In this study, the outputs from four Ocean-Atmosphere Coupled General Circulation Model experiments within the Paleoclimate Modelling Intercomparison Project phase 2 and four sets of Atmospheric General Circulation Model experiments were used to analyze the Asian summer climate during the mid-Holocene (6 ka, about 6000 years ago). Additionally, the role of the orbital forcings and the effects of a cold ocean surface for the 6 ka were investigated by comparing the Atmospheric General Circulation Model simulations forced by different combinations of forcing parameters. The results indicated that in the 6 ka summer, the orbital forcings were the prime drivers of the increased temperature and precipitation and the strengthened summer monsoon over the Asian continent. On the other hand, these different orbital forcings also resulted in a colder Indian Ocean-northwestern Pacific during this period. Our results suggested that this cold ocean surface could reduce the warming amplitude and precipitation enhancement over the Asian monsoon area in the 6 ka summer. The changes caused by the different ocean surface conditions were comparable with simulated 6 ka climate changes. The cold ocean surface also suppressed the Asian summer monsoon circulations. Therefore, the influences from anomalous ocean surface conditions played an important role on regulating the Asian summer climate during the 6 ka. In addition, it was found that the summer climate in the South Asian monsoon area was more sensitive to the changes in the orbital forcings and ocean surface conditions than that in the East Asian monsoon area.
 The mid-Holocene (about 6000 years ago, hereafter abbreviated as 6 ka) is a well-known warm climate interval [Jansen et al., 2007] and is of interest to the Paleoclimate Modelling Intercomparison Project (PMIP) [Joussaume et al., 1999]. During the 6 ka, seasonal difference in the Northern Hemisphere was stronger compared to the present, largely as a result of variations of solar insolation associated with the changes in Earth's orbital parameters [Berger, 1978]. As a consequence of increased solar insolation during the 6 ka, the Northern Hemisphere summers were warmer, and the ice sheets had almost disappeared.
 Numerous studies have addressed the orbital forcing induced climate changes in East Asia [e.g., Zhou and Zhao, 2009; Wang et al., 2010; Jiang et al., 2012; Zhou and Zhao, 2012], Europe [e.g., Bonfils et al., 2004; Brewer et al., 2007], Africa, and India [e.g., Zhao et al., 2005; Patricola and Cook, 2007; Zhao et al., 2007] during the 6 ka. The Pacific and El Niño-Southern Oscillation (ENSO) activities [e.g., Brown et al., 2008; Zheng et al., 2008; Chiang et al., 2009] and large-scale climatic features [Shin et al., 2006; Braconnot et al., 2007a, 2007b] are also quite different from today. It is believed that the 6 ka orbital forcing could have substantially strengthened the summer monsoons and associated precipitation, whereas reduced the Pacific meridional mode activity and the ENSO activity.
 In addition, most research has focused on understanding the related feedback, which is intrinsic in the Earth system and aroused by the 6 ka orbital external forcing. The vegetation feedback could further enhance the orbital forcing induced Northern Hemisphere temperature and rainfall during the 6 ka [Wang, 1999; Gallimore et al., 2005; Otto et al., 2009]. More important, ocean feedback had a substantial impact in altering the response of climate to the 6 ka orbital forcing [Liu et al., 2003; Wei and Wang, 2004; Otto et al., 2009]. Particularly in most coastal regions, the climate response is very different, and even of opposite sign, between the coupled and atmosphere-alone experiments [Braconnot et al., 2000]. Therefore, the climatic impact of the ocean is essential in the 6 ka simulation. Because of the development and perfection of the models, particularly under the framework of the second phase of the PMIP (PMIP2), the Ocean-Atmosphere Coupled General Circulation Model (OAGCM) simulations are to some extent consistent with the proxy reconstructed data [Peyron et al., 2006; Braconnot et al., 2007b; Wang et al., 2010]. The OAGCMs reproduce more realistic 6 ka climate changes than the early Atmospheric General Circulation Models (AGCMs).
 In general, the intensity of monsoons is controlled by the land-ocean thermal contrast [Webster, 1987; Zuo et al., 2012]. Any changes in the Indian Ocean and northwestern Pacific can directly influence the South Asian summer monsoon (SASM) and East Asian summer monsoon (EASM) and associated climate changes in the regions dominated by the Asian monsoonal system [Zhang et al., 2006]. Previous studies have indicated that both the SASM and EASM were significantly strengthened by the orbital forcings alone during the 6 ka [e.g., Braconnot et al., 2002; Wang, 2002; Wang et al., 2010; Zhao and Harrison, 2012]. On the other hand, the Indian Ocean and northwestern Pacific were cooled due to the lower solar insolation over the tropics and subtropics during the 6 ka [Zhao and Harrison, 2012]. Thus, the cold ocean surface could have regulated the strengthened SASM and EASM, making it impossible to neglect the ocean feedback during this period.
 Recently, Zhao and Harrison  compared multimodel results from the PMIP2 OAGCMs with PMIP1 AGCMs and noted that the modification of the 6 ka ocean on the orbital forcings was very complex. The atmospheric response to the 6 ka orbital forcings could be amplified in the EASM area but weakened in the SASM area. The role of the cold ocean surface in altering the 6 ka climate was therefore different for these two monsoon areas. However, this conclusion was mainly based on two generations of climate models. In fact, the influences from the different AGCMs also played a substantial role in the Asian climate [Ohgaito and Abe-Ouchi, 2009]. The simulated difference between AGCMs and OAGCMs could have resulted from differences in the atmospheric component used within these two phases of PMIP. Thus, the issue on how the different ocean surface conditions affect the 6 ka Asian climate needs further investigation. Here we used four groups of AGCM sensitivity experiments, together with four OAGCMs simulations within the PMIP2, to address the related issues. In particular, we focused on the sensitivity of the SASM and EASM to changes in the sea surface temperatures (SSTs) during the 6 ka. Particular attention has been given to the changes in temperature and precipitation induced by the colder ocean surface in the Indian Ocean-northwestern Pacific during this interval.
 The rest of this paper is organized as follows: section 2 describes the model, the experiment design, and brief evaluation of the models’ climatology. Section 3 compares the spatial patterns of SSTs and surface air temperature (SAT) between the 6 ka and preindustrial states (hereafter abbreviated as 0 ka) over the Asia, and analyzes the impacts of the colder SSTs in the Indian Ocean-northwestern Pacific on the inland temperature during the 6 ka. Other climatic component changes and their responses to the colder ocean surface are explored in section 4. Finally, some concluding remarks are presented in section 5.
2 Model Description and Experiment Design
 The climate model we used was the IAP-AGCM Version 4 (IAP4) [Sun et al., 2012]. The IAP4 is a global atmospheric general circulation model, which has been developed from IAP9L-AGCM [Liang, 1996; Zeng and Mu, 2002] at the Institute of Atmospheric Physics, Chinese Academy of Sciences. The old version of this model has been utilized to investigate global and Asian paleoclimate in the 6 ka [e.g., Jin et al., 2006; Wang, 2002; Wei and Wang, 2004] and the Last Glacial Maximum [e.g., Jiang, 2008; Jiang et al., 2003; Ju et al., 2007], and other geological paleo-periods [e.g., Jiang et al., 2005; Zhang et al., 2007]. Its newest version (i.e., IAP4) has a horizontal grid resolution of 1.4° in latitude by 1.4° in longitude and a total of 26 vertical levels are employed, ranging from the surface to 10 hPa. Its higher resolution will be conducive to simulate Asian summer large-scale climate as pointed out by previous studies [Gao et al., 2006; Kusunoki et al., 2006].
 Four sets of simulations were carried out using the IAP4 with prescribed SSTs and sea-ice, which were issued from four PMIP2 OAGCMs, respectively (listed in Table 1a). The outputs from the CCSM, UBRIS, MIROC, and MRIfa were used as ocean surface forcings in this study because of their better representation of the Asian modern climate [Kripalani et al., 2007; Ohgaito and Abe-Ouchi, 2009; Wang et al., 2010; Jiang et al., 2012]. Each set includes three individual simulations, which were integrated with 0 and 6 ka forcings (Table 1b). The differences between them are the orbital forcings and greenhouse gas levels (listed in Table 2) and different ocean surfaces for the 0 and 6 ka periods. Other boundary conditions remained the same. Two simulations were performed following the PMIP set-ups for the periods, 0 and 6 ka (see http://pmip2.lsce.ipsl.fr/). In these simulations, hereafter abbreviated as IAP4[0ka] and IAP4[6ka], the IAP4 was forced by climatological SSTs and sea-ice issued from the OAGCMs for these two periods. In the third simulation, as an IAP4 sensitivity simulation (hereafter abbreviated as IAP4[Sen]), the IAP4 was performed with 0 ka climatological SSTs and sea-ice output from the OAGCMs, but with 6 ka orbital forcings and greenhouse gas levels. All the simulations were consecutively integrated for 40 years. The results analyzed below represent averages for the final 30 years, allowing 10 years for the model to reach a relative equilibrium state. At the same time, the outputs over the last 50 years from these four PMIP2 OAGCMs were also analyzed in this study. The comparison between OAGCMs[6ka] (IAP4[6ka]) and OAGCMs[0ka] (IAP4[0ka]) shows the OAGCMs (IAP4) simulated 6 ka climate changes relative to the 0 ka. The comparison between IAP4[Sen] and IAP4[0ka] reveals changes in climate forced only by the 6 ka orbital parameters, excluding all other feedback, and is therefore equivalent to the PMIP1 AGCM experiments. The comparison between IAP4[6ka] and IAP4[Sen] highlights the role of different ocean surface conditions on the 6 ka Asian climate.
Table 1. The PMIP2 OAGCMs and the Summary of Simulations
Table 2. The Differences in (a) Orbital Parameters and (b) Greenhouse Gas Levels for the 0 and 6 ka Forcings
Angular Precession (°)
Modern- 10 DU
Modern- 10 DU
 Because the Asian summer monsoon was of interest to us, the summer (June, July, and August) climate was investigated in the present study. In addition, the strongest Asian climate changes in the model experiments also occurred in this season during the 6 ka [Wang, 2000, 2002].
 First, it is important to examine whether IAP4 can reliably reproduce Asian summer climatology because this ability is directly related to the utility of this model for investigating the 6 ka climate changes in this region. Figure 1 illustrates that the IAP4[0 ka] ensemble results reproduce the main features of the Asian summer climatology. The spatial correlation coefficients are high, up to 0.995 and 0.716 with respect to the observational data for simulated SAT and precipitation, respectively. The model biases appear to be smaller than most of the coupled models used in the IPCC Fourth Assessment and PMIP2 [Annamalai et al., 2007; Wang et al., 2010]. Additionally, a number of simulations have been performed with this model, and a reasonable representation of many aspects of the Asian modern climate has been obtained [Sun et al., 2012]. Therefore, the IAP4 would be one of the most appropriate models for the study on the 6 ka Asian climate changes and its responses to the 6 ka ocean surface condition.
3 Surface Temperature
3.1 Sea Surface Temperatures
 In the 6 ka experiment, changes in the Earth's orbital parameters intensified the seasonal contrast of the incoming solar radiation at the top of the atmosphere in the Northern Hemisphere. The reverse occurred in the Southern Hemisphere. In addition, the atmospheric CH4 concentration varied from 760 ppb at 0 ka to 650 ppb at 6 ka. These different forcings resulted in different SST patterns during the 6 ka compared to 0 ka. As shown in Figure 2, increased annual SSTs were evident at high latitudes during the 6 ka in the CCSM and UBRIS. The MIROC and MRIfa only simulated very weak warming in the Sea of Okhotsk during the 6 ka. However, large-scale cool anomalies were observed over middle to low latitudes in the Northern Hemisphere and the tropical oceans in all the models. Particularly for the CCSM, the cool anomalies were strongest among the PMIP2 OAGCMs and the maximum cooling centers were located over the Kuroshio-Oyashio extensions, implying a weaker Kuroshio Current during the 6 ka. The simulated summer ocean surface condition was to some extent similar to the annual one. The warming SST anomalies were a little stronger in the northern middle- to high-latitude oceans compared to the annual SSTs. However, a colder Indian Ocean-northwestern Pacific was observed in the summer during the 6 ka. All the other PMIP2 OAGCMs also simulate the similar anomalous SST patterns for the 6 ka. Therefore, together with different orbital forcings, these changes in the ocean surface condition would play an important role in regulating summer climate over the Asian monsoon area during the 6 ka.
3.2 Land SAT and Its Response to the Oceanic Modulation
 The Asian summer climate is extremely sensitive to changes in the orbital forcings. As shown in Figure 3, the simulated 6 ka SAT increased over the Asian continent north of 30°N and decreased over Southeast Asia, the Indian subcontinent, and South China. The warming maximums were located over the middle- to high-latitude Asian continent, where the SAT increased by more than 3°C in the summer during the 6 ka. The OAGCMs and IAP4 simulated similar warming patterns over the Asia for the 6 ka summer. For the South Asia, the SAT increased in the northern part whereas decreased in the southern part. The regionally averaged difference in summer SATs between the 6 and 0 ka ranges across the OAGCMs from −0.26°C (MIROC) to 0.45°C (CCSM), suggesting large discrepancies among the OAGCMs simulated SATs there (Table 3). The ensemble mean of them is 0.08°C. Differently, the ensemble mean of the IAP4 simulated SAT increased by 0.38°C. Although the IAP4 was forced by different ocean surface conditions from these four OAGCMs, it simulated changes in SATs for the 6 ka showed smaller discrepancies. For the East Asia, the regionally averaged 6 ka SATs increased by 0.55°C and 0.37°C in the OAGCMs and IAP4, respectively. The similar northern warming and southern cooling pattern was observed in China. The regionally averaged summer SATs increased by 0.68°C (OAGCMs) and 0.49°C (IAP4) in North China, but decreased by 0.09°C (OAGCMs) and 0.27°C (IAP4) in South China during the 6 ka. Compared to the reconstructed temperature [Shi et al., 1993], neither OAGCMs nor IAP4 could reproduce the warming amplitude in China for the 6 ka.
Table 3. Four PMIP2 OAGCMs and IAP4 Simulated Regionally Averaged 0 ka Summer SAT and its 6 ka Changes (Unit: °C) for the Asian Monsoon Domainsa
Each region is defined as: South Asia (20°N–40°N, 70°E–100° E), East Asia (20°N–50°N, 100°E–140°E), North China (30°N–40°N, 110°E–120°E), and South China (20°N–30°N, 110°E–120°E).
denote the differences are significant at the 99%, 95%, and 90% confidence levels, respectively.
 Figure 3 also shows the differences of the simulated SAT between the IAP[Sen] and the IAP[0ka]. A stronger warming pattern was observed over the continent when we used the 6 ka orbital forcings and 0 ka ocean surface conditions in the IAP4. The simulated 6 ka SAT increased over almost the entire Asian continent, except for some cooling over the SASM area. Further investigation suggested that the cooling over the Indian monsoon area was mainly caused by two reasons. One was the reduced net surface solar flux, which was led by the enhanced water vapor transport and associated increase in the total cloud amount. The other was the enhanced latent heat release caused by the increased evaporation over the SASM area (not shown here). That is, the 6 ka cold ocean surface induced significant decrease in SATs over the Asian continent south of 35°N (also see Figure 3, IAP4 [6ka] minus [Sen]). The negative SAT anomalies could reach up to 1.5°C over the Qinghai-Tibetan Plateau. On average, the regionally averaged SATs decreased by 0.5°C and 0.35°C for South Asia and the East Asia, respectively. Therefore, the cold ocean surface in the 6 ka suppressed the orbital forcings induced warming over Asia, particularly in the low-latitude inland regions. Thus, the ocean surface condition over the Indian Ocean-northwestern Pacific could have played a vital role in regulating the surface energy balance over the Asian monsoon area during the 6 ka.
 As stressed in the previous paragraph, all the 6 ka simulations indicated that the Asian continental summer SAT was influenced significantly by the 6 ka forcings. Moreover, the sensitivity simulations designed to reflect the effect of the changes in the ocean surface also showed substantial changes in the continental SATs (Figure 3, IAP4 [6ka] minus [Sen]), which are not intuitively linked to changes in the ocean surface during the 6 ka. To understand the underlying mechanisms behind the response of SATs to the cold ocean surface during the 6 ka, we analyzed the changes in net surface solar flux (NSF), the associated total cloud amount (TCA) and water vapor transport (WVT).
 Figure 4 (top) shows the differences of NSF between IAP4[6ka] and IAP4[Sen]. Significant negative anomalies were observed over North China, western China, and northern India, while positive anomalies were observed over the Indian subcontinent and the high-latitude regions of the Asian continent. This anomalous NSF pattern was mainly attributed to the changes in TCA (Figure 4, middle). Because of the different ocean surface conditions in the IAP4[6ka] and the IAP4[Sen], anomalous circulation enhanced WVT over midlatitude Asian continent (Figure 4, bottom row), and by that to cause increased TCA and lower NSF there. This partly explained negative SAT anomalies over the Asian continent in the IAP[6ka] compared to the IAP[Sen]. In the Qinghai-Tibetan Plateau, the ice-snow feedback could have amplified the effect of the decreased NSF, leading to further cooling in the region.
 In the Asian monsoon area, on the other hand, the sea-land breeze from the ocean toward the continent prevails in the summer. Therefore, changes in the advection could have affected the surface energy balance there. Particularly for the Indian subcontinent and South China, anomalous cold advection induced by the cold ocean surface was another key factor to cause significant surface cooling over land in the 6 ka summer.
4 The Asian Summer Climatic Response
 In this section, we examine further the major atmospheric variables to understand the causes and related changes in the climate over the SASM and EASM areas in the summer during the 6 ka.
4.1 Changes in the Atmospheric Circulation
 Figure 5 (first two rows) depicts the simulated differences of summer sea level pressure (SLP) by the OAGCMs and IAP4 between the 6 ka and the 0 ka. Due to changes in the orbital forcings, the SLP increased significantly over the northwestern Pacific and Indian Ocean. Over the northwestern Pacific, the maximum SLP anomalies increased by 2 and 3 hPa in the OAGCMs and IAP4, respectively. The simulated western Pacific subtropical high was stronger and shifted northward during the 6 ka. On the contrary, the SLP decreased markedly over the central Asian continent. Thus, the land-sea thermal contrast increased in the summer during the 6 ka, particularly in the EASM area, where both the meridional and zonal land-sea thermal contrasts increased significantly. In addition, when we maintained the same 0 ka ocean surface condition for the 6 ka, similar changes in SLPs were still observed in the IAP[Sen] simulations (Figure 5, [Sen] minus [0ka]). Therefore, the enhanced land-sea thermal contrast was mainly caused by the 6 ka orbital forcings alone. However, the influence from different ocean surface conditions on the SLP was not negligible. The 6 ka colder northwestern Pacific-Indian Ocean increased SLPs in the Indian subcontinent, eastern China and Northwestern Pacific (except for the IAP simulations forced by the MIROC ocean surface) while decreased SLPs in the central Asian continent (Figure 5, [6ka] minus [Sen]). As a result of changes in SLPs in the monsoon areas, the land-sea thermal contrasts between the Indian subcontinent, South China, and their adjacent ocean decreased to some extent, which would lead to weaker monsoon circulations. It was suggested that the 6 ka colder ocean surface could weaken the summer monsoon over the SASM area and South China during the 6 ka. In the EASM area, the meridional land-sea thermal contrast was weakened as mentioned before, whereas the zonal land-sea thermal contrast was strengthened when the 6 ka colder ocean surface was considered, which complicated the climatic response of the EASM to the orbital forcings during the 6 ka.
 The altered land-sea thermal contrast over the SASM and EASM areas are expected to yield different monsoon circulation. Here we examine the changes in the wind fields at 850 hPa. Both the OAGCMs and the IAP4 simulated enhanced summer monsoon circulation over India, Southeast Asia, and East Asia in the 6 ka summer (Figure 6, first two rows). The Somali jet, the southwestern flow from the extension of the South Asian monsoon, the cross-equatorial flow from Southeast Asia, and the steering flow from south of the northwestern Pacific subtropical high were all significantly strengthened. The Asian summer monsoon, both the SASM and the EASM, was intensified during the 6 ka. Similar enhancement of the Asian monsoon circulation was also observed in the IAP[Sen] (Figure 6, [Sen] minus [0ka]). It was indicated that the strengthened Asian summer monsoon was mainly attributed to the 6 ka orbital forcings. In addition, the 6 ka cold ocean surface played an important role in regulating the circulation over the entire Asian region. Because of the different responses of land-sea thermal contrast to the 6 ka SSTs over the SASM and EASM areas, anomalously weak summer monsoon circulation was observed over India, Southeast Asia, and South China, whereas anomalously strong summer circulation was observed over North China (Figure 6, [6ka] minus [Sen]). The results suggested that the colder Indian Ocean-northwestern Pacific reduced the strength of the change in the SASM during the 6 ka. For the EASM, the flows from the southwest were suppressed, while the steering flow from the northwestern Pacific and the southerly wind in the North China were strengthened further. Overall, the 6 ka cold ocean surface suppressed most responses of the Asian summer circulation to the orbital forcings, except for a steering flow from the northwestern Pacific.
4.2 Changes in the Precipitation
 Because of the strengthened summer monsoon circulation during the 6 ka, increased precipitation was evident over the Asian continent (Figure 7, first two rows). Both the OAGCMs and the IAP4 simulated significant positive precipitation anomalies over North China, western China, and northern India. For the South China, CCSM, and IAP4 simulated negative precipitation anomalies whereas other OAGCMs simulated positive ones in the 6 ka summer. In the meantime, negative precipitation anomalies were also observed over South China Sea, the coast of western India and the Bay of Bengal. On average, the precipitation increased by 15.1% in the OAGCMs and by 8.6% in the IAP4 over South Asia in the 6 ka summer (Table 4). Nevertheless, the increased precipitation was relatively small for East Asia (3.3% in the OAGCMs and 1.0% in the IAP4). Due to the enhanced EASM, the rainfall belt was pushed northward over eastern China during the 6 ka. Thus, summer precipitation increased by 17.3% in the OAGCMs and 10.4% in the IAP4 over the North China. Over South China, differently, OAGCMs simulated enhanced summer precipitation (5.7%), whereas the IAP4 simulated weakened precipitation (−8.5%) in the 6 ka summer.
Table 4. Four PMIP2 OAGCMs and IAP4 Simulated Regionally Averaged 0 ka Summer Precipitation (Unit: mm d–1) and Its 6 ka Changes (Unit: %) for the Asian Monsoon Domains
denote the differences are significant at the 99%, 95%, and 90% confidence levels, respectively.
 In the comparison between IAP[6ka] and IAP[Sen], we found that the 6 ka cold ocean surface had larger impacts on the summer precipitation in the SASM area compared to the EASM area. The precipitation decreased markedly over the SASM area (−4.4% on average) due to colder Indian Ocean, particularly over the Indian subcontinent and the Bay of Bengal (Figure 7, [6ka] minus [Sen]). The summer precipitation also decreased in Japan, Korean Peninsula, and northeastern China but increased in some other regions of China. On average, the summer precipitation decreased by 0.7% over the North China and decreased by 2.8% over the South China due to the cold ocean surface. This was consistent with the previous results of the further increased summer precipitation forced by warmer SSTs in the Indian Ocean-northwestern Pacific [Wei and Wang, 2004]. In addition, the similar pattern of the anomalous precipitation in the IAP[6ka], IAP4[Sen], and different OAGCMs suggested that the simulated distribution of the 6 ka precipitation anomalies over the Asian monsoon area was mainly caused by atmospheric responses to the 6 ka orbital forcings.
 The outputs from four OAGCMs and different IAP4 simulations were used to investigate several aspects of the simulated Asian summer climate during the 6 ka, focusing on SATs, precipitation, and monsoons. The model results suggested that the orbital forcings induced significant warmer conditions over the Asian continent in the 6 ka summer; a regionally averaged SATs increase of ~0.38°C above the 0 ka value was observed in the IAP4 ensemble results. In addition, the Asian summer circulation was significantly strengthened during the 6 ka. As a result, the transport of water vapor from the Indian Ocean, South China Sea, and northwestern Pacific to the Asian continent was enhanced, causing a marked increase in precipitation over the SASM and EASM areas. The regionally averaged precipitation increased by 15.1% in the OAGCMs and increased by 8.6% in the IAP4 compared to the 0 ka for the SASM area. For the EASM area, the summer precipitation only increased by 3.3% and 1.0% in the OAGCMs and IAP4, respectively.
 The impact of the cold Indian Ocean-northwestern Pacific on the Asian summer climate during the 6 ka was analyzed. The cold ocean surface was found to reduce the strength of the orbital parameters forced warming over the Asian continent in the 6 ka summer. The IAP4[6ka] simulated regional averaged SAT for the SASM and EASM areas decreased by 0.5°C and 0.35°C compared to the IAP4[Sen], although it still remained higher than the 0 ka. The cold ocean surface also played an important role in regulating the Asian summer circulation. According to the comparison between the IAP4[6ka] and the IAP4[Sen], the 6 ka colder SSTs weakened the response of the SASM to the 6 ka orbital forcings, which is in agreement with the conclusions of Zhao and Harrison . However, the 6 ka SSTs partly weakened or amplified the response of the EASM to the 6 ka orbital forcings. Therefore, the strengths of the change in the entire SASM and most components of the EASM, including the southwestern flow from the extension of SASM and the cross-equatorial flow from Southeast Asia, were weakened. Nevertheless, the steering flow from south of the northwestern Pacific subtropical high, a component of the EASM system, was further strengthened by the 6 ka colder SSTs. Overall, the cold ocean surface weakened the responses of the Asian summer circulation to the orbital forcings during the 6 ka.
 On the other hand, this anomalous circulation also resulted in significant changes in WVT over the Asian monsoon area, and by that to regulate TCA and associated NSF. As a result, this anomalous surface energy budget, together with the cold advection due to 6 ka cold ocean surface condition, caused the cooling (weakened warming) over the low-latitude (midlatitude) Asian continent when using different ocean forcings in the 6 ka simulations. Correspondingly to the anomalous circulation, the summer precipitation decreased by 4.4% and 2.8% over the SASM and EASM areas, respectively. These changes were comparable with 6 ka forcings increased summer precipitation, 8.6% for the SASM area and 1.0% for the EASM area. Thus, the effects of different ocean surface conditions on the 6 ka summer precipitation were not negligible. A similar result has been found by Ohgaito and Abe-Ouchi [2007, 2009]. In addition, the different strengths of the change in summer precipitation between these two monsoon areas implied that the summer precipitation in the SASM area is more sensitive to the changes in the orbital forcings and ocean surface conditions than that in the EASM area. This kind of difference also can be observed in the OAGCMs.
 Numerous studies have addressed that the orbital forcings are the dominant factors to shape the anomalous Asian summer climate during the 6 ka [e.g., Wang et al., 2010; Zhou and Zhao, 2010; Zhao and Harrison, 2012; Jiang et al., 2013]. On this basis, our results further emphasized that the 6 ka ocean surface condition played an important role in altering the summer climate over the Asian monsoon area, particularly for the inland surface energy balance. According to previous studies, the current models still could not simulate the reconstructed warming amplitude in China (annual temperature increased by 2°C ~ 3°C in eastern China and 4°C ~ 5°C on the Qinghai-Tibetan Plateau) during the 6 ka [Wang et al., 2010; Jiang et al., 2012]. The effect of the cold ocean surface in the 6 ka summer could further enlarge the discrepancies in the model-data comparison. Although it could be partly attributed to the uncertainties in the proxy data, the model-data mismatch implied a large degree of uncertainty in the simulated 6 ka Asian climate. As pointed out by Ohgaito and Abe-Ouchi , how to improve the simulation of the 6 ka ocean would be a key to better simulate and understand the Asian climate changes during the 6 ka.
 We sincerely thank the three anonymous reviewers and Dr. Steven Ghan for their helpful comments and suggestions on the earlier versions of the manuscript. We are grateful to Prof. Dabang Jiang, Prof. Jianqi Sun, Dr. He Zhang, Dr. Zhongshi Zhang, and Dr. Tonghua Su for fruitful discussions. Also, we acknowledge the international modeling groups for providing their data for analysis, and the Laboratoire des Sciences du Climat et de l'Environnement (LSCE) for collecting and archiving the model data. This research was supported by the Strategic Priority Research Program (XDA05120703) of the Chinese Academy of Sciences, the National Basic Research Program of China (2010CB951901) and the National Natural Science Foundation of China (41205051). The PMIP2/MOTIF Data Archive is supported by CEA, CNRS, the EU project MOTIF (EVK2-CT-2002-00153) and the Programme National d'Etude de la Dynamique du Climat (PNEDC). More information is available on http://pmip2.lsce.ipsl.fr/ and http://motif.lsce.ipsl.fr/.