman 2005). How climate changes will modify the behavior of ENSO is one of the important questions in future climate projections, but disagreement on future ENSO changes by increasing atmospheric concentrations of greenhouse gases with various models indicates the need for further studies of various aspects (e.g., van Oldenborgh et al. 2005; Guilyardi 2006; Meehl et al. 2006; Merryfield 2006). Paleoclimate studies show evidence of past changes in ENSO variability. An investigation of ENSO with GCMs under different climate forcing gives us good insight to the mechanism of ENSO variability and its changes.

Because of the differences in the earth’s orbital parameters, climate at the mid-Holocene is characterized by an amplification of the seasonal cycle in the Northern Hemisphere and a strengthened summer monsoon activity. Climate model simulations for 6000 yr before present (6 ka) by the Paleoclimate Modeling Intercomparison Project (PMIP) have looked at those aspects by coordinated climate modeling studies (Joussaume and Taylor 1995; Joussaume et al. 1999; PMIP 2000; Braconnot et al. 2000b). Models show an enhanced seasonal variation in surface temperature and a northward shift of the African and Asian summer monsoon rain area. However, data–model comparisons show that models underestimate both the northward shift in the African monsoon belt and the magnitude of the precipitation required to fit into paleoenvironmental data, suggesting the importance of ocean–atmosphere coupling and vegetation feedback.

The PMIP simulations described above used either an atmospheric GCM or an atmosphere–mixed layer ocean model; hence, they do not include ocean dynamics explicitly. Several paleoclimate simulations for the mid-Holocene period using a fully coupled atmosphere–ocean GCM (AOGCM) have been reported (Hewitt and Mitchell 1998; Otto-Bliesner 1999; Bush 1999; Braconnot et al. 2000a; Liu et al. 2000, 2003; Kitoh and Murakami 2002). Because we use AOGCMs to project future climate changes under the scenarios of future greenhouse gases and aerosols concentration, the same AOGCMs should be used in the paleoclimate simulations to demonstrate the effectiveness of these models to reproduce past climate under different forcing and evaluate simulated results against observed proxy records. These coupled model simulations unanimously emphasize the importance of ocean feedback in the climate response to orbital forcing. Braconnot et al. (2004) and Zhao et al. (2005) analyzed post-PMIP1 AOGCM simulations of the mid-Holocene and found the robust differences between AOGCM and AGCM simulations for the African and the Indian summer monsoons. The ocean feedback enhances the African monsoon and shifts the belt of maximum precipitation farther north than in the basic PMIP simulations (Braconnot et al. 2004). The length of the African monsoon increases because of the late-summer positive sea surface temperature (SST) anomalies in the tropical Atlantic. This feature results from a wind-evaporation feedback and changes in the mixed layer depth and increased Ekman transport in the Atlantic that enhance the insolation forcing at the mid-Holocene (Zhao et al. 2005). In the recently started second phase of the PMIP (PMIP2; Braconnot et al. 2003; Crucifix et al. 2005), AOGCMs are now a common tool for paleoclimate simulations, and model–model and model–data intercomparisons of the PMIP2 models are just getting started.

Paleoclimate studies using proxy data revealed a weaker ENSO variability in the mid-Holocene (e.g., Gagan et al. 1998; Tudhope et al. 2001; Cole 2001; McGregor and Gagan 2004; Rein et al. 2005). There are some modeling approaches that investigate ENSO variability in the mid-Holocene using comprehensive AOGCMs. The modeling study by Liu et al. (2000) shows reduced El Niño variability in the Holocene, while Otto-Bliesner (1999) obtained comparable SST variability in the eastern equatorial Pacific at 6 ka compared to that at the present. Using a simple coupled model, Clement et al. (2000) simulated weaker ENSO activity in the mid-Holocene than in the present. Otto-Bliesner et al. (2003), using the National Center for Atmospheric Research (NCAR) Climate System Model (CSM), obtained weaker ENSO for the Holocene and stronger ENSO for the Last Glacial Maximum (LGM). They attributed weaker ENSO to a strengthening of the tropical Pacific trade winds and a weakening of the tropical thermocline. A recent modeling study with CCSM3 by Otto-Bliesner et al. (2006) showed weaker ENSO variability both for the Holocene and LGM. Such paleo-ENSO simulation studies would give us good insight on the mechanism of ENSO variability and its changes, and clues for future ENSO changes by anthropogenic greenhouse gases increase.

In this paper, ENSO characteristics of the latest version of the Meteorological Research Institute (MRI) coupled GCM, which has also been used for the future climate projections, are investigated. There are two types of AOGCMs—those with or those without flux adjustment. Flux adjustments are employed to give a stable and realistic simulation of the present surface climate. Use of flux adjustment alleviates systematic errors (bias) in the models, but may distort ocean–atmosphere feedbacks (Neelin and Dijkstra 1995) and their sensitivity to changed forcing (McAvaney et al. 2001). However, models without flux adjustment, more or less, reveal a cold bias over the equatorial Pacific in their control (present day) simulations, and this climate drift (bias) may affect the model’s response to forcing. Our motivation in this study is to simulate the 6-ka climate by the MRI-Coupled General Circulation Model version 2.3 (CGCM2.3) with and without flux adjustment and to investigate its sensitivity on ENSO modulation.

We describe the model and experimental setup in section 2. Simulation of preindustrial climate is given in section 3, while simulation of 6-ka mean climate is shown in section 4. A comparison of simulated ENSO between the preindustrial control and the 6-ka runs under different situations with or without flux adjustment is investigated in section 5. Finally, conclusions are given in section 6.

2. Experiments

a. Model

The model used for the present study is the latest version of the MRI climate model, MRI-CGCM2.3 (Yukimoto et al. 2006a). The atmospheric component has horizontal T42 resolution (an approximately 280-km transform grid) and 30 layers in vertical, with the top at 0.4 hPa. The version of the model used in this study exhibits a climate sensitivity of 2.6 K (global annual mean surface air temperature change for a doubling of atmospheric CO2). The oceanic component is a Bryan–Cox-type global ocean general circulation model (OGCM). The horizontal resolution is 2.5° longitude and 2.0° latitude poleward of 12°S and 12°N, with finer resolution up to 0.5° near the equator. The vertical 23 levels are unevenly placed between the surface and the deepest bottom at 5000 m, which forms a realistic topography. The sea ice model calculates compactness and thickness prognostically based on thermodynamics and horizontal advection and diffusion. The advection velocities are determined from the surface ocean current multiplied by an empirical constant (set to one-third at present). The detailed model description and model performance in the control simulation are described in Yukimoto et al. (2006a), while analyses of the historical twentieth-century simulation and the twenty-first-century scenario experiments can be found in Yukimoto et al. (2006b).

b. 0- and 6-ka runs

The preindustrial control (0 ka) run and the mid-Holocene (6 ka) run are performed based on the PMIP2 (Crucifix et al. 2005) protocol (information available online at http://www-lsce.cea.fr/pmip2/). The 0-ka run uses preindustrial values in greenhouse gases concentration (280 ppm in CO2, 760 ppb in CH4, and 270 ppb in N2O). A solar constant of 1365 W m−2 is used. In the 6-ka run, only the orbital parameters are changed to the values in 6 ka (Berger 1978), while keeping the greenhouse gases concentration values the same as in the 0-ka run. The same prescribed ice sheets (Greenland and Antarctica) and vegetation cover are used for 0 and 6 ka. We therefore investigate the impact of seasonal-latitudinal changes in insolation on climate in this experiment. Because of the insolation change, the Northern Hemisphere gets more insolation in summer and less in winter. Both the 0- and 6-ka runs have been performed either with flux adjustment or without flux adjustment (Table 1). Flux adjustment values are obtained in the calibration run where sea surface temperature and sea surface salinity (SSS) are restored to the observed present-day climatology. In addition, we employed an adjustment for wind stress in the equatorial region to reproduce a realistic climatological thermocline along the equator. Details are shown in Yukimoto et al. (2006a). Hereafter, the 0- and 6-ka runs with flux adjustment are denoted as 0ka_fa and 6ka_fa, while those without flux adjustment are 0ka_nfa and 6ka_nfa, respectively.

The model has been integrated for 300 yr for each experiment. Time series of the global annual mean surface air temperature for each experiment is plotted in Fig. 1. There is very little drift in the global annual mean surface air temperature in the control run with flux adjustment (0ka_fa). The 6ka_fa run shows a slightly cooler temperature than in 0ka_fa during the entire period of integration. The last 100-yr mean of the difference (6ka_fa minus 0ka_fa) is −0.14°C. The global annual mean surface air temperature in the non-flux-adjusted runs drops about 2.5°C in first 50 yr. They are in near equilibrium after a 200-yr integration. The last 100-yr mean of the difference in non-flux-adjusted runs (6ka_nfa minus 0ka_nfa) is larger than that in flux-adjusted runs, and is about −0.47°C. The global annual mean SST change in 6 ka for flux- and non-flux-adjusted runs is −0.15° and −0.36°C, respectively. In the following, we use the data in the last 100 yr of the integration (years 201–300) for the analyses.

3. Simulation of the present climate

The annual mean SST and precipitation in each control run (0ka_fa and 0ka_nfa) are displayed in Fig. 2, together with the corresponding observations. The Hadley Centre Sea Ice and SST dataset (HadISST; Rayner et al. 2003) for the period of 1971–2000 and the Xie and Arkin (1996) precipitation for the period of 1979–97 are used for the observations. Thanks to flux adjustments, the horizontal distribution of the annual mean SST in 0ka_fa (Fig. 2b) is close to that of the observations (Fig. 2a). The simulated SST is about 1°C lower in the tropical western Pacific and the northern Indian Ocean compared to the observations. It should be noted that because we use the preindustrial value of greenhouse gases concentration for 0-ka runs, simulated SST should be lower than the present-day values. The geographic distribution of annual mean precipitation in 0ka_fa (Fig. 2e) captures the observed characteristics well, with the global mean value (2.63 mm day−1) comparable to that of the observations (2.64 mm day−1). Simulated precipitation in the western Pacific is underestimated on the equator and is overestimated in the northern intertropical convergence zone (ITCZ) and around the south Pacific convergence zone (SPCZ). These drawbacks appear in all seasons (Yukimoto et al. 2006a).

Simulated global annual mean SST in 0ka_nfa is 1.5°C lower than that in 0ka_fa. Cold bias is amplified in the surface air temperature where the non-flux-adjusted run (0ka_nfa) is 2.4°C lower than the flux-adjusted run (0ka_fa) (see Fig. 1). The difference in annual mean SST between 0ka_nfa and 0ka_fa is large in the North Atlantic Ocean and in the central equatorial Pacific region (Fig. 2c). The cold bias in the latter region is common to most non-flux-adjusted GCMs (Davey et al. 2002). In these models, the westward shift of the warm water pool and the Walker circulation is evident, associated with the extended equatorial upwelling region, resulting in the cold bias in SST. The longitude of the large zonal SST gradient in 0ka_nfa is located west of that in 0ka_fa, which is consistent with the westward displacement of surface easterlies. Associated with the cold bias in SST, the global annual mean precipitation in 0ka_nfa (2.44 mm day−1) is less than that in 0ka_fa. A double ITCZ in the tropical eastern Pacific and a very dry region along the equatorial Pacific reaching to the western Pacific are also noted. The cold bias in the non-flux-adjusted run results in more extensive sea ice. This is evident in Northern Hemisphere spring season, when the northwestern Pacific and the North Atlantic north of about 45°N is covered by sea ice in March in 0ka_nfa, while that in 0ka_fa is much closer to the observations (Yukimoto et al. 2006a).

4. Simulation of the 6-ka mean climate

Figure 3 displays the seasonal cycle of the simulated change in zonal mean surface air temperature in 6 ka compared to that in 0 ka. Incoming solar radiation at the top of the atmosphere in 6 ka intensifies the seasonal distribution of insolation in the Northern Hemisphere by about 5%, while weakens it in the Southern Hemisphere (Berger 1978). Surface air temperature responses in 6 ka in both the flux- (Fig. 3a) and non-flux- (Fig. 3b) adjusted runs roughly follow the difference in insolation with a lag of about 1 month, which is comparable to a result obtained by Hewitt and Mitchell (1998), while Braconnot et al. (2000a) obtained a 2–3-month lag.

Comparison between Figs. 3a and 3b reveals a larger cooling in the Northern Hemisphere high latitudes in winter and spring in the non-flux- than in the flux-adjusted runs. In the former, there is a cooling of more than 2.5°C between 50° and 60°N in late winter and early spring. This larger temperature decrease in 6 ka in the non-flux-adjusted run is associated with the large extension of the sea ice area in the North Pacific in this season. An anomalously large sea ice extent in the non-flux-adjusted control run (0ka_nfa) becomes more exaggerated by the lower-than-present insolation in 6 ka in this season. Extensive sea ice spreads less saline upper-ocean water in the North Pacific in 0 ka, which is effective, responding to negative insolation anomalies, resulting in more sea ice and large negative temperature anomalies in 6 ka. These responses are not seen in the flux-adjusted run.

Figures 4a, 4c and 4e display the geographical distributions of the 6-ka changes in the annual mean surface air temperature, precipitation, and SSS in the flux-adjusted run. Regions with statistical significance at 95% level are shaded. The surface air temperature and precipitation responses in 6 ka are qualitatively similar to those obtained in the previous studies. As described already regarding Fig. 1, the annual mean surface air temperature decreases in 6 ka. This is seen in the low- and midlatitudes, while there is positive temperature anomaly in the Arctic region. These changes are mainly due to insolation difference by orbital changes where the tropical atmosphere receives about 1 W m−2 less in the 6-ka period compared to the present, while the polar regions receive about 4 W m−2 more (Liu et al. 2003). Large positive surface air temperature anomalies and positive SSS anomalies are evident in the Barents Sea, where sea ice retreat occurs throughout the year, with largest retreat during summer and fall. Annual mean precipitation anomalies are positive over a belt from the Sahel to India. Over Africa, this is accompanied by negative precipitation anomalies to the south, corresponding to a northward shift and an intensification of the summer monsoon rainfall. More rainfall in India is contrasted to less rainfall in Southeast Asia. Surface temperature anomalies are negative over the region with positive precipitation anomalies where the ground is wetter and there is more cloudiness. A contrast in SSS change is distinct between the western tropical Pacific and the Indian Ocean. The lower SSS in 6 ka is simulated in the Bay of Bengal and the Arabian Sea because of larger precipitation and runoff. This is contrasted with higher salinity in the western tropical Pacific where precipitation has decreased. This contrast in SSS change has also been simulated by a previous study (Kitoh and Murakami 2002).

The geographical pattern of the responses in the non-flux-adjusted run (Figs. 4b, 4d and 4f) is qualitatively similar to that in the flux-adjusted run, except in the North Pacific and Arctic regions. The absolute value of the temperature change is larger in the non-flux-adjusted run than in the flux-adjusted run, particularly in the Pacific. A large negative temperature anomaly over the North Pacific (Fig. 4b) is associated with more extensive sea ice in 6 ka. Because of a large cold bias, the non-flux-adjusted control experiment (0ka_nfa) has more snow-covered area over the continent compared to that in the flux-adjusted one (0ka_fa). The 6-ka experiment produces more snow cover over the Eurasian continent in spring resulting from less insolation. This snow–albedo feedback is more effective in the non-flux-adjusted simulations because of their cold bias, resulting in larger surface air temperature decreases (Fig. 4b versus Fig. 4a). On the contrary, smaller temperature anomalies in the Barents Sea than in the flux-adjusted run may also be due to colder climate in the non-flux-adjusted run, where sea ice change is seen only in the July–October season (not shown). The non-flux-adjusted run also shows an increase in the Afro-Indian monsoon rainfall. The responses in the Indian monsoon rainfall are almost similar to each other, with the June–September (JJAS) Indian monsoon rainfall anomalies of 1.01 and 0.84 mm day−1 in the flux- and non-flux-adjusted runs, respectively. On the other hand, precipitation anomalies over North Africa extend more northward in the non-flux-adjusted run (Fig. 4d). A clear SSS contrast between the Indian Ocean and the South China Sea is again seen in Fig. 4f, corresponding to larger precipitation and runoff in the Indian monsoon region and negative precipitation anomalies in the western tropical Pacific. Large negative SSS anomalies in the North Pacific come from the melting of extensive sea ice there.

5. ENSO

In this section, we investigate the ENSO modulation in 6 ka in our experiments. As indicated in section 1, most AOGCMs previously published simulated a decrease in El Niño amplitude in 6 ka compared to that in 0 ka.

a. Flux-adjusted run

We first investigate the reproducibility of ENSO in the control run with flux adjustment. Figure 5a shows the time series of the leading EOF mode of the near-global (50°S–50°N) SST in the 0ka_fa run. This explains 17.4% of the total variance and corresponds to the model El Niño. There are positive SST anomalies over the equatorial central and eastern Pacific, the Indian Ocean, and the Atlantic Ocean (Fig. 5b). Negative SST anomalies are found over the North Pacific and the Southern Oceans. Figures 5c and 5d show the associated regression fields of the precipitation and the wind vector at 850 hPa. In these figures, regions with statistically significant correlation with the first EOF time series are shaded. It is immediately found that the large precipitation anomalies are associated with the SST anomalies in the central tropical Pacific, accompanied by negative precipitation anomalies over the Maritime Continent and the Philippines Sea region, implying an eastward displacement of the rainfall center in model El Niño years. Westerly wind anomalies prevail over the entire equatorial Pacific, while the Indian Ocean is covered by easterly wind anomalies. There are significant cyclonic circulation anomalies in the North Pacific. These features are in agreement with the observed features in El Niño years (Ropelewski and Halpert 1989). In summary, the flux-adjusted model reproduces a reasonable El Niño and associated precipitation and atmospheric circulation changes.

Figures 5e–5h show the leading EOF mode in SST and associated precipitation and circulation fields in the 6-ka run with flux adjustment. The leading EOF mode in the 6ka_fa run explains 19.6% of the total variance, and its SST spatial pattern as well as regressed precipitation and 850-hPa winds are very similar to the 0-ka counterparts.

Seasonal cycle of the standard deviations of monthly Niño-3.4 SST anomalies is displayed in Fig. 6a both for 0ka_fa and 6ka_fa. Observation for the modern (not for the preindustrial) climate is also plotted. Here the HadISST1.1 (Rayner et al. 2003) for the period of 1901–2000 is used. Both the 0ka_fa and 6ka_fa runs show that the seasonal cycle of the Niño-3.4 SST variability is similar to that of the observation, both in its annual amplitude and phase with larger variability in the latter half of the calendar year, although the simulated magnitude is slightly smaller than that of the observation. Standard deviations of monthly Niño-3.4 SST anomalies are 0.695°C in 0 ka and 0.682°C in 6 ka, respectively. The Niño-3.4 SST variability in 6 ka tends to be smaller in the first half and larger in the latter half of the year than that in 0 ka, but its difference is not statistically significant according to the F test. Therefore, the flux-adjusted version of our model shows almost no change in ENSO at 6 ka.

It is well known that the Indian summer (June–September) monsoon rainfall (IMR) becomes below normal in El Niño years and becomes above normal in La Niña years. This negative correlation between the SST anomalies over the equatorial central/eastern Pacific and the IMR is called the ENSO–monsoon relationship (e.g., Webster et al. 1998). Here we calculated lag correlations between IMR and Niño-3.4 SST anomalies. For observations, all-India JJAS monsoon rainfall data archived at the Indian Institute of Tropical Meteorology [IITM; Indian Regional/Subdivisional Monthly Rainfall Dataset (IMR), which is available online at http://www.tropmet.res.in/] and the HadISST1.1 (Rayner et al. 2003) are used. The observed lagged relationship during 1901–2000 (a thin solid line in Fig. 7a) shows significantly negative correlations during summer through following winter when Niño-3.4 SST lags IMR for two seasons (Yasunari 1990).

The ENSO–monsoon relationship in the flux-adjusted runs is compared with the observations in Fig. 7a. The lag correlation in the model agrees well with the observed ENSO–monsoon relationship in the sense that the IMR and Niño-3.4 SST show significantly negative correlations during concurrent and succeeding seasons. The 6-ka result tends to have larger negative correlations in early part of year 0 and smaller correlations in the latter half of year 0 and the beginning of year 1. The difference in correlation coefficients between 6 and 0 ka is statistically significant during February–April of year 0 and in February of year 1. This slightly earlier evolution may be related to differences in the seasonal cycle in monthly Niño-3.4 SST standard deviation (Fig. 6a).

b. Non-flux-adjusted run

Model ENSO in the non-flux-adjusted runs is quite different from the flux-adjusted one. Figure 8 shows the leading EOF mode in SST and associated regression fields in the 0ka_nfa and 6ka_nfa runs. The leading EOF mode explains 10.3% and 7.3% of the total variance in the 0ka_nfa run and 6ka_nfa run, respectively. These are substantially smaller than the flux-adjusted ones. The spatial pattern of the leading SST EOF mode in 0ka_nfa (Fig. 8b) shows a uniform distribution in the entire equatorial Pacific. The polarity of the SST anomalies in the Indian Ocean is the same sign in the tropical Pacific as in observed ENSO events and in the flux-adjusted runs, but its correlation is weaker. The Indian Ocean SST anomalies in ENSO events have been understood to be a result of the atmospheric bridge response through modified surface wind strength (e.g., Lau and Nath 2003). The weaker atmospheric bridge response in the Indian Ocean in this run can be understood from weak wind anomalies (Fig. 8d), which are linked with weak precipitation anomalies in the Pacific associated with model ENSO (Fig. 8c). A very weak precipitation anomaly over the equator and a double-ITCZ-like anomaly in the model ENSO are related to a distorted climatological precipitation field in the base state of this non-flux-adjusted run (Fig. 2f). Cyclonic circulation anomalies and associated negative SST anomalies in the North and South Pacific are reproduced well. The temporal phase in the ENSO–monsoon relationship (Fig. 7b) in 0ka_nfa is distorted compared to the observation and the flux-adjusted ones, although significantly negative concurrent correlation is reproduced well. Apparently, the flux-adjusted case captures the observed characteristics of ENSO (magnitude, spatial pattern, and ENSO–monsoon relationship) better than the non-flux-adjusted case.

The magnitude of 6 ka ENSO in the non-flux-adjusted run becomes weaker than the 0-ka counterpart. Standard deviations of monthly Niño-3.4 SST anomalies are 0.631°C in 0ka_nfa and 0.536°C in 6ka_nfa, respectively. The seasonal cycle of the interannual standard deviation of the Niño-3.4 SST anomalies (Fig. 6b) reveals statistically significant differences in May and October–January. Spatial patterns of SST, precipitation, and 850-hPa winds in 6ka_nfa (Figs. 8f–h) are more or less similar to those in 0ka_nfa (Figs. 8b–d). However, as the explained variance becomes smaller in 6ka_nfa, the spatial extent of SST anomalies with significant correlations (correlation coefficient greater than 0.2 is significant at 95% level) becomes narrower, particularly in the Indian Ocean. Significant precipitation anomalies and 850-hPa wind anomalies also appear only in the vicinity of equatorial Pacific. The concurrent ENSO–monsoon relationship still holds in 6ka_nfa (Fig. 7b), although it does not appear in Fig. 8g where all of the season’s data are used. There are significant differences in lagged correlation coefficients between 0ka_nfa and 6ka_nfa during October of year −1 to April of year 0 (Fig. 7b).

The weaker ENSO activity in 6 ka compared to that in 0 ka in the non-flux-adjusted runs is contrasted with those in the flux-adjusted runs. Possible reasons of these differences will be elaborated in the following subsection.

c. Possible mechanism of a different response

As demonstrated by previous studies (e.g., Hu et al. 2004; Zelle et al. 2005), the mean state change plays a key role in determining the El Niño amplitude through change in sensitivity of SST variability to surface wind stress. As shown in section 3, the non-flux-adjusted runs have a cold bias in the equatorial Pacific SST, easterlies penetrating into the western Pacific, the longitude of large zonal SST gradient being located too westward.

In Fig. 6, we showed that the Niño-3.4 SST variability in 6 ka decreased significantly than that in 0 ka in the non-flux-adjusted run, but not in the flux-adjusted run. It was also shown that there are seasonal differences in these changes. In this subsection, we explore the mechanism of these changes. First, the seasonal cycle of the Niño-3.4 SST is plotted in Fig. 9b. The Niño-3.4 SST in 0ka_fa shows a similar seasonal cycle to the observation, which is a 100-yr average (1901–2000) of the HadISST1.1. It is noted that the 0ka_fa values are about 1°C lower than the observations throughout the year, a part of which may come from a difference between the preindustrial simulation and the twentieth-century observations. The cold bias in the non-flux-adjusted run is evident by lower Niño-3.4 SST in 0ka_nfa throughout the year compared to the flux-adjusted one. The Niño-3.4 SST in 0ka_nfa shows a much larger seasonal cycle than that in 0ka_fa, and has a large cold bias in September–December.

The only forcing in the 6-ka experiment is insolation changes resulting from the earth’s orbital changes with season and latitude. In the Tropics, insolation in 6 ka is larger than the present during June–October and smaller during November–April (Fig. 9a). The difference is about +20 W m−2 during July–September and −20 W m−2 during January–March. Responding to this anomalous forcing, with a lag of about 1–2 months, the tropical SST in 6 ka tends to decrease in the first half of the year and increase in the latter half of the year (Fig. 9b, also see Fig. 3). Because the annual mean difference in incoming solar radiation is about −1 W m−2 in the Tropics, the annual mean SST in 6 ka can be lower than that in 0 ka. Actually, the annual mean Niño-3.4 SST in 6ka_fa is lower than that in 0ka_fa by 0.30°C. Superposed on this annual mean difference, a sinusoidal SST anomaly with an amplitude of about 0.7°C, with a minimum in April and a maximum in September, results in the seasonal cycle of the Niño-3.4 SST in 6ka_fa. The non-flux-adjusted run has a colder bias in its preindustrial run (0ka_nfa), and a cooling in 6 ka (–0.62°C) is larger than the flux-adjusted one. Because of this large annual mean cooling and smaller magnitude of annual cycle response (about 0.4°C), the Niño-3.4 SST in 6ka_nfa does not become higher than that in 0ka_nfa during the latter half of the year despite positive orbital solar forcing. These SST changes are consistent with the SST variability changes. Lower SST in the first half of the year and higher SST in the latter half of the year may contribute to the difference of the SST variability within a year in flux-adjusted cases (Fig. 6a), while lower SST throughout the year in 6 ka than in 0 ka may contribute to an overall weaker SST variability in the non-flux-adjusted cases (Fig. 6b).

To understand different behavior of the equatorial Pacific SST seasonal cycle between the simulations, Hovmöller diagrams of monthly mean SST and surface wind stress averaged for 5°S–5°N are compared in Fig. 10. The seasonal cycle of SST in 0ka_fa (Fig. 10a) is very close to the observation (not shown), with a large annual cycle in the eastern Pacific (amplitude greater than 0.5°C east of 150°W, with a maximum of about 1.5°C near 100°W) and a semiannual cycle with maxima in May and November in the western Pacific. Easterly winds converge toward a warm water pool over 140°–160°E. In the non-flux-adjusted run (0ka_nfa, Fig. 10c), a westward displacement of the Walker circulation is evident and easterlies penetrate further west of 140°E.

In 6ka_fa, zonal wind anomalies in the western Pacific change their sign between February–May and August–December (Fig. 10b). During February–May, a colder land surface temperature, relative to that over the ocean, makes positive SLP anomalies over land. Associated with these pressure differences are westerly surface wind anomalies in the western Pacific. Horizontal wind convergence is found around the date line in this season. As insolation anomaly changes its sign from negative to positive during the latter half of the year, land surface warms more than the sea surface, and monsoonal low pressure anomalies prevail over land so that easterly trade winds become stronger in the western Pacific during August–December. The reversal of zonal wind anomalies with season is also found in 6ka_nfa, but its zonal extent is confined west of 140°E because of the westward shift of the whole system.

Next we consider the longitude–depth cross sections of the upper-ocean temperature and upwelling/downwelling in the equatorial region in the two contrasting seasons. The temperature and zonal and vertical velocity averaged for 5°S–5°N of the upper 300 m in the Indian and the Pacific Oceans in the 0-ka experiments and the difference between 6- and 0-ka runs are shown in Fig. 11 for April and in Fig. 12 for October.

In the flux-adjusted runs, there exists a contrasting response in the vertical velocity and upper-ocean temperature anomalies over the central equatorial Pacific between April and October. In April, converging surface wind anomalies are responsible for downwelling anomalies (i.e., less intense upwelling) in the central equatorial Pacific, with positive temperature anomalies in the thermocline depth (Fig. 11b). The warm equatorial thermocline is partly caused by the warm subsurface water from the Southern Hemisphere (Liu et al. 2003). Surface wind anomalies change their sign in October because of opposite monsoonal circulation anomalies (Fig. 10), and more intense upwelling is evident in 6ka_fa than in 0ka_fa, with negative temperature anomalies in the thermocline (Fig. 12b). Temperature anomalies close to the surface are negative in April and positive in October in association with insolation changes. Thus, the vertical temperature gradient in the upper equatorial Pacific Ocean is more diffuse in April in 6ka_fa compared to that in 0ka_fa, while it is tighter in October. These seasonal changes in the mean ocean state may explain the seasonal change in the SST variability between 0ka_fa and 6ka_fa (Fig. 6a).

In the non-flux-adjusted run, reversal of monsoonal wind anomalies in 6 ka appears around the Maritime Continent, which is located much more westward than the flux-adjusted counterpart (Fig. 10). In April, vertical velocity anomalies are not so significant in the central equatorial Pacific (Fig. 11d). However, there are significant upwelling anomalies east of the date line in October in 6ka_nfa (Fig. 12d) with stronger easterly winds (Fig. 10d). The significant upwelling anomalies result in a cooler upper ocean and lower SST in 6ka_nfa than in 0ka_nfa throughout the year despite larger insolation during the latter half of the year. The thermocline structure is, therefore, more diffuse and SST variability is less in 6ka_nfa than in 0ka_nfa throughout the year.

In summarizing this section, relative role of atmospheric and oceanic sensitivity is compared. Following Timmermann et al. (1999), an atmospheric sensitivity index α is defined as
i1520-0442-20-11-2484-e1
where T is the Niño-3.4 SST anomaly, τx is the zonal wind stress anomaly averaged for 4°S–4°N, 150°E–150°W, and “cov” and “var” denotes covariance and variance, respectively. Similarly, an oceanic sensitivity index, β, is defined as
i1520-0442-20-11-2484-e2
Larger wind stress anomalies per unit change of SST anomaly (larger atmospheric sensitivity) would gain more latent heat supply from the ocean, leading to a larger moisture content, more convection, and thus a larger atmospheric response. On the other hand, larger oceanic sensitivity indicates the importance of ocean dynamics for El Niño variability, where a tight thermocline prefers to have greater-amplitude El Niño variability (Timmermann et al. 1999; Meehl et al. 2001).

These atmospheric and oceanic sensitivity indices for each experiment are shown in Fig. 13. In the flux-adjusted runs, the atmospheric sensitivity index in 6 ka is larger than that in 0 ka, while the direction of change in oceanic sensitivity index is opposite. Larger atmospheric sensitivity in 6 ka is due to increased trade winds. Lower oceanic sensitivity in 6 ka is in accord with the more diffuse thermocline in 6 than in 0 ka (Figs. 11 and 12). Therefore, the 6-ka atmosphere tends to increase El Niño amplitude; on the other hand, the 6-ka ocean tends to decrease El Niño amplitude.

A qualitatively similar feature is also seen in non-flux-adjusted runs. However, in this case, quantitative changes between 0 and 6 ka are different. The atmospheric sensitivity shows virtually no change, although the oceanic sensitivity largely decreases from 0 to 6 ka. This relative change of the oceanic sensitivity strength from 0 to 6 ka between flux- and non-flux-adjusted runs is in accord with thermocline changes as seen in Figs. 11 and 12. It is also noted that magnitude of atmospheric sensitivity is smaller in non-flux- than in flux-adjusted runs, while oceanic sensitivity is comparable between them. Non-flux-adjusted cases have lower atmospheric sensitivity resulting from weaker basic-state easterlies in the eastern Pacific where the thermocline gradient is large as a result of westward displacement of the warm pool and converging easterlies. Cooler SST (cold bias) in 0ka_nfa may be the reason of this weaker Bjerknes feedback.

6. Conclusions

We have integrated the MRI-CGCM2.3 under the preindustrial (0 ka) and mid-Holocene (6 ka) conditions. Both used the same atmospheric greenhouse gases concentrations as that at the preindustrial level, and the same land vegetation cover and prescribed ice sheets. The only difference between the 0- and 6-ka runs in this study is the earth’s orbital parameters. A version with flux adjustment and a version without flux adjustment are employed to compare the possible effect of differences in the basic state (i.e., the control climate) and effect of adjustment on air–sea coupling. It is found that the El Niño variability in 6 ka only slightly decreases from that in 0 ka with flux-adjusted runs. On the other hand, the El Niño variability in 6 ka becomes significantly weaker from that in 0 ka without flux adjustment. During the first half of the year, both the flux- and non-flux-adjusted cases show a similar mechanism for weak El Niño variability. On the other hand, during the latter half of the year, the flux-adjusted case tends to increase El Niño variability as SST becomes warmer because of solar forcing, but the colder upper ocean in the non-flux-adjusted case suppresses this anomalous heating, resulting in smaller variability throughout the year.

Regarding changes in El Niño variability in 6 ka, atmospheric and oceanic sensitivity tend to work in the opposite way; the atmospheric sensitivity tends to increase variability while the oceanic sensitivity tends to decrease variability. The former may be related to intensifying trade winds in the 6- compared to the 0-ka climate, while cooler SST and more diffuse thermocline structure in 6 ka resulting from surface cooling and warming below the thermocline may explain the latter. Because flux- and non-flux-adjusted runs have different magnitudes in their atmospheric and oceanic (and thus total) sensitivity, modulation of simulated El Niño variability in 6 versus 0 ka differs whether we use flux adjustment or not, at least in our model. We obtained comparable El Niño variability and ENSO characteristics with flux adjustment, while a reduced amplitude of El Niño variability in 6 ka is found in the non-flux-adjusted cases.

The use of flux adjustment has two contrasting effects. Models without flux adjustment often show an extensive and westward-penetrating cold tongue in the equatorial Pacific, together with a distorted SST pattern in model El Niño. Cold bias may also affect the strength of ocean–atmosphere feedback because convection depends on temperatures. Flux adjustment is used to alleviate these systematic errors (bias) in the model basic state. On the other hand, flux adjustment distorts the way the ocean and atmosphere couple each other (Neelin and Dijkstra 1995). The observed proxy data are available only for limited regions and thus do not cover the true El Niño variability in the equatorial Pacific. However, they as a whole suggest a weaker El Niño variability in the mid-Holocene. Although it is impossible to determine whether the differences in the 6-ka ENSO response are a consequence of differences in the basic state or an intrinsic consequence of the use of flux adjustments, the non-flux-adjusted case seems to better agree to our knowledge based on the observation, suggesting that nondistortion of sensitivity to forcing may be important, even if systematic errors (bias) exist in the non-flux-adjusted cases. However, because the present model is still not satisfactory in simulating ENSO behavior in its control simulation under the non-flux-adjusted case, such as a weaker variability and a westward displacement of the variability maximum, more research is needed for deducing ENSO changes during the Holocene.

Acknowledgments

The authors thank O. Arakawa for his contribution on computational assistance. We appreciate valuable comments to the original manuscript by three anonymous reviewers.

REFERENCES

  • Berger, A., 1978: Long-term variation of daily insolation and Quaternary climatic changes. J. Atmos. Sci., 35 , 23622367.

  • Braconnot, P., O. Marti, S. Joussaume, and Y. Leclainche, 2000a: Ocean feedback in response to 6 kyr BP insolation. J. Climate, 13 , 15371553.

    • Search Google Scholar
    • Export Citation
  • Braconnot, P., S. Joussaume, N. de Noblet, and G. Ramstein, 2000b: Mid-Holocene and Last Glacial Maximum African monsoon changes as simulated within the Paleoclimate Modelling Intercomparison Project. Global Planet. Change, 26 , 5166.

    • Search Google Scholar
    • Export Citation
  • Braconnot, P., and Coauthors, 2003: The second phase of the Paleoclimate Modeling Intercomparison Project (PMIP-II). CLIVAR Exchanges, No. 28, International CLIVAR Project Office, Southampton, United Kingdom, 19–20.

  • Braconnot, P., and Coauthors, 2004: Evaluation of PMIP coupled ocean-atmosphere simulations of the Mid-Holocene. Past Climate Variability through Europe and Africa, R.W. Battarbee, F. Gasse, and C. E. Stickley, Eds., Springer, 515–533.

    • Search Google Scholar
    • Export Citation
  • Bush, A. B. G., 1999: Assessing the impact of Mid-Holocene insolation on the atmosphere-ocean system. Geophys. Res. Lett., 26 , 99102.

    • Search Google Scholar
    • Export Citation
  • Clement, A. C., R. Seager, and M. A. Cane, 2000: Suppression of El Niño during the mid-Holocene by changes in the Earth’s orbit. Paleoceanography, 15 , 731737.

    • Search Google Scholar
    • Export Citation
  • Cole, J., 2001: A slow dance for El Niño. Science, 291 , 14961497.

  • Crucifix, M., P. Braconnot, S. P. Harrison, and B. Otto-Bliesner, 2005: Second Phase of Paleoclimate Modelling Intercomparison Project. Eos, Trans. Amer. Geophys. Union, 86 , 264265.

    • Search Google Scholar
    • Export Citation
  • Davey, M. K., and Coauthors, 2002: STOIC: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn., 18 , 403420.

    • Search Google Scholar
    • Export Citation
  • Gagan, M. K., L. K. Ayliffe, D. Hopley, J. A. Cali, G. E. Mortimer, J. Chappell, M. T. McCulloch, and M. J. Head, 1998: Temperature and surface-ocean water balance of the mid-Holocene tropical western Pacific. Science, 279 , 10141018.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., 2006: El Niño-mean state-seasonal cycle interactions in a multi-model ensemble. Climate Dyn., 26 , 329348.

  • Hewitt, C. D., and J. F. B. Mitchell, 1998: A fully coupled GCM simulation of the climate of the mid-Holocene. Geophys. Res. Lett., 25 , 361364.

    • Search Google Scholar
    • Export Citation
  • Hu, Z-Z., E. K. Schneider, U. S. Bhatt, and B. P. Kirtman, 2004: Potential mechanism for response of El Niño-Southern Oscillation variability to change in land surface energy budget. J. Geophys. Res., 109 .D21113, doi:10.1029/2004JD004771.

    • Search Google Scholar
    • Export Citation
  • Joussaume, S., and K. E. Taylor, 1995: Status of the Paleoclimate Modeling Intercomparison Project (PMIP). Proc. First Int. AMIP Scientific Conf., WCRP-92, WMO/TD-732, Monterey, CA, WMO, 425–430.

  • Joussaume, S., and Coauthors, 1999: Monsoon changes for 6000 years ago: Results of 18 simulations from the Paleoclimate Modeling Intercomparison Project (PMIP). Geophys. Res. Lett., 26 , 859862.

    • Search Google Scholar
    • Export Citation
  • Kitoh, A., and S. Murakami, 2002: Tropical Pacific climate at the Mid-Holocene and the Last Glacial Maximum simulated by a coupled ocean-atmosphere GCM. Paleoceanography, 17 .1047, doi:10.1029/2001PA000724.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate, 9 , 20362057.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 2003: Atmosphere–ocean variations in the Indo-Pacific sector during ENSO episodes. J. Climate, 16 , 320.

  • Liu, Z., J. Kutzbach, and L. Wu, 2000: Modeling climate shift of El Niño variability in the Holocene. Geophys. Res. Lett., 27 , 22652268.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., E. Brady, and J. Lynch-Steiglitz, 2003: Global ocean response to orbital forcing in the Holocene. Paleoceanography, 18 .1041, doi:10.1029/2002PA000819.

    • Search Google Scholar
    • Export Citation
  • McAvaney, B. J., and Coauthors, 2001: Model evaluation. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 471–523.

    • Search Google Scholar
    • Export Citation
  • McGregor, H. V., and M. K. Gagan, 2004: Western Pacific coral δ18O records of anomalous Holocene variability in the El Niño-Southern Oscillation. Geophys. Res. Lett., 31 .L11204, doi:10.1029/2004GL019972.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., P. R. Gent, J. M. Arblaster, B. L. Otto-Bliesner, E. C. Brady, and A. Craig, 2001: Factors that affect the amplitude of El Niño in global coupled climate models. Climate Dyn., 17 , 515526.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., H. Teng, and G. Branstator, 2006: Future changes of El Niño in two global coupled climate models. Climate Dyn., 26 , 549566.

    • Search Google Scholar
    • Export Citation
  • Merryfield, W. J., 2006: Changes to ENSO under CO2 doubling in a multimodel ensemble. J. Climate, 19 , 40094027.

  • Neelin, J. D., and H. A. Dijkstra, 1995: Ocean-atmosphere interaction and the tropical climatology. Part I: The dangers of flux correction. J. Climate, 8 , 13251342.

    • Search Google Scholar
    • Export Citation
  • Otto-Bliesner, B. L., 1999: El Niño/La Niña and Sahel precipitation during the middle Holocene. Geophys. Res. Lett., 26 , 8790.

  • Otto-Bliesner, B. L., E. C. Brady, S-I. Shin, Z. Liu, and C. Shields, 2003: Modeling El Niño and its tropical teleconnections during the last glacial-interglacial cycle. Geophys. Res. Lett., 30 .2198, doi:10.1029/2003GL018553.

    • Search Google Scholar
    • Export Citation
  • Otto-Bliesner, B. L., E. C. Brady, G. Clauzet, R. Tomas, S. Levis, and Z. Kothavala, 2006: Last Glacial Maximum and Holocene climate in CCSM3. J. Climate, 19 , 25262544.

    • Search Google Scholar
    • Export Citation
  • PMIP, 2000: Paleoclimate Modeling Intercomparison Project (PMIP). Proc. Third PMIP Workshop, WCRP-111, WMO/TD-1007, La Huardiere, Canada, WMO, 271 pp.

  • 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 .4407, doi:10.1209/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Rein, B., A. Lückge, L. Reinhardt, F. Sirocko, A. Wolf, and W-C. Dullo, 2005: El Niño variability off Peru during the last 20,000 years. Paleoceanography, 20 .PA4003, doi:10.1029/2004PA001099.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., and M. S. Halpert, 1989: Precipitation patterns associated with the high index phase of the Southern Oscillation. J. Climate, 2 , 268284.

    • Search Google Scholar
    • Export Citation
  • Timmermann, A., J. Oberhuber, A. Bacher, M. Esch, M. Latif, and E. Roeckner, 1999: Increased El Niño frequency in a climate model forced by future greenhouse warming. Nature, 398 , 694697.

    • Search Google Scholar
    • Export Citation
  • Tudhope, A. W., and Coauthors, 2001: Variability in the El Niño-Southern Oscillation through a glacial-interglacial cycle. Science, 291 , 15111517.

    • Search Google Scholar
    • Export Citation
  • van Oldenborgh, G. J., S. Philip, and M. Collins, 2005: El Niño in a changing climate: A multi-model study. Ocean Sci., 1 , 8195.

  • Webster, P. J., V. O. Magaña, T. N. Palmer, J. Shukla, R. A. Tomas, M. Yanai, and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103 , 1445114510.

    • Search Google Scholar
    • Export Citation
  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate, 9 , 840858.

    • Search Google Scholar
    • Export Citation
  • Yasunari, T., 1990: Impact of Indian monsoon on the coupled atmosphere/ocean system in the tropical Pacific. Meteor. Atmos. Phys., 44 , 2941.

    • Search Google Scholar
    • Export Citation
  • Yeh, S-W., and B. P. Kirtman, 2005: Pacific decadal variability and decadal ENSO amplitude modulation. Geophys. Res. Lett., 32 .L05703, doi:10.1029/2004GL021731.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., and Coauthors, 2006a: Present-day climate and climate sensitivity in the Meteorological Research Institute coupled GCM version 2.3 (MRI-CGCM2.3). J. Meteor. Soc. Japan, 84 , 333363.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., A. Noda, T. Uchiyama, S. Kusunoki, and A. Kitoh, 2006b: Climate changes of the twentieth through twenty-first centuries simulated by the MRI-CGCM2.3. Pap. Meteor. Geophys., 56 , 924.

    • Search Google Scholar
    • Export Citation
  • Zelle, H., G. J. van Oldenborgh, G. Burgers, and H. A. Dijkstra, 2005: El Niño and greenhouse warming: Results from ensemble simulations with the NCAR CCSM. J. Climate, 18 , 46694683.

    • Search Google Scholar
    • Export Citation
  • Zhao, Y., and Coauthors, 2005: A multi-model analysis of the role of the ocean on the African and Indian monsoon during the mid-Holocene. Climate Dyn., 25 , 777800.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Time series of the annual mean global surface air temperature for 0- (bold solid lines) and 6-ka (thin solid lines) runs.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 2.
Fig. 2.

(a) Observed annual mean SST, (b) annual mean SST in 0ka_fa, (c) annual mean SST in 0ka_nfa, (d) observed annual mean precipitation, (e) annual mean precipitation in 0ka_fa, and (f) annual mean precipitation in 0ka_nfa.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 3.
Fig. 3.

(a) Latitude–month cross section of zonal mean surface air temperature difference between 6ka_fa and 0ka_fa. (b) Same as (a), but for the difference between 6ka_nfa and 0ka_nfa.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 4.
Fig. 4.

Annual mean differences in (a), (b) surface air temperature (°C); (c), (d) precipitation (mm day−1); and (e), (f) sea surface salinity (psu). Regions with statistically significant difference at 95% are shaded.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 5.
Fig. 5.

(a), (b) The first EOF of the near-global (50°S–50°N) SST of SST in 0ka_fa, and (c), (d) associated regression fields of the precipitation and the wind vector at 850 hPa. Regions with correlation coefficients larger than +0.14 or less than −0.14 (statistically significant at 95% level) are shaded. (e)–(h) Same as in (a)–(d), except for 6ka_fa.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 6.
Fig. 6.

(a) Seasonal change of the interannual standard deviation of the Niño-3.4 SST anomalies. Thick solid line denotes 0ka_fa. Thick dashed line denotes 6ka_fa. Thin solid line denotes the observed values for the period 1950–1993. (b) Same as in (a), but for 0ka_nfa and 6ka_nfa.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 7.
Fig. 7.

(a) Lag correlations between IMR and Niño-3.4 SST anomalies from the previous year (−1) to the following year (+1). Thick solid and thick dashed lines denote 0ka_fa and 6ka_fa results, respectively. Thin solid line denotes observed relationship for 1950–93. Vertical dashed lines denote the IMR season. (b) Same as in (a), but for 0ka_nfa and 6ka_nfa.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 8.
Fig. 8.

Same as in Fig. 5, but for the non-flux-adjusted run (0ka_nfa and 6ka_nfa).

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 9.
Fig. 9.

(a) Seasonal change of the difference in downward shortwave radiation at the top of the atmosphere between 6 and 0 ka. Average between 5°S and 5°N is shown in thick solid line. Thin solid line denotes annual mean difference (−1.09 W m−2). (b) Seasonal change of the Niño-3.4 SST. Thin solid line with a rectangle mark denotes the observed values for the period 1901–2000 based on the HadISST1.1.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 10.
Fig. 10.

(a) Longitude–time cross section of SST and surface wind stress averaged for 5°S–5°N in 0ka_fa run. Zonal and meridional components of surface wind stress are shown as vectors. (b) Same as (a), but for the difference between 6ka_fa and 0ka_fa. (c), (d) Same as (a), (b), but for the non-flux-adjusted run.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 11.
Fig. 11.

Ocean temperature and zonal and vertical velocity in April for upper 300 m averaged for 5°S–5°N for (a) 0ka_fa, and (b) the difference of 6ka_fa minus 0ka_fa. In (b), velocity differences with statistically significant difference at 90% are plotted. (c), (d) Same as in (a), (b), but for the non-flux-adjusted run (0ka_nfa and 6ka_nfa).

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 12.
Fig. 12.

Same as in Fig. 11, except for October.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Fig. 13.
Fig. 13.

Atmospheric and oceanic sensitivity.

Citation: Journal of Climate 20, 11; 10.1175/JCLI4141.1

Table 1.

Experiments done by the MRI-CGCM2.3.2.

Table 1.
Save
  • Berger, A., 1978: Long-term variation of daily insolation and Quaternary climatic changes. J. Atmos. Sci., 35 , 23622367.

  • Braconnot, P., O. Marti, S. Joussaume, and Y. Leclainche, 2000a: Ocean feedback in response to 6 kyr BP insolation. J. Climate, 13 , 15371553.

    • Search Google Scholar
    • Export Citation
  • Braconnot, P., S. Joussaume, N. de Noblet, and G. Ramstein, 2000b: Mid-Holocene and Last Glacial Maximum African monsoon changes as simulated within the Paleoclimate Modelling Intercomparison Project. Global Planet. Change, 26 , 5166.

    • Search Google Scholar
    • Export Citation
  • Braconnot, P., and Coauthors, 2003: The second phase of the Paleoclimate Modeling Intercomparison Project (PMIP-II). CLIVAR Exchanges, No. 28, International CLIVAR Project Office, Southampton, United Kingdom, 19–20.

  • Braconnot, P., and Coauthors, 2004: Evaluation of PMIP coupled ocean-atmosphere simulations of the Mid-Holocene. Past Climate Variability through Europe and Africa, R.W. Battarbee, F. Gasse, and C. E. Stickley, Eds., Springer, 515–533.

    • Search Google Scholar
    • Export Citation
  • Bush, A. B. G., 1999: Assessing the impact of Mid-Holocene insolation on the atmosphere-ocean system. Geophys. Res. Lett., 26 , 99102.

    • Search Google Scholar
    • Export Citation
  • Clement, A. C., R. Seager, and M. A. Cane, 2000: Suppression of El Niño during the mid-Holocene by changes in the Earth’s orbit. Paleoceanography, 15 , 731737.

    • Search Google Scholar
    • Export Citation
  • Cole, J., 2001: A slow dance for El Niño. Science, 291 , 14961497.

  • Crucifix, M., P. Braconnot, S. P. Harrison, and B. Otto-Bliesner, 2005: Second Phase of Paleoclimate Modelling Intercomparison Project. Eos, Trans. Amer. Geophys. Union, 86 , 264265.

    • Search Google Scholar
    • Export Citation
  • Davey, M. K., and Coauthors, 2002: STOIC: A study of coupled model climatology and variability in tropical ocean regions. Climate Dyn., 18 , 403420.

    • Search Google Scholar
    • Export Citation
  • Gagan, M. K., L. K. Ayliffe, D. Hopley, J. A. Cali, G. E. Mortimer, J. Chappell, M. T. McCulloch, and M. J. Head, 1998: Temperature and surface-ocean water balance of the mid-Holocene tropical western Pacific. Science, 279 , 10141018.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., 2006: El Niño-mean state-seasonal cycle interactions in a multi-model ensemble. Climate Dyn., 26 , 329348.

  • Hewitt, C. D., and J. F. B. Mitchell, 1998: A fully coupled GCM simulation of the climate of the mid-Holocene. Geophys. Res. Lett., 25 , 361364.

    • Search Google Scholar
    • Export Citation
  • Hu, Z-Z., E. K. Schneider, U. S. Bhatt, and B. P. Kirtman, 2004: Potential mechanism for response of El Niño-Southern Oscillation variability to change in land surface energy budget. J. Geophys. Res., 109 .D21113, doi:10.1029/2004JD004771.

    • Search Google Scholar
    • Export Citation
  • Joussaume, S., and K. E. Taylor, 1995: Status of the Paleoclimate Modeling Intercomparison Project (PMIP). Proc. First Int. AMIP Scientific Conf., WCRP-92, WMO/TD-732, Monterey, CA, WMO, 425–430.

  • Joussaume, S., and Coauthors, 1999: Monsoon changes for 6000 years ago: Results of 18 simulations from the Paleoclimate Modeling Intercomparison Project (PMIP). Geophys. Res. Lett., 26 , 859862.

    • Search Google Scholar
    • Export Citation
  • Kitoh, A., and S. Murakami, 2002: Tropical Pacific climate at the Mid-Holocene and the Last Glacial Maximum simulated by a coupled ocean-atmosphere GCM. Paleoceanography, 17 .1047, doi:10.1029/2001PA000724.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 1996: The role of the “atmospheric bridge” in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate, 9 , 20362057.

    • Search Google Scholar
    • Export Citation
  • Lau, N-C., and M. J. Nath, 2003: Atmosphere–ocean variations in the Indo-Pacific sector during ENSO episodes. J. Climate, 16 , 320.

  • Liu, Z., J. Kutzbach, and L. Wu, 2000: Modeling climate shift of El Niño variability in the Holocene. Geophys. Res. Lett., 27 , 22652268.

    • Search Google Scholar
    • Export Citation
  • Liu, Z., E. Brady, and J. Lynch-Steiglitz, 2003: Global ocean response to orbital forcing in the Holocene. Paleoceanography, 18 .1041, doi:10.1029/2002PA000819.

    • Search Google Scholar
    • Export Citation
  • McAvaney, B. J., and Coauthors, 2001: Model evaluation. Climate Change 2001: The Scientific Basis, J. T. Houghton et al., Eds., Cambridge University Press, 471–523.

    • Search Google Scholar
    • Export Citation
  • McGregor, H. V., and M. K. Gagan, 2004: Western Pacific coral δ18O records of anomalous Holocene variability in the El Niño-Southern Oscillation. Geophys. Res. Lett., 31 .L11204, doi:10.1029/2004GL019972.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., P. R. Gent, J. M. Arblaster, B. L. Otto-Bliesner, E. C. Brady, and A. Craig, 2001: Factors that affect the amplitude of El Niño in global coupled climate models. Climate Dyn., 17 , 515526.

    • Search Google Scholar
    • Export Citation
  • Meehl, G. A., H. Teng, and G. Branstator, 2006: Future changes of El Niño in two global coupled climate models. Climate Dyn., 26 , 549566.

    • Search Google Scholar
    • Export Citation
  • Merryfield, W. J., 2006: Changes to ENSO under CO2 doubling in a multimodel ensemble. J. Climate, 19 , 40094027.

  • Neelin, J. D., and H. A. Dijkstra, 1995: Ocean-atmosphere interaction and the tropical climatology. Part I: The dangers of flux correction. J. Climate, 8 , 13251342.

    • Search Google Scholar
    • Export Citation
  • Otto-Bliesner, B. L., 1999: El Niño/La Niña and Sahel precipitation during the middle Holocene. Geophys. Res. Lett., 26 , 8790.

  • Otto-Bliesner, B. L., E. C. Brady, S-I. Shin, Z. Liu, and C. Shields, 2003: Modeling El Niño and its tropical teleconnections during the last glacial-interglacial cycle. Geophys. Res. Lett., 30 .2198, doi:10.1029/2003GL018553.

    • Search Google Scholar
    • Export Citation
  • Otto-Bliesner, B. L., E. C. Brady, G. Clauzet, R. Tomas, S. Levis, and Z. Kothavala, 2006: Last Glacial Maximum and Holocene climate in CCSM3. J. Climate, 19 , 25262544.

    • Search Google Scholar
    • Export Citation
  • PMIP, 2000: Paleoclimate Modeling Intercomparison Project (PMIP). Proc. Third PMIP Workshop, WCRP-111, WMO/TD-1007, La Huardiere, Canada, WMO, 271 pp.

  • 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 .4407, doi:10.1209/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Rein, B., A. Lückge, L. Reinhardt, F. Sirocko, A. Wolf, and W-C. Dullo, 2005: El Niño variability off Peru during the last 20,000 years. Paleoceanography, 20 .PA4003, doi:10.1029/2004PA001099.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., and M. S. Halpert, 1989: Precipitation patterns associated with the high index phase of the Southern Oscillation. J. Climate, 2 , 268284.

    • Search Google Scholar
    • Export Citation
  • Timmermann, A., J. Oberhuber, A. Bacher, M. Esch, M. Latif, and E. Roeckner, 1999: Increased El Niño frequency in a climate model forced by future greenhouse warming. Nature, 398 , 694697.

    • Search Google Scholar
    • Export Citation
  • Tudhope, A. W., and Coauthors, 2001: Variability in the El Niño-Southern Oscillation through a glacial-interglacial cycle. Science, 291 , 15111517.

    • Search Google Scholar
    • Export Citation
  • van Oldenborgh, G. J., S. Philip, and M. Collins, 2005: El Niño in a changing climate: A multi-model study. Ocean Sci., 1 , 8195.

  • Webster, P. J., V. O. Magaña, T. N. Palmer, J. Shukla, R. A. Tomas, M. Yanai, and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. J. Geophys. Res., 103 , 1445114510.

    • Search Google Scholar
    • Export Citation
  • Xie, P., and P. Arkin, 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Climate, 9 , 840858.

    • Search Google Scholar
    • Export Citation
  • Yasunari, T., 1990: Impact of Indian monsoon on the coupled atmosphere/ocean system in the tropical Pacific. Meteor. Atmos. Phys., 44 , 2941.

    • Search Google Scholar
    • Export Citation
  • Yeh, S-W., and B. P. Kirtman, 2005: Pacific decadal variability and decadal ENSO amplitude modulation. Geophys. Res. Lett., 32 .L05703, doi:10.1029/2004GL021731.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., and Coauthors, 2006a: Present-day climate and climate sensitivity in the Meteorological Research Institute coupled GCM version 2.3 (MRI-CGCM2.3). J. Meteor. Soc. Japan, 84 , 333363.

    • Search Google Scholar
    • Export Citation
  • Yukimoto, S., A. Noda, T. Uchiyama, S. Kusunoki, and A. Kitoh, 2006b: Climate changes of the twentieth through twenty-first centuries simulated by the MRI-CGCM2.3. Pap. Meteor. Geophys., 56 , 924.

    • Search Google Scholar
    • Export Citation
  • Zelle, H., G. J. van Oldenborgh, G. Burgers, and H. A. Dijkstra, 2005: El Niño and greenhouse warming: Results from ensemble simulations with the NCAR CCSM. J. Climate, 18 , 46694683.

    • Search Google Scholar
    • Export Citation
  • Zhao, Y., and Coauthors, 2005: A multi-model analysis of the role of the ocean on the African and Indian monsoon during the mid-Holocene. Climate Dyn., 25 , 777800.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Time series of the annual mean global surface air temperature for 0- (bold solid lines) and 6-ka (thin solid lines) runs.

  • Fig. 2.

    (a) Observed annual mean SST, (b) annual mean SST in 0ka_fa, (c) annual mean SST in 0ka_nfa, (d) observed annual mean precipitation, (e) annual mean precipitation in 0ka_fa, and (f) annual mean precipitation in 0ka_nfa.

  • Fig. 3.

    (a) Latitude–month cross section of zonal mean surface air temperature difference between 6ka_fa and 0ka_fa. (b) Same as (a), but for the difference between 6ka_nfa and 0ka_nfa.

  • Fig. 4.

    Annual mean differences in (a), (b) surface air temperature (°C); (c), (d) precipitation (mm day−1); and (e), (f) sea surface salinity (psu). Regions with statistically significant difference at 95% are shaded.

  • Fig. 5.

    (a), (b) The first EOF of the near-global (50°S–50°N) SST of SST in 0ka_fa, and (c), (d) associated regression fields of the precipitation and the wind vector at 850 hPa. Regions with correlation coefficients larger than +0.14 or less than −0.14 (statistically significant at 95% level) are shaded. (e)–(h) Same as in (a)–(d), except for 6ka_fa.

  • Fig. 6.

    (a) Seasonal change of the interannual standard deviation of the Niño-3.4 SST anomalies. Thick solid line denotes 0ka_fa. Thick dashed line denotes 6ka_fa. Thin solid line denotes the observed values for the period 1950–1993. (b) Same as in (a), but for 0ka_nfa and 6ka_nfa.

  • Fig. 7.

    (a) Lag correlations between IMR and Niño-3.4 SST anomalies from the previous year (−1) to the following year (+1). Thick solid and thick dashed lines denote 0ka_fa and 6ka_fa results, respectively. Thin solid line denotes observed relationship for 1950–93. Vertical dashed lines denote the IMR season. (b) Same as in (a), but for 0ka_nfa and 6ka_nfa.

  • Fig. 8.

    Same as in Fig. 5, but for the non-flux-adjusted run (0ka_nfa and 6ka_nfa).

  • Fig. 9.

    (a) Seasonal change of the difference in downward shortwave radiation at the top of the atmosphere between 6 and 0 ka. Average between 5°S and 5°N is shown in thick solid line. Thin solid line denotes annual mean difference (−1.09 W m−2). (b) Seasonal change of the Niño-3.4 SST. Thin solid line with a rectangle mark denotes the observed values for the period 1901–2000 based on the HadISST1.1.

  • Fig. 10.

    (a) Longitude–time cross section of SST and surface wind stress averaged for 5°S–5°N in 0ka_fa run. Zonal and meridional components of surface wind stress are shown as vectors. (b) Same as (a), but for the difference between 6ka_fa and 0ka_fa. (c), (d) Same as (a), (b), but for the non-flux-adjusted run.

  • Fig. 11.

    Ocean temperature and zonal and vertical velocity in April for upper 300 m averaged for 5°S–5°N for (a) 0ka_fa, and (b) the difference of 6ka_fa minus 0ka_fa. In (b), velocity differences with statistically significant difference at 90% are plotted. (c), (d) Same as in (a), (b), but for the non-flux-adjusted run (0ka_nfa and 6ka_nfa).

  • Fig. 12.

    Same as in Fig. 11, except for October.

  • Fig. 13.

    Atmospheric and oceanic sensitivity.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 131 39 6
PDF Downloads 52 20 10