The Indian monsoon system is a complex phenomenon. Its seasonal evolution and variability on multiple time scales involve individual and coupled land, ocean, and atmospheric processes, in addition to the annual progression of the Earth’s revolution around the Sun. From June to September, the country receives roughly 75% of the annual rainfall (Supplementary Figure 1). The seasonally phase-locked rainfall and associated circulation are referred to as the southwest monsoon or Indian summer monsoon (ISM). Any changes in the average seasonal rainfall patterns have a profound effect on agriculture, and therefore the livelihood of 1.5 billion people in South Asia (Gadgil and Rupa Kumar, 2006). The Indian summer monsoon rainfall (ISMR) is weakening over several regions (e.g., Krishnan et al., 2020), with a simultaneous increase in the extreme events in the recent ∼150 years (Boyaj et al., 2020, and the references therein). Observation-based studies in the last 50–100 years suggest that, on interannual time scales, tropical oceanic phenomena such as the ENSO dominantly influence the ISM variability (see Webster et al., 1998, and the recent reviews by Ashok et al., 2019; Mohanty et al., 2020, and the references therein). However, Proxy studies indicate that ISMR also varied on the decadal, centennial, and millennial to multi-millennial time scales, in addition to intra-annual, and interannual scales (e.g., Ramesh et al., 2010; Chakraborty et al., 2012). The relevance of internal climate variabilities, such as ENSO or multi-decadal variations in the tropical pacific circulation, is shown to be important even in the last millennium, through analysis of PMIP3 outputs (e.g., Tejavath et al., 2020).

On the other hand, external forcings are suggested to play major roles in manifesting various past climate periods. For example, changes in the strength of monsoon over the glacial-interglacial transition are suggested to be due to changes in orbital parameters (Kutzbach and Otto-Bliesner, 1982; Kutzbach and Guetter, 1986). Earth’s precession cycle of 23 kyr plays a dominant role in modulating the tropical precipitation by changing the seasonal and meridional distribution of incoming solar radiation (Kutzbach, 1981; Pokras and Mix, 1987). Speleothem reconstructions suggest that changing precession leads to antisymmetric precipitation patterns in the southern and northern hemispheres (Wang et al., 2007). Several proxy and modeling studies show that the strength of the monsoon has changed in proportion with changes in insolation on orbital scales (Kutzbach, 1981; Tuenter et al., 2005; Wang et al., 2007; Kutzbach et al., 2008; Weber and Tuenter, 2011; Shi, 2016; Kathayat et al., 2017).

The distinct external forcings have been said to have manifested the MH climate in India. A study by Gupta et al. (2005) shows that slight changes in solar radiation have brought changes in the tropical monsoon systems during the Holocene. Studies also show that Earth’s climate is sensitive to small changes in solar radiation on the centennial to the millennial time scale during the Holocene (Rind and Overpeck, 1993; Shindell et al., 2001; Hu et al., 2003). Signatures of changes in insolation are reflected in the changes of the northern hemisphere monsoon systems (Staubwasser et al., 2003). Past climate records based on proxy reconstructions from India and neighborhood surroundings of the Indian region (see the review by Dixit and Tandon, 2016; Banerji et al., 2020; Tejavath et al., 2020 for the details and references) report centennial to millennial-scale changes in the ISMR during the Holocene (≃ 12 kyr BP to present). These studies, in general, show that from the early Holocene period (9 kyr BP) to late Holocene (3 kyr BP), the ISMR shows a long-term decreasing trend. A single model study by Kumar et al. (2019) and a multi-model study by Tejavath et al. (2020) using PMIP3 (Paleoclimate Modelling Intercomparison Project 3) show that during the mid-Holocene (MH; ≃ 6 kyr BP), the ISM was stronger relative to the present day. An analysis of the PMIP3 simulations by Tejavath et al. (2020) suggests that a robust large-scale convergence over the Indian region at 850 hPa may have been the reason for strong ISM during the MH. A study by Cobb et al. (2013) purports a highly variable ENSO during the Holocene period. However, several other proxy reconstruction studies (e.g., Gill et al., 2016, 2017) suggest that there was reduced ENSO activity during the MH period, and dominated by La Niña type SSTs over the tropical eastern pacific. There is another argument that the ENSO activity during the MH was indeed relatively subdued (An and Choi, 2014; Chen et al., 2019), and that such suppression is due to the enhanced Asian monsoon activity (Chen et al., 2019; Crétat et al., 2020). An ocean-atmospheric coupled model study by Crétat et al. (2020) shows orbitally driven trends during the MH to the current day (i.e., 0 kyr BP). The study also claims an increased influence of ENSO on ISMR and Indian Ocean Dipole (IOD) during the MH.

Studies based on paleo reconstruction data identify two significant periods in the last millennium (LM). These two periods, commonly known as Medieval Climate Anomaly (MWP; CE 950–1350) (e.g., Lamb, 1965)—a relatively warmer period, followed by a relatively more cooling period referred to as the Little Ice Age (LIA, CE 1500–1850) (e.g., Grove, 1988). Studies based on proxy and modeling show wetter and stronger ISM during MWP, and drier and weaker ISM during LIA relative to the LM mean (see Dixit and Tandon, 2016; Tejavath et al., 2019 for further references).

Various mechanisms have been proposed for ISM variability during the Holocene period, in addition to the changes in the orbital parameters (e.g., precession). For example, during the early Holocene climatic period, the mean latitudinal position of the Intertropical convergence zone (ITCZ) shifted northwards, resulting in an increase in the ISMR in response to the changes in the solar insolation (Fleitmann et al., 2007). Early climate model sensitivity experiments show that the timing of Northern Hemisphere (NH) summer monsoon strength has, or should have the near-zero phase relative to maxima of NH precession-driven radiation (Clemens and Prell, 2007). A proxy reconstruction study by Dixit et al. (2014a, b) shows that solar insolation played a significant role in strengthening the ISM during the early Holocene period. A study by Polanski et al. (2012) based on proxy and modeling shows that solar insolation changes strongly influenced the MH climate. Several other modeling studies suggest enhancing the northern hemisphere monsoon systems due to changes in insolation (Braconnot et al., 2007a,b; Bosmans et al., 2012; Marzin et al., 2013; Zheng et al., 2008). A recent modeling study suggests that a decrease in Saharan vegetation may have resulted in droughts in northernmost parts of India and excessive rains in southern India (Griffiths et al., 2020). A study by Leuschner and Sirocko (2003) shows that precessional and obliquity cycles coincide with the major events observed in the insolation-based ISM index. Interestingly, a study by Li and Harrison (2008) suggest in their coupled model simulations, the orbital forcing mostly enhances the Asian summer precipitation but in contrast, the neighboring ocean reduces the orbitally-induced summer precipitation, leading to the postponement of the time of summer monsoon onset over the Asian monsoon region.

However, the above studies including the coupled modeling studies, have not specifically examined the relevance of the external forcings to the ISM during MH in detail. Furthermore, the relative variations in the ISM during MH and LM have been less-studied, particularly from the modeling perspective. Keeping this in mind, using the community atmospheric model version 5 (Hersbach et al., 2015), we carry out several time-slice experiments for MH, MWP, LIA, and Historical (deemed as “present-day”) periods. Importantly, we explore the sensitivity of the simulations by changing the orbital parameters; a major external forcing said to be important for the MH climate period, as discussed in detail in the next section.

This manuscript is organized as follows. In the following section, we briefly describe the model, our experimental setup, and methodology. In section “Results,” we present our results and discussion. We present our conclusions and future scope for related studies in the final section, which is section “Conclusions and Scope for Future Studies.”

The Community Atmospheric Model Version 5 (CAM5), is the atmospheric component of the well-known Community Earth System Model Version 1.2.0 (CESM1.2.0; Hersbach et al., 2015). The CESM1.2.0, formerly also known as CCSM4, is a fully coupled climate model used to generate simulations of the Earth’s past, present, and future climate states. For our experiments, we configured the horizontal resolution of the CAM5 at 1.9° latitude × 2.5° longitude grids, with 30 vertical levels. This version of CAM is comparable and compared to the CESM simulations for the Coupled Model Intercomparison Project 5 (CMIP5) and the Paleo Model Intercomparison Project 3 (PMIP3) simulations. The PMIP3 is a collective initiative endorsed by the World Climate Research Programme (WCRP) and JSC/CLIVAR working group on coupled models and the International Geosphere and Biosphere Programme (IGBP; PAGES) (Braconnot et al., 2012).

We carried out four suites of multi-ensemble control simulations of a 30-year span for each of the MH, MWP, LIA, and Historical (Present day) periods. Each suite of the control experiments in turn contains three ensembles. Each ensemble simulation starts with January initial conditions that are different from one ensemble to another and continues for a year. The external boundary conditions used in this study are period-specific (i.e., MH, MWP, LIA, and HS), similar to those mentioned in the PMIP3 website https://pmip3.lsce.ipsl.fr/, and briefly presented in Table 1. The climatological SSTs used in this study for each period are adapted from the simulation outputs from the corresponding PMIP3 simulations that were generated with CCSM4 (Figure 1). Apart from the control simulations, we have also carried out multiple ensemble simulations for the MH period by forcing the model with different orbital forcings for different periods, e.g., present-day orbital, 8.2 kyr BP orbital, and LGM orbital forcings. The forcings are mentioned in Table 2. For comparison, the same set of initial conditions have been used in all the complementary experiments.

Further to these 30-year long control and orbital sensitivity simulations, we have also carried out suites of simulations of 1-year span (simulation starts in January and ends in December) with 10-ensembles each with the same set of initial conditions across various orbital forcings. The results from the 1-year length simulations are in very good agreement with those from the 30-year long simulations, as evidenced by the evolution of the seasonal cycles (Supplementary Figure 2). Therefore, we shall not discuss the results from these runs.

Apart from the control and orbital sensitivity experiments, we have also carried out two sensitivity experiments to gauge the potential relevance of the El Niños during MH, and two more to assess the impact of the La Niñas during the same period. This involved, for example in case of the La Niña sensitivity experiments, imposing the La Niña anomalies on the climatological SSTs over the tropical pacific ocean (120°E to 80°W and 30°S to 30°N) to obtain the lower boundary SSTs representing the La Niña SSTs during the MH, and in another experiment, imposing SST anomalies throughout the tropical Indo-pacific ocean (40°E to 80°W and 30°S to 30°N). The SST anomalies imposed in these SST sensitivity experiments for the MH period are shown in Supplementary Figure 3. The SST anomalies during the La Niñas (El Niños) were obtained by compositing monthly SST anomalies of all “typical” La Niñas (El Niños). These typical events are identified as those for which the magnitude of the simulated JJAS NINO3 index is above (below) one standard deviation.

In our analysis, the climatology is taken from the entire 30-year runs, because the initial conditions are taken from PMIP3 CCSM4 simulation, and they are therefore spun up well. As we are starting our simulation from January, the simulated atmosphere will adjust to the lower boundary SSTs within a couple of months. As it is, such experiments for the current day monsoonal season typically start in the month of May (e.g., Ashok et al., 2001, 2004, 2009, 2012; Guan et al., 2003).

We have calculated the monthly climatological cycles of rainfall and surface temperatures over the Indian land region bounded by 66.5°E–101.5°E; 6.5°N–39.5°N. We analyze the simulated fields of velocity potential-which represent divergence, vorticity, and moisture flux convergence. The moisture flux convergence has been computed as the sum of moisture convergence and advection. The fields of velocity potential, vorticity, and moisture convergence have been computed from the simulated outputs of horizontal winds and moisture through the application of spherical harmonics using the NCL routines. The details can be found at https://www.ncl.ucar.edu/Applications/wind.shtml. The orbital parameters are calculated using the PMIP3 protocol https://pmip3.lsce.ipsl.fr/ and are present in Supplementary Table 1.

While several proxy studies, such as Rawat et al. (2015) and Band et al. (2018) suggest that the ISM during the mid-Holocene (MH; ∼6000 yr BP) was characterized by a wet climate relative to the modern time. The few modeling studies (Kumar et al., 2019; Tejavath et al., 2020, etc.) also suggest that the ISM was indeed stronger at that time. Interestingly, proxy-based (Mukherjee et al., 2016) and model-based studies suggest that external factors such as changes in the orbital parameters, changes in solar forcing induced by volcanic eruptions, land surface, and vegetation changes, etc., may have played an important role in evolving a distinct climate of the Earth during MH relative to the current day (Crétat et al., 2020). The present study, through various experiments with the Community Atmospheric Model (CAM5), examines the potential role that the distinct orbital parameters during the MH may have played in the manifestation of mean ISM conditions during the MH.

We have carried out 30-year long multiple ensemble simulations for MH, MWP, LIA, and Historical period (Present-day; HS) climate periods. The results show that the ISMR during the MWP is higher than that of the LIA, but lesser than that of the HS, in agreement with various PMIP3 coupled simulations (Tejavath et al., 2019). Our AGCM experiments simulate relatively higher surplus ISMR during the MH period relative to the HS, supporting proxy observation-based studies and from few modeling studies.

Results from our novel experiments show that the higher ISMR simulated in the MH is due to a stronger monsoon circulation, northward migration of the ITCZ, strengthening of the subtropical high over the western pacific relative to the HS. To examine the potential role of orbital parameters on the ISM variability during the MH, we carried out several sensitivity simulations for the MH period by changing the orbital parameter with those of the HS, LGM, and 8.2 kyr BP. These sensitivity experiments demonstrate that the higher orbital parameter during the MH, which meant more solar insolation in the northern hemisphere played a significant role in enhancing the ISM, and consequently, high ISMR.

This suggests that the higher insolation during the MH, around 6 kyr BP, associated with relatively favorable external forcings (e.g., orbital parameters) likely played a significant role in enhancing the ISM during the MH compared to the HS. Of course, this is by no means a complete explanation for the high MH rainfall simulated and recorded in several proxies. For example, we have not carried out any experiments to ascertain the potential role of land-surface and vegetation changes in the mid-Holocene, which is said to have facilitated the mid-Holocene climate in West Africa (Messori et al., 2018; Crétat et al., 2020; Griffiths et al., 2020). Such experiments need a model with an interactive land-surface and vegetation model. The other factors, such as the internal variability such as ENSO, whether generated due to internal coupled dynamics or forced by orbital forcings may have played their roles. In this context, a few sensitivity simulations carried out by us suggest that our AGCM simulates below normal ISM rainfall during the MH as a response to concurrent El Niño-like SSTs imposed in the tropical Pacific (e.g., Ashok et al., 2004). This is similar to the current day ENSO impact on ISM. But we elicit a relatively low Indian summer rainfall response during the months of June and July when we force the model La Niña type of SSTs in the tropical Pacific ocean or even when the La Niña-associated concurrent SST anomalies are also imposed in the tropical Indian Ocean. Importantly, similar La Niña experiments for the MWP and LIA result in a position ISMR anomaly. This suggests that, at least in our simulations, the La Ninas have not significantly contributed to the relatively wet ISM during the MH, and that the orbital forcings may be relatively more important.

To get a comprehensive idea of the relative roles of various internal and external factors and their influences on the ISM during the MH, a more extensive study with a fully coupled land-biosphere-ocean-atmospheric model will be carried out soon. We also plan to revisit the above results by repeating these experiments with multiple AGCMs in order to ascertain our claims in this paper.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

CT took lead in carrying out the simulations, analyzed the data, and wrote the manuscript. KA and SC supervised and co-wrote the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.