1 Introduction

One of the mostly concerned issues, especially under a global warming scenario, is the future trend of the Indian summer monsoon (ISM) rainfall. Although variations of the ISM on different timescales are well studied, the trend during the past 100 and 200 years is strongly debatable. For instance, most climate model simulations show increasing ISM rainfall trends under a global warming scenario (Meehl and Washington 1993; Douville et al. 2000; May 2002). Sediment records from the Arabian Sea also indicated a significant increasing ISM trend during the past 400 years (Anderson et al. 2002). However, the meteorological ISM rainfall data all over India show no clear long term trend from 1871 to the present. Moreover, increasing evidence shows declined rather than intensified ISM rainfall during the past 100 and 200 years. For example, the meteorological rainfall records of northeastern India, Nepal, and Bangladesh, show a clear decreasing trend during the past century (Duan et al. 2004; Naidu et al. 2009). The past 50 years witnessed an even severe decline of monsoon rainfall over North Africa, e.g. Sahel (Giannini et al. 2003) and parts of monsoonal China (Xiang et al. 2007). It is also notable that the ISM intensity recorded in Dasuopu ice core shows a decreasing trend from about 1850 to the present (Duan et al. 2004; Thompson et al. 2000). Moreover, a speleothem ISM record in Oman (Burns et al. 2002) also shows a general decreasing trend during the past century. Considering these contradicting results, what is the real ISM trend during the past 100 and 200 years, and what is the underlying cause? It is urgent to develop more local and regional ISM records to make clear the trend of ISM during the past 100 and 200 years, and to understand the nature of this trend and how it might continue in the near future.

2 Background and method

During boreal summer, ISM enters into China through the “water vapor channel”, southeastern Tibetan plateau, and moves northwards along the eastern edge of the Tibetan plateau (Fig. 1). Tree-ring cores (Picea likiangensis var. rubescens) were collected using an increment borer at Shuajingsi town (32°01′N, 102°37′E; Alt.: around 3500 m near the timberlines), Hongyuan County 2004 (Xu et al. 2010). Mean annual temperature at Hongyuan is 1.1 °C, and mean January and July temperature are −10.3 and 10.9 °C, respectively. Mean relative humidity is about 70 ± 3 %. Mean annual precipitation is 700 mm, and about 80 % is centered during June to September.

Fig. 1
figure 1

Sampling sites. Red star denotes the tree-ring sampling site at Shuajingsi town (SJS), while yellow dashed curve and filled green stars indicate the transect and the carex sampling sites. Yellow circles denote the locations of Hongyuan (HY), Songpan (SP), Guangyuan (GY), Chengdu (CD), and Yushu (YS). Also shown are locations of Dasuopu (DSP) and Rongbuk (RBK) ice cores (filled blue triangles). Arrow-lines in upper panel show the streamlines averaged from June to August at 850 hPa during 1968–1996 based on the NCEP/NCAR reanalysis data (Kalnay et al. 1996). Gray arrow in lower panel sketchily shows the “vapor channel” at the southeast Tibet plateau

Full size image

Cross dating was performed following the skeleton plot method (Stokes & Smiley 1968) and twelve cores without missing and false rings from six trees were selected for stable isotope analysis (Xu et al. 2010). The selected tree-ring cores were sampled with annual resolution and then each year was pooled (Borella et al. 1998), frozen with liquid nitrogen and hammered to powder. For larger rings an attempt was made to visually trim equal masses from each year into the pool, while for narrow rings whole years were pooled together regardless of mass, which may result in trees with narrower rings having less impact on the final δ18O value of the pooled sample. Alpha cellulose was prepared using a modified sodium chlorite oxidation method (Hong et al. 2000), including a final NaOH rinse of the samples. Delta 13C of these α-cellulose samples were previously studied, and the results verified the reliability of both the chronology and the pre-treatments (Xu et al. 2010). In this study, we used these α-cellulose samples to measure oxygen isotopes on a mass spectrometer (Finnigan Mat-252), with experimental precision better than ±0.2 ‰. We measured each sample more than twice, and got an average δ18O value for each calendar year. Experiments were performed in 2007 at the State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences. The last 50 samples (sample 1–50; covering 1955–2004) were repeatedly measured 2009 at the same laboratory to verify the reproducibility.

Another two cores, core SJS-17 and SJS-12, also without missing and false rings, but with very narrow ring widths for some sections, were cut off every five rings. Delta 18O series of these two individual cores were established in 2009 using the same method to verify the robustness of chronology and data measurements. SJS-17 series covers 1600–1898, with two missing points due to the insufficiency of samples; while SJS-12 covers 1895–2001. As shown in Fig. 2, the Hongyuan δ18Otree ring series highly synchronize with δ18O series of the other two individual cores, supporting both the reliability of chronology and the reproducibility of δ18O measurements.

Fig. 2
figure 2

Hongyuan tree ring δ18O timeseries. Gray and pink curves represent the Hongyuan δ18Otree ring series and the duplicates of the last 50 years, respectively. The blue curve denotes five year running mean of Hongyuan δ18Otree ring series. Green curve with blank circles and orange curve with filled triangles represent δ18O series of SJS-17 and SJS-12 samples (see text), respectively

Full size image

In 2007, we also selectively collected living sedges (Carex sp.) growing without groundwater supply along the easten edge of the Tibet plateau (Fig. 1) and measured δ18Ocarex, 2009, using a similar method than for the tree ring samples. The purpose was to understand the “δ18Orainfall–climatic factor” relationship from the “δ18Ocarex–climatic factor” relationship, assuming δ18Ocarex and δ18Orainfall have similar spatial variations (see below).

3 Climatic significance of Hongyuan δ18Otree ring

As already demonstrated by Dansgaard (1964), temperature is generally the dominating factor controlling δ18Orainfall at high latitudes (examples can be seen in Greenland and Antarctica ice cores); while precipitation amounts significantly influence δ18Orainfall at low latitudes. At middle to middle-low latitudes, δ18Orainfall is generally influenced both by temperature and precipitation. From an oxygen isotopic fractionation perspective, if the oxygen isotopic fractionation is dominated by temperature, then δ18Orainfall is expected to be positively correlated to temperature (δ = aT + b, where a > 0). On the other hand, if δ18O is dominated by the precipitation amount effect, then δ18Orainfall should be negatively correlated to precipitation (Dansgaard 1964). Observation data show that monthly δ18Orainfall at Chengdu area is negatively correlated to precipitation (r = −0.53, α = 0.11; Fig. 3a). In summer, δ18Orainfall is low and precipitation is high; while in winter, δ18Orainfall is high but precipitation is low (Fig. 3a). This suggests that δ18Orainfall is dominated by the amount effect around the study area, and therefore is linked to monsoon precipitation. The δ18O variations in speleothem carbonates in southern China have also been suggested to be controlled by precipitation, and are widely used as indicators of monsoon intensity (Yuan et al. 2004; Wang et al. 2005; Zhang et al. 2008). Weak ISM is related to low precipitation and high δ18Orainfall values; while strong ISM corresponds with high precipitation and low δ18Orainfall values.

Fig. 3
figure 3

a Monthly rainfall δ18O and precipitation at Chengdu (see location in Fig. 1), b averaged May to September relative humidity and precipitation of Hongyuan during 1958–2003. c scatter plots between Carex δ18O and precipitation along the eastern edge of the Tibetan Plateau (Fig. 1). d. scatter plots between tree ring δ18O and relative humidity (five years running mean)

Full size image

For plants that directly use rainfall water as source water, the δ18O in plant cellulose is largely enriched compared with δ18Orainfall. This enrichment, called as biological fractionation, is relatively constant (about 27‰), and therefore the variability of δ18Oplant is mainly derived from that of δ18Orainfall (Roden et al. 2000; Miller et al. 2006; Xu et al. 2006). As shown in Fig. 3c, the spatial variation of δ18Ocarex is negatively correlated to precipitation. Since Carex we selectively collected has short roots and uses rainfall water directly as source water, the δ18Ocarex variability can be used as a substitute of δ18Orainfall variability. Therefore, a negative spatial correlation pattern of “δ18Orainfall–precipitation” can be inferred from that of “δ18Ocarex–precipitation”, providing independent evidence that the rainfall δ18O around the study site is dominated by precipitation, and thus ISM intensity.

As shown in Fig. 3d, the Hongyuan δ18Otree ring is significantly correlated to relative humidity. Since relative humidity and precipitation show similar annual variations (Fig. 3b) and monthly distribution patterns (correlation coefficient is 0.932, α < 0.001; figure not shown), the Hongyuan δ18Otree ring can be related to monsoon precipitation, and thus ISM intensity. Higher δ18Otree ring relates to weaker ISM intensity, and vice versa. Such “δ18Otree ring–monsoon intensity” relationship is also consistent with that found by Sano et al. (2010) in the Nepal Himalaya and that found by Grießinger et al. (2011) in the southern Tibetan plateau most recently.

4 Comparison of δ18Otree ring with other precipitation records

One of the direct rainfall archives are ice cores. The accumulation rate of Rongbuk ice core, which is closely related to the northeastern Indian summer precipitation, recorded the variation of ISM intensity during the past about 160 years (Zhang et al. 2004). As shown in Fig. 4, the Hongyuan δ18Otree ring curve shows high correlation with the Rongbuk accumulation rate curve (r = −0.39; α < 0.001), supporting the use of δ18Otree ring as an indicator of monsoon precipitation. Thompson et al. (2000) suggested that the dust content in Dasuopu ice core is a sensitive indicator of ISM intensity. As shown in Fig. 4 (see shaded columns), most of the climatic events recorded in Dasuopu ice core also correlate well with those inferred from Hongyuan δ18Otree ring, especially from 1860 to the present. The Chinese Drought/Flood (D/F) indices, which were derived from historical literature records (Zhang et al. 2003), have been widely used as precipitation proxy indices. As shown in Fig. 4, the climatic events recorded in δ18Otree ring are generally consistent with those recorded in the D/F index of Guangyuan. Similar climatic events can also be found in the Oman speleothem record regardless of some small age difference during the past 300 years (Burns et al. 2002).

Fig. 4
figure 4

Comparison between precipitation indices of Hongyuan δ18Otree ring, Rongbuk ice accumulation (Zhang et al. 2004), Dasuopu dust concentration (Thompson et al. 2000), Drought/Flood index of Guangyuan (5-years running mean) (Zhang et al. 2003), and δ18O of Oman speleothem S3 (Burns et al. 2002). Shaded columns indicate ISM failures and dotted lines indicate the long-term rainfall trend. All curves (except the Drought/Flood index) are annually scaled with five-point running means

Full size image

The ISM inferred from δ18Otree ring dropped dramatically during 1760–1800, and reached the lowest values at about 1788–1795, which is very similar with that recorded in speleothem (Burns et al. 2002). At about 1867, ISM decreased again, and was minimum about 1879, consistent with the widely reported “1877 Indian big famine” (Pant and Rupa Kumar 1997). Such a clear monsoon failure can also be seen in ice core records and the Drought/Flood indices. At about 1895–1902, a strong ISM reduction can be detected in the δ18Otree ring curve, similar to those recorded in Rongbuk and Dasuopu ice cores, Guangyuan D/F index, and Oman S3 (Fig. 4). Around 1927 and 1942 are another two obvious ISM reductions, which are also clear in Rongbuk and Dasuopu ice cores and in the D/F index series. The δ18Otree ring and D/F indices show significant drops from the middle 1960s, reaching the twentieth century minimum at the middle 1970s, in accordance with the Rongbuk snow accumulation curve (Fig. 4). This strong monsoon drop can also be detected from meteorological records over a wide geographical extension, from North Africa (e.g. Sahel) to northern India, to southwest China, and even to north China (Yan et al. 1990).

One striking feature of the δ18Otree ring curve is that from 1700 to about 1860, the δ18Otree ring shows no clear long term trend, while an obviously decreasing trend can be seen from about 1860 to the present (Fig. 4). The dust content in Dasuopu ice core shows a similar long term behavior with δ18Otree ring, i.e. generally stable from 1700 to 1860, and decreasing from about 1860 to the present. The accumulation rate at Dasuopu, which is highly correlated with the northeastern Indian precipitation (Duan et al. 2004), also shows a significant decrease after about 1850. Such decreasing trends can also be found in southern Oman speleothem records during the past century (Fig. 4) (Burns et al. 2002). Meteorological rainfall records from northeastern India (Naidu et al. 2009), Nepal, and Bangladesh (Duan et al. 2004) also show decreasing trends during the past 100 years. Summer precipitation inferred from tree ring δ18O in southern Tibet also indicates a decreasing monsoon precipitation trend during the past 100 and 200 years (Grießinger et al. 2011). These evidences illustrate that the ISM intensity during the past 100 and 200 years is gradually decreasing rather than increasing.

5 Discussion

5.1 Periodicities in δ18Otree ring series

On short term timescales, more and more evidence shows strong linkages between annual variations in ISM and the El Niño Southern Oscillation (ENSO) (Kumar et al. 2006). The interannual rainfall in Sahel region is also obviously influenced by ENSO (while its low frequency variability is closely linked to SSTs of the tropical Ocean; see below) (Giannini et al. 2003). We previously found an inverse relationship between temperature at the study site and SSTNiño3 (an indicator of El Niño events) on multi-annual scale, which is attributed to the potential heating of ISM rainfall associated with ENSO (Xu et al. 2010). However, the annual variations in δ18Otree ring do not show significant correlation with SSTNiño3. The power spectrum analysis also failed to detect significant high frequency components in the δ18Otree ring series. This is possibly due to the influences of some local processes on source water δ18O on short term timescale, e.g. evaporation and hydrological mixture between surface water and soil water, which masks the high ISM frequency.

It is interesting to show by Fast Fourier Transform (FFT) analysis that there exist three significant periodicities in δ18Otree ring series, namely 110.9, 59.0, and 22.6-years (Fig. 5). A significant 22 year cycle can also be obtained from the FFT analysis of the Rongbuk ice accumulation series (calculated from Zhang et al. 2004). Hiremath and Mandi (2004) found a 22-years cycle in ISM rainfall. Duan et al. (2004) found a 25-years cycle in the Dasuopu ice accumulation series. These similar cycles suggest a common quasi-22-years cycle in the ISM variability, which is possibly a response to the 22-years variation of solar activity. FFT analysis shows a significant 25-years cycle in the Yushu (see location in Fig. 1) temperature series reconstructed from tree ring widths (Liang et al. 2008). The coherence on 22-years cycle between our δ18Otree ring series and the Yushu temperatures (Fig. 5) further validates the application of δ18Otree ring as an indicator of precipitation. An interpretation of the δ18Otree ring as an indicator of temperature would result in an anti-phase relationship between temperature variations at Hongyuan and Yushu, which contradicts the high in-phase relationship as indicated by the meteorological records (figure not shown).

Fig. 5
figure 5

a Periodicities (a and b) in Hongyuan δ18Otree ring series (examined by Fast Fourier Transform method on the software of Autosignal). Green, red, and blue curves denote 111, 59.0, and 22.6-years cycles respectively. c, d, e show the comparison between 22.6-years component of Hongyuan δ18Otree ring and 25-years components of Yushu temperature reconstructed from tree ring width (Liang et al. 2008), that between 59.0-years component of Hongyuan δ18Otree ring and the 21-years running mean of June–Sept Indian summer monsoon rainfall (Parthasarathy et al. 1995), and that between 111-years component of Hongyuan δ18Otree ring and δ18O of S3 (Burns et al. 2002). See text for details

Full size image

The significant 110.9-years cycles in δ18Otree ring series is a quasi-100-years cycle, which has also been detected in Hongyuan peat records near the study site (Xu et al. 2006). This 110.9-years component shows very similar trend with the ISM record from Oman (Burns et al. 2002) (Fig. 5), supporting strongly that the low frequency climate at Hongyuan area is dominated by ISM. This typical period in ISM variations is also detected in speleothem records of Dongge cave (Wang et al. 2005), and in sediment records from the Arabian Sea (Anderson et al. 2002; von Rad 1999; Agnihotri et al. 2002), suggesting that the quasi-100-years cycle is a common periodicity in the ISM regions.

The 59.0-years cycle is also clear in δ18Otree ring series and is similar in pattern with the quasi-60-years cycle in ISM rainfall record (Fig. 5) (Parthasarathy et al. 1995; Agnihotri and Dutta 2003). Similar periodicities can also be found in the Dongge cave speleothem record (Dykoski, et al. 2005), and in the cross spectrum between δ18Ostalagmite and atmospheric 14C concentration in both Dongge cave (Wang et al. 2005) and Hoti cave (Neff et al. 2001). A 62-years cycle can also be detected in the FFT spectrum of the Rongbuk ice accumulation series (calculated from Zhang et al. 2004). The ISM proxy indices in northeastern Arabian Sea also show similar periodicities (von Rad 1999; Agnihotri et al. 2002).

A lot of the studies show coherence between solar activity and ISM on centennial to decadal timescales. For example, Wang et al. (2005) showed that the decadal to centennial scale Asian summer monsoon intensity synchronizes with solar activity. Neff et al. (2001) suggested that “one of the primary controls on centennial- to decadal-scale changes in tropical rainfall and monsoon intensity during this time are variations in solar radiation”. Agnihotri et al. (2002) showed that on decadal and centennial scale the ISM shows similar variations with solar activity (Periodicities of 200, 105 and 60 yr are observed in the proxy records coinciding with those known for solar cycles). Fleitmann et al. (2003) suggested “decadal to multidecadal variations in monsoon precipitation being linked to solar activity”. The 22.6, 59.0, and 110.9-years periodicities extracted from Hongyuan δ18Otree ring series seem to be linked to solar activity.

5.2 Possible mechanism of the decreasing ISM trend

As shown in Fig. 6, the decreasing trend of ISM inferred from our δ18Otree ring series is inversely correlated with the increasing trend of total solar irradiance (Lean 2000) during the past 100 and 200 years. Duan et al. (2000) also found a negative correlation between the long term trend of Dasuopu ice accumulation and solar activity. This suggests that solar activity is at least not the direct forcing for the decreasing ISM trend from 1860 to the present. An increasing temperature trend at the study site (inferred from δ13Ctree ring, Xu et al. 2010) has been reported most recently, consistent with the observed global warming (Fig. 6). The meteorological records around the study site also show increases in temperature but slight decreases in precipitation. For example, the mean annual temperature increased by about 0.95 centigrade during the past 50 years, while the mean annual precipitation decreased by about 40 mm at Hongyuan during the same interval. Why does the ISM intensity decrease in opposition to an increasing temperature trend? What is the nature of the long term decreasing ISM trend during the past 100 and 200 years?

Fig. 6
figure 6

Comparison between Hongyuan δ18Otree ring (green; precipitation proxy index; see text for details), total solar irradiance (TSI, red) (Lean 2000), Hongyuan δ13C (blue; temperature proxy index) (Xu et al. 2010), Indian Ocean coral δ18O (pink; temperature proxy index; see Cole et al. (2000) for details), and all Indian temperature records (1901–1990; sky blue) (Pant and Rupa Kumar 1997)

Full size image

The north or south movement of Intertropical Convergence Zone (ITCZ) has been widely suggested to be related to Asian summer monsoon intensity (Wang et al. 2005; Fleitmann et al. 2007). In general, higher solar insolation leads to northward movement of ITCZ, and intensified Asian summer monsoon on long term time scale, e.g. millennium timescale. However, for short term timescales, the movement of ITCZ is still not well documented. Recently, Sachs et al. (2009) suggested that the ITCZ zone shifted southwards significantly during 1400–1850 AD (the Little Ice Age), which implies a northward return after 1850 AD. Thus, if the ISM trend from 1860 to the present is also dominated by ITCZ, then an increasing, rather than decreasing ISM trend can be expected, which is contrary with the observed decline trend.

Numerous studies link the reduced ISM rainfall with changes in SST of the Indian Ocean (e.g. Giannini et al. 2003; Naidu, et al. 2009; Chung and Ramanathan 2006). Accompanied with the global warming, SSTs increase significantly. For example, various sources of SST data reveal that the tropical Indian Ocean, particularly the equatorial portion, has warmed substantially since the 1950s (see Chung and Ramanathan 2006 and references therein). Coral δ18O records indicate that the sea surface temperature increased both at Malindi and Seychelles (tropical West Indian Ocean) during the past two centuries (Cole et al. 2000). Wilson et al. (2006) showed that SST increased significantly between 30°N and 30°S in Pacific and Indian Oceans from about 1860 to the present. The Hongyuan precipitation (inferred from δ18Otree ring) is highly negatively correlated with the Indian Ocean SST (inferred from coral δ18O; r = −0.44, α < 0.001; Fig. 6; Cole et al. 2000), suggesting a possible linkage between increasing tropical ocean SST and decreasing ISM trend. Giannini et al. (2003) also linked the Sahel drought during the past two decades to the increased SST over the tropical oceans.

It is interesting to note that the SST increase over the equatorial Indian Ocean during the past 50 years is much higher than that over the northern Indian Ocean (Chung and Ramanathan 2006), which leads to an obvious decrease in summertime SST gradient. Ramanathan et al. (2005) ascribed this decreased SST gradient to the green house gases and the aerosol; the former leads to a uniform warming of the tropical Indian Ocean, while the latter has a cooling effect and masks the green house gas warming effect in the northern Indian Ocean. “From a dynamical consideration, a weakening of the SST gradients in the north Indian Ocean can weaken the monsoon circulation, by weakening both the cross-equatorial (south to north) surface pressure gradient and north-to-south mid- and upper troposphere pressure gradient” (Chung and Ramanathan 2006). Therefore, decreasing rather than increasing monsoon circulation can be expected. Naidu et al. (2009) also associated the decreased ISM rainfall with the decreased meridional SST gradient under a global warming era.

The decreased ISM is also possibly related to the decrease in land-sea thermal contrast under an abnormal global warming scenario. Although temperature over the northern hemisphere has been widely reported to increase during the past 100 and 200 years, the increment over land regions where the Indian low pressure develops may be smaller compared with that over sea surface. This is possibly because the increasing atmospheric brown cloud over the Indian sub-continent and northern Indian Ocean weakened the increase in temperature (Ramanathan et al. 2005). For example, SST at western Indian Ocean increased significantly during the past 100 years inferred from Cole et al. (2000), yet the increase in surface temperature over the Indian sub-continent is rather slight from 1900 to 2000 (Fig. 6). Temperature at Pune, western India, even shows a decreasing trend during the past one hundred years (Gadgil and Dhorde 2005). Except for the aerosol cooling effect, we speculate that the Himalayan snow cover, which has been strongly shrinked during the past decades (e.g. Zhao and Moore 2006), may have dampened the increase in temperature, especially under an abnormal global warming scenario. For example, the spring temperature variations reconstructed from tree ring chronologies over the western Himalaya during the past four centuries showed no clear long term trend, but even indicated “a long-term cooling trend since the late 17th century to early 20th” (Yadav and Singh 2002). The observed data also indicated no clear (or slightly decreasing) long term temperature trend over northwestern India and western Himalaya during the past century (Yadav et al. 2004; Ganguly and Iyer 2009). Meehl and Washington (1993) also showed that the SST of the Indian Ocean increased by 0.28 °C but the south Asian land temperature increased only 0.07 °C from 1971 to 1990. Therefore, a decreased long term land-sea thermal contrast between the equatorial Indian Ocean and Indian sub-continent to western Himalaya under an abnormal global warming scenario can be expected. This decreased land-sea thermal contrast may have led to the decreased summer monsoon rainfall over a wide geographic region during the past three decades, like in the Sahel zone (Giannini et al. 2003), India, Nepal, Bangladesh (Naidu et al. 2009; Chung and Ramanathan 2006), and the southwest to north China (Xiang et al. 2007).

The increased SST leads to more atmospheric moisture transported northward to the summer monsoon regions due to intensified evaporation, which is expected to provide more precipitable water over the ISM region. The decreased monsoon circulation may have counterbalanced the increased atmospheric moisture, and leads to no clear long term rainfall trend over the west and central Indian sub-continent. However, for northeastern India, and for regions along the monsoon flow direction (like areas along the above mentioned “vapor channel” to southwest China), the weakened monsoon flow will eventually lead to decline in monsoon rainfall since much vapor has been precipitated before reaching this area. Therefore, if the global warming continues, the SST over tropical Indian Ocean will persist increasing, while temperature over the north Indian Ocean, the Indian sub-continent, and western Himalaya would keep relatively steady, which will lead to continuous decreases in SST gradient over the Indian Ocean and decreases in land-sea thermal contrast, and eventually lead to continuously weakening of ISM intensity.

6 Summary

We reconstructed the ISM intensity during the past 270 years from tree ring δ18O at Hongyuan, eastern edge of the Tibet Plateau. The monsoon failures inferred from δ18Otree ring synchronize well with those recorded in other ISM indices. Significant periodicities of 22.6, 59.0, and 110.9-years were detected in the Hongyuan δ18Otree ring series, which may be the responses to solar activities. A notable feature of the reconstructed ISM intensity is the gradually decreasing trend from about 1860 to the present, which is inversely related to the increasing temperature trend contemporaneously. Such “decreasing ISM intensity–increasing temperature” tendency has been suggested to be related to the decrease in sea surface temperature gradient between tropical and north Indian Ocean, and the decrease in land-sea thermal contrast between tropical Indian Ocean and “Indian sub-continent–western Himalaya”. Under this circumstance, the ISM intensity would be expected to continue decreasing if the global warming continues.