Abstract

The onset of the South Asian summer monsoon (SASM) indicates the beginning of the rainy season in the South Asia region. It is not only critical for the local agriculture and animal husbandry but also important for water and life security. Precipitation in the early rainy season (May) increases rapidly and has a large interannual variability, especially in the Tibetan Plateau (TP) region. One of the starting mechanisms of the monsoon system is the land–sea thermal contrast (LSTC) between the Indian Ocean (IO) and South Asia region. Therefore, the IO can be considered as a crucial factor for the intensity of the monsoon system, as well as the TP precipitation. In this study, the relationships between IO sea surface temperature (SST) and TP precipitation on the interannual time scale are investigated. Correlation maps show that IO SST variability contains a portion that is independent from the tropical Pacific Ocean SST and is negatively correlated with the TP precipitation. Here the authors define an LSTC index to determine the thermal condition over the IO and South Asia region. The SASM reveals an out-of-phase relationship with LSTC between land and ocean, which means it would be suppressed by the enhanced LSTC. The daily data are used to further analyze the relationship between the SASM and TP precipitation in detail. Results show that the anomalous TP precipitation in May is mainly caused by the Bay of Bengal monsoon and that the Indian monsoon is responsible for the TP precipitation in June. More specifically, warmer SST enlarges the LSTC between the IO and South Asia region. The SASM is weaker than the mean state, resulting in less precipitation over the TP. In negative years the opposite occurs.

1. Introduction

The Tibetan Plateau (TP), with an average elevation of over 4000 m above sea level and occupying an area of around 2.5 million km2, is the highest highland in the world. The TP plays critical roles in dominating the regional and global circulation in Asia as well as the Northern Hemisphere through dynamical and thermal forcing (Bothe et al. 2012; Murakami 1987; Yanai et al. 1992). The TP is known as the “third pole” and the “Asian water tower” (X. Xu et al. 2008). There are 36 793 existing glaciers on the TP, with a total area of 49 873.44 km2 and ice volume of 4561.3857 km3, which are at the headwaters of many famous rivers in Asia (Yao et al. 2012). As an important linkage of water and hydrological cycle, the TP precipitation has been widely studied. You et al. (2015) compare multiple datasets with gridded precipitation observations over the TP. Annual precipitation, especially in the east and central TP, has increased during the past 40 years, while in the west TP it has exhibited a decreased trend (Z. Xu et al. 2008). Gao et al. (2014) investigated the moisture flux over the TP and found that the dynamic component plays an important role in the changes of precipitation minus evapotranspiration. The TP had become wetter as a whole during 1979–2011, but the spatial variability is large, which is mainly caused by the complex topography. Moreover, divided by the Tanggula Mountains, there is an antiphase of summer (June–August) precipitation variation between the northeast and southeast TP (Liu and Yin 2001), and a wave train induced by the North Atlantic Oscillation may be a major reason for this dipole oscillation of precipitation (Liu et al. 2015). The TP is considered as a box in some studies to calculate the water vapor balance. The South Asian summer monsoon (SASM) brings most of the water vapor into the TP through the southern border and dominates the summer precipitation over the southeast TP (Feng and Zhou 2012). Schiemann et al. (2009) described the seasonal cycle and the interannual variability of the westerly jet in the TP region. A northerly jet position in the TP region is related to less precipitation over the TP in April. Lin et al. (2016) found that the midlatitude westerly and SASM show opposite effects on the water vapor transport to the TP region. Some major features of the climate systems, such as the South Asian high, monsoon trough, and Somali jet also play important roles on the moisture supply (Findlater 1969; Krishnamurti and Bhalme 1976; Wei et al. 2014). In summary, the TP precipitation is mainly affected by three factors: topography, moisture supply, and the surrounding climate systems. However, there has been little focus on the TP precipitation in the early rainy season. The relationship between Indian Ocean (IO) SST, SASM, and TP precipitation also needs further investigation.

Under the global warming background, the IO has undergone dramatic changes in recent years. Since the 1960s, the sea level has increased in most of the IO except for the south tropical IO (Han et al. 2010). The SST in the equatorial IO displays warming trends after 1950s, with the total trend amounting to 0.5°C by the end of twentieth century (Du and Xie 2008). As the heat transport from the Pacific Ocean to the IO carried by the Indonesian Throughflow has increased, the IO heat content has increased abruptly during the past decade (Lee et al. 2015). Following the El Niño that occurred in the prior winter, the IO SST increases and persists through boreal summer. This basinwide warming phenomenon can be revealed by the first empirical orthogonal function (EOF) of IO SST and is referred to as the IO basin mode (IOBM) (Yang et al. 2007). Many studies indicate that the IOBM not only is a passive response to El Niño but also acts as a capacitor that anchors atmospheric anomalies over the Indo–western Pacific Ocean (Xie et al. 2009). The positive IOBM event can enhance the summer monsoon and has a strong link to the large-scale atmospheric circulation, which increases the moisture convergence and more precipitation over South Asia (Yang et al. 2010). Chiang and Lintner (2005) found that the tropics’ tropospheric temperature is responsible for observed SST and precipitation variability in the tropics outside the El Niño–Southern Oscillation (ENSO) region, which may be another reason for the formation of IOBM.

In general, the monsoons are considered an atmospheric response to seasonal changes in land–sea thermal contrast (LSTC). It is not only critical for agriculture in the environmentally sensitive region but also important for water and life security (X. Xu et al. 2008). One of the starting mechanisms of the monsoons is the thermal difference between land and ocean; the monsoon onset is concurrent with the reversal of meridional temperature gradient (Webster et al. 1998). The SASM, which is the strongest element of the global monsoon system (Wang 2006), affects the precipitation over South Asia in different ways. The increased aerosols, such as dust and black carbon, in north India during the early rainy season may lead to an advance of the SASM (Lau and Kim 2006; Lau et al. 2006). The early or late onset of SASM can directly control the pattern and amount of precipitation over South Asian regions (Xing et al. 2015). As a crucial factor of LSTC, the IO SST plays an important role on the generation and evolution of SASM (Li 1996). In terms of the tropical biennial oscillation, the monsoon rainfall is mainly affected by local moisture convergence that is associated with the IO SST variability (Li et al. 2001). Cherchi et al. (2007) found that the monsoon system acts as a bridge between IO SST and TP precipitation. A significant portion of SST variability in the equatorial IO can affect the precipitation over India. In this study, we investigate the relationship between IO SST and TP precipitation on the interannual time scale. Then the detailed process of SASM influencing the TP precipitation is further analyzed via daily data.

The paper is organized as follows. Section 2 gives the data and method used in this study. The relationship between IO SST and TP precipitation is described in section 3. Section 4 reveals the influence of IOBM-like SST on SASM. In section 5, the effects of SASM on TP precipitation are investigated. Finally, conclusions and discussion are presented in section 6.

2. Data and methods

We choose the stations with elevation over 3000 m above sea level and within the range of 25°–40°N, 75°–105°E to represent the TP precipitation (Fig. 1). The monthly and daily data from 59 stations are provided by the National Meteorological Information Center, China Meteorological Administration (NMIC/CMA). This dataset for the period 1979–2014 is selected. After carefully assessing the data quality, reliability, and homogeneity, the adjustments have improved the reliability of the data and decreased uncertainties in the study of observed climate change in China, and readers are referred to Li et al. (2012) for more detailed information. The SST data are obtained from the monthly mean Hadley Centre Global Sea Ice and SST (HadISST) dataset (Rayner et al. 2003). The atmospheric data are used from the European Centre for Medium‐Range Weather Forecasts (ECMWF) interim reanalysis (ERA-Interim) (Dee et al. 2011; Simmons et al. 2007). All the reanalysis data had been regridded to a 2.5° × 2.5° horizontal resolution. Observations and reanalysis data used in the study cover the period 1979–2014.

Fig. 1.

Spatial distribution of 59 surface meteorological stations (black dots) across the TP. The colors denote the altitude (m). Dark blue lines indicate the grid boxes.

Fig. 1.

Spatial distribution of 59 surface meteorological stations (black dots) across the TP. The colors denote the altitude (m). Dark blue lines indicate the grid boxes.

As shown in Fig. 1, most of the stations are located in the east part of TP. To better represent the precipitation over the entire TP, we calculate the precipitation series by the following means: first, the TP region is divided into 5° × 5° resolution grid boxes; second, all available precipitation data are averaged within each grid box; and, finally, the regional mean precipitation is calculated by averaging all grid boxes (grid boxes without stations are dismissed). A similar method can be found in a previous study (Abatzoglou and Barbero 2014). The SASM index is used to characterize the interannual variation of SASM. It is defined as the difference between zonal wind at 850 and 200 hPa over the domain 0°–20°N, 40°–110°E (Webster and Yang 1992). We also investigate the relationship between SASM and IO SST using a land–sea thermal contrast (LSTC) index at a range of 0°–40°N, 50°–100°E:

 
formula

where Tland is the vertical integrated temperature (VIT) over the land and Tocean is the VIT over the IO. The apparent heat source Q1 and its vertically integrated values 〈Q1〉 are calculated based on the following equations (Yanai 1973):

 
formula
 
formula

where T is the temperature; θ is the potential temperature; V is the horizontal wind vector; ω is the vertical p velocity; κ = R/cp; R and cp are the gas constant and the specific heat at constant pressures of dry air, respectively; p0 = 1000 hPa; and ps and pt are the pressures at the surface and at 200 hPa, respectively. The IOBM is represented as the first EOF modes of SST anomaly over the IO (40°S–30°N, 40°–110°E), a significant positive (negative) IOBM event is selected when the first principal component PC1 exceeds 1 (below −1). The averaged SST anomalies in November–January in the Niño-3 region (5°S–5°N, 90°–150°W) are used to represent the canonical ENSO event. The partial correlation is used to examine the relationships between IO SST and TP precipitation after removing the previous effects of ENSO. It involves computation of the linear dependence of a predictand upon a predictor after the linear relationship with a second predictor has been removed from both the predictand and predictor [detailed information about this method can be found in Cai et al. (2011) and Yuan et al. (2009)]. The statistical significance is tested by the two-tailed t test. All the precipitation and SST data are normalized and the linear trend has been removed before analysis.

3. Relationship between IO SST and TP precipitation

Figure 2 gives the seasonal cycle of the TP precipitation. The peak appears in July, with a maximum of 98.3 mm. Followed by August and June, the amounts of precipitation are 88.6 and 66.7 mm, respectively. The TP precipitation in May–October account for 90.3% of total precipitation, and thus it is defined as the rainy season. With a large amount of moisture brought into the TP, the precipitation increases rapidly from April to July. As the onset of SASM occurs around May (Xing et al. 2015), the standard deviation of precipitation in May becomes much larger compared to the prior months. The coefficient of variation (the standard deviation divided by the mean) is 0.21, which is the second largest in the whole rainy season except for that in October. This means that despite the average precipitation in May (32.8 mm) being less than that in June–-September, the interannual variability caused by the SASM greatly influences the local agricultural economics. In this study, we focus on the TP precipitation in the early rainy season and detect the potential influence from the IO.

Fig. 2.

Seasonal cycle in TP precipitation presented as monthly average (1979–2014; mm). The error bars represent one standard deviation from the mean.

Fig. 2.

Seasonal cycle in TP precipitation presented as monthly average (1979–2014; mm). The error bars represent one standard deviation from the mean.

We calculated the correlation coefficients to find out the crucial region in the IO. Figure 3a gives the correlation map between TP precipitation and the simultaneous SST in May. There is a board association between SSTs in the equatorial and northwestern IO and TP precipitation, as highlighted by significant negative correlations exceeding −0.5. Given the IO SST has a positive response to the El Niño event that occurred in the prior winter, its variation may contain a portion of the tropical Pacific Ocean SST variation. Thus, we further use partial correlation analysis to extract a sole influence of the IO SST. As shown in Fig. 3b, after removing the El Niño effect, the correlation coefficient becomes slightly smaller in the western and tropical IO, but the basic distribution exhibited in Fig. 3a still remains. As both the two correlation maps show great consistency of SST variation, we assume that the IOBM may be a key factor responsible for it. The EOF analysis has been applied to the IO SST in May. The IOBM accounts for 33.1% of the total variance. Positive characteristic values exist in most of the IO basin, which indicates the IO has a concurrent variation (Fig. 3c). This EOF1 mode is similar to the correlation map. Meanwhile, the correlation coefficient between the first principal component (PC1) and TP precipitation in May is −0.42 (99% confidence level), and the correlation coefficient turns to −0.33 after removing the Niño-3 index (95% confidence level). So the IOBM can be considered as a key factor when studying the effect of IO SST on TP precipitation. Thus, we select five positive IOBM years (1983, 1987, 1991, 1998, and 2010) and seven negative IOBM years (1984, 1985, 1989, 1999, 2000, 2008, and 2011) where the PC1 exceeds 1 (below −1) for the composite analyses (Fig. 3d).

Fig. 3.

(a) Correlation patterns of normalized and detrended precipitation in May with IO SST. (b) Partial correlation patterns of normalized and detrended precipitation in May with IO SST after removing Niño-3 index. Black dots indicate areas of 95% confidence level. (c) The EOF1 pattern of normalized and detrended IO SST in May. (d) Time series of the TP precipitation in May (P_M) and the PC1 of TP precipitation.

Fig. 3.

(a) Correlation patterns of normalized and detrended precipitation in May with IO SST. (b) Partial correlation patterns of normalized and detrended precipitation in May with IO SST after removing Niño-3 index. Black dots indicate areas of 95% confidence level. (c) The EOF1 pattern of normalized and detrended IO SST in May. (d) Time series of the TP precipitation in May (P_M) and the PC1 of TP precipitation.

To verify the negative correlation between PC1 and precipitation in May, we extracted the TP precipitation difference between the positive and negative IOBM years. The seasonal cycle (Fig. 4a) shows that precipitation in positive IOBM years is less than that in negative IOBM years in May–June. The precipitation differences in these two months are both statistically significant at the 95% confidence level, indicating that the IO SST in May can influence the TP precipitation not only in May but also in June. Figure 4b is the spatial difference of TP precipitation in May–June. Except for a small part of the northeast TP, most of the regions exhibit negative precipitation anomalies. There are three significant negative precipitation centers with the values below −20 mm: two locations in the central TP and the other in the southeast TP.

Fig. 4.

(a) Seasonal cycle of monthly TP precipitation (mm) in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean. The bars are the precipitation difference between positive and negative IOBM years. Shading indicates the differences are statistically significant at the 95% confidence level. (b) Spatial difference of TP precipitation in May–June between positive and negative years (mm). Shading indicates areas of 95% confidence level.

Fig. 4.

(a) Seasonal cycle of monthly TP precipitation (mm) in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean. The bars are the precipitation difference between positive and negative IOBM years. Shading indicates the differences are statistically significant at the 95% confidence level. (b) Spatial difference of TP precipitation in May–June between positive and negative years (mm). Shading indicates areas of 95% confidence level.

To investigate the atmospheric response to the SST anomaly, the difference in composite anomalies of winds and geopotential height at 200 hPa (Fig. 5a) are presented. The anomalous westerlies, with positive geopotential anomalies, cover most of South Asia and the tropical IO. There is an anomalous cyclonic flow on the west of the TP. It reveals that the South Asian high moves to a farther southward (northward) position in positive (negative) IOBM years. This is because the South Asian high is the dominant climate system at the upper level (100 hPa) over the TP in summer (Liu et al. 2013). With the onset of SASM, it moves up the TP from the Indochina peninsula. We note here that the conclusions are similar if we check the difference pattern at 100 hPa. At 850 hPa, the most significant signal is the anomalous easterlies from Indonesia to East Africa along 10°N. Meanwhile the anomalous easterlies shifted toward the southeast over the western IO. There is an anomalous anticyclonic flow on the north of the Bay of Bengal. Both the anomalous winds at upper and lower levels are oriented against the climatology flow, which means the positive (negative) IOBM event prevents (facilitates) the onset of the monsoon system in the early rainy season. As the SASM exhibits significant anomalies between the positive and negative IOBM years, we assume that SASM is the key factor responsible for the influence of IO SST on TP precipitation. Here two questions are presented: Why does positive IOBM induce weaker SASM? How does the SASM affect the TP precipitation?

Fig. 5.

Composite difference of u, υ wind (m s−1; vector) and geopotential height (m; contour) in May–June at (a) 200 and (b) 850 hPa between positive and negative IOBM years. Black vectors and the shading indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP.

Fig. 5.

Composite difference of u, υ wind (m s−1; vector) and geopotential height (m; contour) in May–June at (a) 200 and (b) 850 hPa between positive and negative IOBM years. Black vectors and the shading indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP.

4. IOBM affecting SASM and the linkage with LSTC

The SASM is an atmospheric response to seasonal changes in LSTC. It presents as reverse wind direction between the upper and lower troposphere. Figure 6a gives the climatological VIT over South Asia and IO in May. The VIT is about 2.5 × 106 K kg m−2 over the entire IO and with a decreasing gradient from south to north. The VIT over the TP is much smaller than the other region, it is about 1.5 × 106 K kg m−2. It reveals that in the early rainy season, the atmosphere over the ocean is basically warmer than that over the land. Then we calculate the correlation coefficient between the PC1 and the VIT (Fig. 6b, color). Positive correlations exceeding 0.9 can be found over most of the IO, whereas it decreases to zero over the land. As the PC1, corresponding to the IOBM, can represent the entire IO SST variations, it indicates that Tocean (VIT over the IO) has a strong positive response to the IO SST. Warmer IO SST corresponds to higher Tocean. The contours in Fig. 6b indicate the correlation coefficient between the SASM index and the VIT. It reveals an out-of-phase relationship between land and ocean. The SASM is positively correlated to Tland (VIT over the land), whereas it is negatively correlated to Tocean. To better understand the relationship between IO SST, VIT, and SASM, an LSTC index defined by the difference between Tocean and Tland is calculated. Figure 7 reveals the relationship between SASM and the LSTC. The black line is the climatological mean of SASM index. The red and blue lines represent the SASM index in positive and negative IOBM years, respectively. The correlation coefficient between SASM index and PC1 is −0.60 (99% confidence level). In the rainy season (May–October), a larger value means a stronger SASM. The bars are the difference of LSTC index between the positive and negative IOBM years. The SASM (LSTC) index is dotted (shaded) if the difference between positive and negative IOBM years pass the 95% confidence level. It is notable that the SASM and LSTC indices are only applicable during May–October. The correlation coefficient between SASM index and LSTC index in May is 0.61 (99% confidence level). In positive IOBM years, IO SST is warmer and the Tocean is higher. As the climatological Tocean is higher than Tland, thus the positive SST anomalies may enlarge the LSTC. Meanwhile, the SASM is negatively correlated to Tocean. In other words, the SASM would be suppressed by the larger LSTC. As shown in Fig. 7, the LSTC anomaly is significant in May and the SASM anomalies are significant in May and June. Larger (smaller) LSTC is corresponding to weaker (stronger) SASM in positive (negative) IOBM years. It is in accordance with the composite of precipitation (Fig. 4a).

Fig. 6.

(a) Climatological VIT in May from 1979 to 2014 (106 K kg m−2). (b) Correlation patterns of PC1 and VIT (color) and correlation patterns of SASM index and VIT (contour).

Fig. 6.

(a) Climatological VIT in May from 1979 to 2014 (106 K kg m−2). (b) Correlation patterns of PC1 and VIT (color) and correlation patterns of SASM index and VIT (contour).

Fig. 7.

Seasonal cycle of SASM index in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean (left axis; m s−1). The bars indicate the composite difference of LSTC index (right axis; 103 K kg m−2) between positive and negative IOBM years. Dots and shading indicate the difference is statistically significant at the 95% confidence level.

Fig. 7.

Seasonal cycle of SASM index in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean (left axis; m s−1). The bars indicate the composite difference of LSTC index (right axis; 103 K kg m−2) between positive and negative IOBM years. Dots and shading indicate the difference is statistically significant at the 95% confidence level.

It is notable that the IOBM is closely related to the Tocean, whereas it exhibited little correlation with Tland in Fig. 6b. As the LSTC index we defined in this study consists of Tocean and Tland, the variation of Tland is also worth investigating to clarify the thermal condition between land and ocean. In positive IOBM years, the amount of TP precipitation is less than normal, which may suppress the diabatic heating over the TP region. The mid- to upper-troposphere thermal condition is shown in Fig. 8; it is defined as the vertical averaged air temperature between 200 and 600 hPa (Goswami and Ajaya Mohan 2001). Figure 8a gives the climatological mid- to upper-troposphere temperature over the South Asia region in May. It is basically warm over the Indian Ocean. A warm center with value above −34°C is observed over the Bay of Bengal. The troposphere temperature gradient, which is one important feature associated with the onset and withdrawal of the monsoon system, can be found over the land. The diabatic heating released by the monsoon precipitation is the most important parameter for the warm center. Thus, we select seven positive TP precipitation years (1980, 1989, 1999, 2000, 2002, 2004, and 2007) and five negative TP precipitation years (1979, 1995, 2005, 2012, and 2014) where the precipitation series in May exceeds 1 (below −1). We use negative precipitation years minus positive precipitation years for the composite analysis to be in accordance with the previous analysis. Composite difference of mid- to upper-troposphere temperature (20°–40°N, 40°–100°E) in May is constructed to see how the anomalous precipitation affects the thermal condition over the land. The longitude–time cross section (Fig. 8b) shows that the western part is in cold phase whereas the eastern part is in warm phase at the beginning of May. Then the region in cold phase expand over time and form a cold center with center value below −3°C, which is statistically significant at 95% confidence level, indicating that the less than normal precipitation over the TP can induce the cooling to its west. The cooling in turn further enlarges the LSTC, which would further suppress the SASM.

Fig. 8.

(a) Climatological mid- to upper-troposphere temperature in May (°C). Black dashed line denotes the area boundary of TP. (b) Longitude–time cross section (20°–40°N) of composite difference of mid- to upper-troposphere temperature in May (°C). Black dots indicate areas of 95% confidence level.

Fig. 8.

(a) Climatological mid- to upper-troposphere temperature in May (°C). Black dashed line denotes the area boundary of TP. (b) Longitude–time cross section (20°–40°N) of composite difference of mid- to upper-troposphere temperature in May (°C). Black dots indicate areas of 95% confidence level.

The composite difference of atmospheric circulation and SST in May is also calculated. As the IOBM index is negatively correlated with the TP precipitation in May. Both the composite differences in SST and atmospheric circulation selected from the IOBM index (Fig. 5) are similar to those selected from the TP precipitation series (not shown). At the upper level, the positive geopotential height anomaly induced by the IOBM is over 40 m, which is higher than that calculated based on the TP precipitation series (20 m), whereas in Fig. 5a the negative geopotential height center (with center value below −40 m) west of the TP is not as strong as that induced by the TP precipitation (with center value below −70 m). At lower levels, the anomalous easterlies induced by the IOBM (Fig. 5b) are stronger than that calculated based on the TP precipitation series. This means that the most significant effect of IOBM is over the tropical region; the TP precipitation is affected through the SASM. On the other hand, over the South Asia region, the anomalous TP precipitation may in turn give a positive feedback.

5. Influence of SASM on TP precipitation

Figure 9a gives climatological 〈Q1〉 and the u, υ wind at 925 hPa in May. It reveals the circulation pattern at lower levels in the early rainy season. The onset of SASM brings a huge amount of moisture to the South Asia region. The 〈Q1〉 center covers from the south Indian peninsula to the north Indochina peninsula with center value exceeding 300 W m−2. The 〈Q1〉 includes the total effect of the sensible heating, latent heating, and radiation. Wei et al. (2014) found that the heating anomalies over the South Asia region are mainly condensation latent heating induced by convection. Thus, 〈Q1〉 can basically represent the precipitation pattern, which means the convections are active over these regions. Figure 9b gives the composite difference of 〈Q1〉 between positive and negative IOBM years; significant negative anomalies can be found from the Bay of Bengal to the southeastern part of the TP. The SASM is weak during positive IOBM event, and the 〈Q1〉 center is suppressed in the South Asia region. Thus, the anomalous anticyclonic flow can be found because of the weakened convective activities.

Fig. 9.

(a) Climatological 〈Q1〉 (color; W m−2) and u, υ winds (vector; m s−1) at 925 hPa in May. (b) Composite difference of 〈Q1〉 in May–June. Black dots indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP.

Fig. 9.

(a) Climatological 〈Q1〉 (color; W m−2) and u, υ winds (vector; m s−1) at 925 hPa in May. (b) Composite difference of 〈Q1〉 in May–June. Black dots indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP.

We further calculate the vertical integrated water vapor flux (from surface to 500 hPa) and its divergence in Fig. 10a. Except for the TP region and the north Indochina peninsula, the equivalent height of surface is below 500 m. There is an anticyclone anomaly over the northern Bay of Bengal and the anomalous easterlies can be found around 10°N. The anomalous moisture transport is similar to the wind difference at 850 hPa. The anomalous divergence center is associated with the anticyclonic pattern, which means the convergence center is suppressed by the weakened SASM, leading to less moisture transport from the Arabian Sea and Bay of Bengal to the TP region. The anticyclonic flow and the divergence center are significant at the 95% confidence level. Figure 10b exhibits the averaged meridional overturning circulation with the range of 85°–105°E. There is a significant updraft over the tropical IO at lower levels. An anomalous downdraft is significant over the Bay of Bengal, as well as the TP region. The anomalous descending flow corresponds to the anomalous divergence center and the negative 〈Q1〉 anomaly, which indicates the convection activities are weak over these regions. The significant downdraft over the TP is around 30°N, which is in accordance with the composite difference in precipitation (Fig. 4b).

Fig. 10.

(a) Composite difference of vertically integrated water vapor flux [vector; kg (m s)−1] and its divergence [colors; 105 kg (m s)−1] in May–June. The black vectors and the dotted lines indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP. (b) Composite difference of meridional cross section of υ (streamlines; m s−1) and vertical velocity (streamlines; 100 pa s−1) in May–June. Shading indicates areas of 95% confidence level. The black mass indicates the topography.

Fig. 10.

(a) Composite difference of vertically integrated water vapor flux [vector; kg (m s)−1] and its divergence [colors; 105 kg (m s)−1] in May–June. The black vectors and the dotted lines indicate areas of 95% confidence level. Black dashed line denotes the area boundary of TP. (b) Composite difference of meridional cross section of υ (streamlines; m s−1) and vertical velocity (streamlines; 100 pa s−1) in May–June. Shading indicates areas of 95% confidence level. The black mass indicates the topography.

The SASM is a large system that covers most of the South Asia region in summer and includes several small monsoon systems (i.e., the Bay of Bengal monsoon and the Indian monsoon). Figure 4a shows that the anomalous SST can influence the TP precipitation not only in May but also in June. To analyze the detailed process of the monsoon system in May and June, we further check the daily precipitation evaluation as well as the OLR evaluation to investigate the relationship of the Bay of Bengal monsoon and the Indian monsoon with the TP precipitation. The 5-day running mean of climatological precipitation shows an increasing trend in May and a more rapid growth in June (Fig. 11a). In the first two pentads of May, the precipitation in positive and negative IOBM years show no significant difference. From the third pentad, the amount of precipitation in negative IOBM years becomes much larger than that in positive IOBM years. The difference of precipitation lasts for about 25 days and ends up at the first pentad of June. Figure 11b gives the monsoon indices represented by the regionally averaged OLR. The Bay of Bengal monsoon is defined as the averaged OLR over 10°–30°N, 90°–105°E. In the tropics, the OLR flux less than 240 W m−2 indicates the presence of deep cumulus convection (Li 1996). In other words, the monsoon onset and its strength can be represented by the OLR value. The red and blue lines are the Bay of Bengal monsoon indices in positive and negative IOBM years, respectively. In positive IOBM years, the monsoon index is around 240 W m−2 and starts to decrease until the third pentad of May. In negative IOBM years, the monsoon index is below 240 W m−2 since the first pentad of May. Meanwhile, the blue line is less than the red line during May and the first pentad of June. This reveals that the Bay of Bengal monsoon onset in negative IOBM years is earlier than that in positive IOBM years. The differences of Bay of Bengal monsoon indices are in good accordance with the difference of TP precipitation in May. We conclude that the anomalous TP precipitation in May is mainly caused by the Bay of Bengal monsoon. Another significant precipitation difference can be found from the second pentad of June to the fifth pentad of June. The amount of TP precipitation in negative IOBM years is larger than that in positive IOBM years. The difference of Indian monsoon indices (defined as the averaged OLR over 10°–30°N, 70°–90°E) become significant from the fourth pentad of May. Based on the criterion of 240 W m−2, the onset of Indian monsoon in negative IOBM years occurs at the first pentad of June, which is one pentad earlier than that in positive IOBM years. The yellow line is lower than the green line in the first four pentads of June. As the Bay of Bengal monsoon indices show no difference in June, the anomalous TP precipitation in June is due to the strength of Indian monsoon. We also calculate the relationship between the IOBM SST and the two monsoon systems. The correlation coefficient between PC1 and the Bay of Bengal monsoon index is −0.46 (99% confidence level), and the correlation coefficient between PC1 and the Indian monsoon index is −0.34 (95% confidence level). As the PC1 is also negatively correlated with the SASM, the strength of the Bay of Bengal monsoon and Indian monsoon are in accordance with the strength of SASM.

Fig. 11.

(a) Daily precipitation evaluation in May and June in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean (mm). (b) Daily OLR evaluation in May and June over the Bay of Bengal (BOB) in positive (red dashed line) and negative (blue dotted line) IOBM years and daily OLR evaluation in May and June over the Indian peninsula (IND) in positive (green dashed line) and negative (yellow dotted line) IOBM years (W m−2). OLR values less than 240 W m−2 (black line) represent deep cumulus convection.

Fig. 11.

(a) Daily precipitation evaluation in May and June in positive (red dashed line) and negative (blue dotted line) IOBM years. The black solid line indicates the climatological mean (mm). (b) Daily OLR evaluation in May and June over the Bay of Bengal (BOB) in positive (red dashed line) and negative (blue dotted line) IOBM years and daily OLR evaluation in May and June over the Indian peninsula (IND) in positive (green dashed line) and negative (yellow dotted line) IOBM years (W m−2). OLR values less than 240 W m−2 (black line) represent deep cumulus convection.

6. Conclusions and discussion

Tropical IO SST anomalies are a remote influence on precipitation in the monsoon region as well as on the moisture supply to the TP. In this study, effects of IO SST on TP precipitation in the early rainy season are discussed, which was proposed based on the onset of SASM. It acts as a bridge between the IO SST and the TP precipitation. Meanwhile, the characteristic of the climatological TP precipitation in the early rainy season is also worth investigating. The amount of precipitation in May increases rapidly and exhibits large interannual variation. In the early rainy season, the soil moisture variability is consistent with precipitation over the TP region, and it is likely to have persistence of several months. Therefore, the precipitation in the following months can also be influenced through anomalous surface energy and moisture fluxes, which is termed as soil moisture memory (Rai et al. 2015). The correlation map between precipitation in May and the simultaneous IO SST exhibits a general negative distribution over the equatorial and northwestern IO. Zheng et al. (2011) investigated the development of IOBM under global warming. Results showed that both the intensity and persistence of IOBM would be enhanced in the simulation forced by increased greenhouse gas concentrations. In our case, IOBM is demonstrated to be responsible for the TP precipitation in the early rainy season; how this mechanism changes needs further investigation.

Composites of wind anomalies show that the SASM is the most significant signal between positive and negative years. It acts as a bridge between IO SST and TP precipitation. The positive/negative TP precipitation is highly associated with the preonset/postonset of the monsoon system. The significant IOBM events dominate the thermal condition over the ocean, which induce SST anomaly over the tropical IO. The strength of SASM is linked not only to the surface temperature but also to the air temperature over the mid- to upper troposphere (Rai et al. 2015). The SASM is negatively correlated to Tocean while positively correlated to Tland, so it would be suppressed as a result of the enhanced LSTC. The LSTC anomaly is significant in May, and the SASM shows significant difference in May–June. We also test another SASM index, which is defined as the averaged OLR over 10°–25°N, 70°–120°E (Li 1996). It shows similar results compared with Webster and Yang’s (1992) index. The correlation coefficient between OLR index and PC1 is −0.42. Both these two indices can represent the strength of SASM.

The mechanism of how SASM affects the TP precipitation is also revealed. A huge amount of moisture was brought to the South Asia region in the early rainy season. A 〈Q1〉 center was formed over the South Asia region. The postonset of SASM in the positive IOBM years associated with less precipitation over India corresponds to less condensation heat release, which would cause anomalous cooling in the upper troposphere over the northern Indian peninsula (Wei et al. 2014). The 〈Q1〉, mainly composed by the latent heating, shows a significant negative anomaly over the northern Indochina peninsula. An easterly over the Bay of Bengal and the Arabian Sea is oriented against the climatological flow, which induces an anomalous anticyclone over the southeast TP at 850 hPa. There is an anomalous cyclone at 200 hPa. The South Asian high, which plays an important role in the effect of SASM on the rainfall over China, moves to a farther south position. The combined effect between upper and lower levels forces a meridional overturning flow from the equator to the TP. The downdraft branch weakens the convection, reduces water vapor flux input, causes anomalous water vapor divergence, and, finally, results in less precipitation over the TP. Figure 12 gives a schematic of the processes mentioned above. The detailed processes of the monsoon system affecting the TP precipitation are also investigated. The anomalous precipitation in May is mainly affected by the Bay of Bengal monsoon, while the Indian monsoon is responsible for the anomalous precipitation in June.

Fig. 12.

Schematic diagram showing the effect of IO SST on the TP precipitation in the early rainy season. Red shading in the IO indicates the positive IOBM event. The double-headed arrow represents the enhanced LSTC. Green arrows represent the wind anomalies at upper and lower levels. Blue shading indicates the negative 〈Q1〉 anomaly. The anomalous downdraft (yellow arrow), anticyclonic flow (red arrow), and divergent flow (purple arrows) are also presented.

Fig. 12.

Schematic diagram showing the effect of IO SST on the TP precipitation in the early rainy season. Red shading in the IO indicates the positive IOBM event. The double-headed arrow represents the enhanced LSTC. Green arrows represent the wind anomalies at upper and lower levels. Blue shading indicates the negative 〈Q1〉 anomaly. The anomalous downdraft (yellow arrow), anticyclonic flow (red arrow), and divergent flow (purple arrows) are also presented.

In this study, we use an LSTC index to represent the thermal condition between land and ocean. We test several variables, such as 2-m temperature and surface temperature. Results show that the VIT is an appropriate variable for the LSTC index. However, this index still can be improved. Wu; Liu (2016) made a comprehensive summary about the role of TP in Asian climate; the thermal control of the Asian summer monsoon was expanded in three aspects, which are the LSTC, mechanical insulation of TP, and thermal forcing of TP, respectively. The southern part of the SASM is controlled mainly by LSTC, its northern part by the thermal forcing of the Iranian Plateau and the East Asian monsoon, and its eastern part by the thermal forcing of the Tibetan Plateau (Wu et al. 2012). Xu et al. (2009) found that both the tropical zonal land–sea distribution and Asian mountains play a crucial role for establishing summer monsoon convection over the South Asian region. Mao et al. (2004) found that the three-dimensional structure of subtropical anticyclone can be well described by the westerly–easterly boundary surface. Therefore, the mean temperature gradient can be used to define the onset of the Asian summer monsoon. In our study, Tland and Tocean are divided by the coastlines. We are trying some new ways to complete its definition, and some other variables may be included in the further study. As shown in Fig. 6b, the SASM reveals an out-of-phase relationship between land and ocean. It is negatively correlated to Tocean whereas it is positively correlated to Tland. The onset of SASM is due to the combined effect of TP and IO. Here we only investigate the relationships between IO SST and Tocean; more work about the interaction between Tland and Tocean is needed to better understand the mechanisms of the TP precipitation. On the other hand, the anomalous TP precipitation can in turn modulate the SASM. As shown in Fig. 8b, the cooling induced by the precipitation can enlarge the LSTC, which would further suppress the strength of the SASM. There may exist a positive feedback between the SASM and TP precipitation. The interaction between TP precipitation, SASM, and SST still needs further study as it may change over time. Moreover, previous study found that if the precipitation in India were less than normal in the early rainy season, the precipitation in the following months would be abnormal (Rai et al. 2015). It is worth investigating whether it is the same in the TP region.

Acknowledgments

This study is supported by the National Key R&D Program of China (2017YFA0603804 and 2016YFA0601702), National Natural Science Foundation (41771069), Jiangsu Natural Science Funds for Distinguished Young Scholar (BK20140047), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and Jiangsu Shuang-Chuang Individual and Team Award. We are very grateful to the reviewers for their constructive comments and thoughtful suggestions.

REFERENCES

REFERENCES
Abatzoglou
,
J. T.
, and
R.
Barbero
,
2014
:
Observed and projected changes in absolute temperature records across the contiguous United States
.
Geophys. Res. Lett.
,
41
,
6501
6508
, doi:.
Bothe
,
O.
,
K.
Fraedrich
, and
X.
Zhu
,
2012
:
Tibetan Plateau summer precipitation: Covariability with circulation indices
.
Theor. Appl. Climatol.
,
108
,
293
300
, doi:.
Cai
,
W.
,
P.
Van Rensch
,
T.
Cowan
, and
H. H.
Hendon
,
2011
:
Teleconnection pathways of ENSO and the IOD and the mechanisms for impacts on Australian rainfall
.
J. Climate
,
24
,
3910
3923
, doi:.
Cherchi
,
A.
,
S.
Gualdi
,
S.
Behera
,
J. J.
Luo
,
S.
Masson
,
T.
Yamagata
, and
A.
Navarra
,
2007
:
The influence of tropical Indian Ocean SST on the Indian summer monsoon
.
J. Climate
,
20
,
3083
3105
, doi:.
Chiang
,
J. C.
, and
B. R.
Lintner
,
2005
:
Mechanisms of remote tropical surface warming during El Niño
.
J. Climate
,
18
,
4130
4149
, doi:.
Dee
,
D.
, and Coauthors
,
2011
:
The ERA‐Interim reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
, doi:.
Du
,
Y.
, and
S. P.
Xie
,
2008
:
Role of atmospheric adjustments in the tropical Indian Ocean warming during the 20th century in climate models
.
Geophys. Res. Lett.
,
35
,
193
202
, doi:.
Feng
,
L.
, and
T.
Zhou
,
2012
:
Water vapor transport for summer precipitation over the Tibetan Plateau: Multidata set analysis
.
J. Geophys. Res.
,
117
,
D20114
, doi:.
Findlater
,
J.
,
1969
:
A major low‐level air current near the Indian Ocean during the northern summer
.
Quart. J. Roy. Meteor. Soc.
,
95
,
362
380
, doi:.
Gao
,
Y.
,
L.
Cuo
, and
Y.
Zhang
,
2014
:
Changes in moisture flux over the Tibetan Plateau during 1979–2011 and possible mechanisms
.
J. Climate
,
27
,
1876
1893
, doi:.
Goswami
,
B. N.
, and
R. S.
Ajaya Mohan
,
2001
:
Intraseasonal oscillations and interannual variability of the Indian summer monsoon
.
J. Climate
,
14
,
1180
1198
, doi:.
Han
,
W.
, and Coauthors
,
2010
:
Patterns of Indian Ocean sea-level change in a warming climate
.
Nat. Geosci.
,
3
,
546
550
, doi:.
Krishnamurti
,
T. N.
, and
H. N.
Bhalme
,
1976
:
Oscillations of a monsoon system. Part I. Observational aspects
.
J. Atmos. Sci.
,
33
,
1937
1954
, doi:.
Lau
,
K. M.
, and
K. M.
Kim
,
2006
:
Observational relationships between aerosol and Asian monsoon rainfall, and circulation
.
Geophys. Res. Lett.
,
33
,
L21810
, doi:.
Lau
,
K. M.
,
M. K.
Kim
, and
K. M.
Kim
,
2006
:
Asian summer monsoon anomalies induced by aerosol direct forcing: The role of the Tibetan Plateau
.
Climate Dyn.
,
26
,
855
864
, doi:.
Lee
,
S.-K.
,
W.
Park
,
M. O.
Baringer
,
A. L.
Gordon
,
B.
Huber
, and
Y.
Liu
,
2015
:
Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus
.
Nat. Geosci.
,
8
,
445
449
, doi:.
Li
,
C.
,
1996
:
The onset and interannual variability of the Asian summer monsoon in relation to land–sea thermal contrast
.
J. Climate
,
9
,
358
375
, doi:.
Li
,
Q.
,
J.
Peng
, and
Y.
Shen
,
2012
:
Development of China homogenized monthly precipitation dataset during 1900–2009
.
J. Geogr. Sci.
,
22
,
579
593
, doi:.
Li
,
T.
,
Y.
Zhang
,
C. P.
Chang
, and
B.
Wang
,
2001
:
On the relationship between Indian Ocean sea surface temperature and Asian summer monsoon
.
Geophys. Res. Lett.
,
28
,
2843
2846
, doi:.
Lin
,
H.
,
Q.
You
,
Y.
Zhang
,
Y.
Jiao
, and
K.
Fraedrich
,
2016
:
Impact of large‐scale circulation on the water vapour balance of the Tibetan Plateau in summer
.
Int. J. Climatol.
,
36
,
4213
4221
, doi:.
Liu
,
B.
,
G.
Wu
,
J.
Mao
, and
J.
He
,
2013
:
Genesis of the South Asian high and its impact on the Asian summer monsoon onset
.
J. Climate
,
26
,
2976
2991
, doi:.
Liu
,
H.
,
K.
Duan
,
M.
Li
,
P.
Shi
,
J.
Yang
,
X.
Zhang
, and
J.
Sun
,
2015
:
Impact of the North Atlantic Oscillation on the dipole oscillation of summer precipitation over the central and eastern Tibetan Plateau
.
Int. J. Climatol.
,
35
,
4539
4546
, doi:.
Liu
,
X.
, and
Z.-Y.
Yin
,
2001
:
Spatial and temporal variation of summer precipitation over the eastern Tibetan Plateau and the North Atlantic Oscillation
.
J. Climate
,
14
,
2896
2909
, doi:.
Mao
,
J.
,
J. C. L.
Chan
, and
G.
Wu
,
2004
:
Relationship between the onset of the South China Sea summer monsoon and the structure of the Asian subtropical anticyclone
.
J. Meteor. Soc. Japan
,
82
,
845
859
, doi:.
Murakami
,
T.
,
1987
: Effects of the Tibetan Plateau. Monsoon Meteorology, C. Chang and T. Krishnamurti, Eds., Oxford University Press, 235–270.
Rai
,
A.
,
S. K.
Saha
,
S.
Pokhrel
,
K.
Sujith
, and
S.
Halder
,
2015
:
Influence of preonset land atmospheric conditions on the Indian summer monsoon rainfall variability
.
J. Geophys. Res. Atmos.
,
120
,
4551
4563
, doi:.
Rayner
,
N.
, and Coauthors
,
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:.
Schiemann
,
R.
,
D.
Lüthi
, and
C.
Schär
,
2009
:
Seasonality and interannual variability of the westerly jet in the Tibetan Plateau region
.
J. Climate
,
22
,
2940
2957
, doi:.
Simmons
,
A.
,
S.
Uppala
,
D.
Dee
, and
S.
Kobayashi
,
2007
: ERA-Interim: New ECMWF reanalysis products from 1989 onwards. ECMWF Newsletter, No. 110, ECMWF, Reading, United Kingdom, 25–35.
Wang
,
B.
,
2006
: The Asian Monsoon. Springer, 845 pp.
Webster
,
P. J.
, and
S.
Yang
,
1992
:
Monsoon and ENSO: Selectively interactive systems
.
Quart. J. Roy. Meteor. Soc.
,
118
,
877
926
, doi:.
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
,
14 451
14 510
, doi:.
Wei
,
W.
,
R.
Zhang
,
M.
Wen
,
X.
Rong
, and
T.
Li
,
2014
:
Impact of Indian summer monsoon on the South Asian high and its influence on summer rainfall over China
.
Climate Dyn.
,
43
,
1257
1269
, doi:.
Wu
,
G.
, and
Y.
Liu
,
2016
:
Impacts of the Tibetan Plateau on Asian climate
.
Multiscale Convection-Coupled Systems in the Tropics: A Tribute to Dr. Michio Yanai, Meteor. Monogr
., No. 56, Amer. Meteor. Soc., 7.1–7.29, doi:.
Wu
,
G.
,
Y.
Liu
,
B.
He
,
Q.
Bao
,
A.
Duan
, and
F.
Jin
,
2012
:
Thermal controls on the Asian summer monsoon
.
Sci. Rep.
,
2
,
404
, doi:.
Xie
,
S. P.
,
J.
Hafner
,
H.
Tokinaga
,
Y.
Du
,
T.
Sampe
,
K. M.
Hu
, and
G.
Huang
,
2009
:
Indian Ocean capacitor effect on Indo–western Pacific climate during the summer following El Niño
.
J. Climate
,
22
,
730
747
, doi:.
Xing
,
N.
,
J.
Li
, and
L.
Wang
,
2015
:
Effect of the early and late onset of summer monsoon over the Bay of Bengal on Asian precipitation in May
.
Climate Dyn.
,
4
,
1961
1970
, doi:.
Xu
,
X.
,
C.
Lu
,
X.
Shi
, and
S.
Gao
,
2008
:
World water tower: An atmospheric perspective
.
Geophys. Res. Lett.
,
35
,
525
530
, doi:.
Xu
,
Z.
,
T.
Gong
, and
J.
Li
,
2008
:
Decadal trend of climate in the Tibetan Plateau—regional temperature and precipitation
.
Hydrol. Processes
,
22
,
3056
3065
, doi:.
Xu
,
Z.
,
C.
Fu
, and
Y.
Qian
,
2009
:
Relative roles of land–sea distribution and orography in Asian monsoon intensity
.
J. Atmos. Sci.
,
66
,
2714
2729
, doi:.
Yanai
,
M.
,
1973
:
Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets
.
J. Atmos. Sci.
,
30
,
611
627
, doi:.
Yanai
,
M.
,
C.
Li
, and
Z.
Song
,
1992
:
Seasonal heating of the Tibetan Plateau and its effects on the evolution of the Asian summer monsoon
.
J. Meteor. Soc. Japan
,
70
,
319
351
, doi:.
Yang
,
J.
,
Q.
Liu
,
S. P.
Xie
,
Z.
Liu
, and
L.
Wu
,
2007
:
Impact of the Indian Ocean SST basin mode on the Asian summer monsoon
.
Geophys. Res. Lett.
,
34
,
L02708
, doi:.
Yang
,
J.
,
Q.
Liu
, and
Z.
Liu
,
2010
:
Linking observations of the Asian monsoon to the Indian Ocean SST: Possible roles of Indian Ocean basin mode and dipole mode
.
J. Climate
,
23
,
5889
5902
, doi:.
Yao
,
T.
, and Coauthors
,
2012
:
Different glacier status with atmospheric circulations in Tibetan Plateau and surroundings
.
Nat. Climate Change
,
2
,
663
667
, doi:.
You
,
Q.
,
J.
Min
,
W.
Zhang
,
N.
Pepin
, and
S.
Kang
,
2015
:
Comparison of multiple datasets with gridded precipitation observations over the Tibetan Plateau
.
Climate Dyn.
,
45
,
791
806
, doi:.
Yuan
,
C.
,
T.
Tozuka
,
T.
Miyasaka
, and
T.
Yamagata
,
2009
:
Respective influences of IOD and ENSO on the Tibetan snow cover in early winter
.
Climate Dyn.
,
33
,
509
520
, doi:.
Zheng
,
X. T.
,
S. P.
Xie
, and
Q.
Liu
,
2011
:
Response of the Indian Ocean basin mode and its capacitor effect to global warming
.
J. Climate
,
24
,
6146
6164
, doi:.

Footnotes

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).