Characteristics of Atmospheric Heat Sources in the Tibetan Plateau-Tropical Indian Ocean Region

The atmospheric heat sources (AHS), also known as the apparent heat sources, are the sum of the surface sensible heat, precipitation condensation latent heat, and net atmospheric radiation (Yanai et al. ^[1]). The AHS are the heat engines driving the planetary scale circulation (Yanai and Tomita ^[2]). There is strong AHS in the Asian monsoon region in summer (Li and Yanai ^[3]), and the spatial distribution of AHS determines the atmospheric circulation (Wu and Liu ^[4]) which affects the weather systems and precipitation (Yao et al. ^[5]).

The thermal condition of the Tibetan Plateau (TP) and the tropical Indian Ocean (TIO) are the two key interactive systems that modulate Asian monsoon. As the 'Third Pole'of the world, the TP has an average elevation of more than 4 km and spans about 31 longitudes. The formation and maintenance of the South Asian High (Ye and Zhang ^[6]) and the onset of the South Asian summer monsoon (SASM) are closely related to the thermodynamic effects of the TP (Wu and Zhang ^[7]). Studies on the TP AHS, including quantitative estimation of the AHS (Luo and Xu ^[8]; Luo et al. ^[9]; Xie et al. ^[10]), and determination of its vertical structure (Yanai and Tomita ^[2]; Zhong et al. ^[11]), spatial and temporal distributions (Jiang et al. ^[12]; Yanai et al. ^[1]; Zhao and Cheng ^[13]) have got many conclusions. Hu and Duan have suggested that the influence of the AHS in the TP on the East Asian summer monsoon precipitation is more important than the tropical Indian Ocean Basin mode ^[14]. The thermal condition of the TIO also has an impact on the interannual variability of the East Asian summer monsoon (Huang et al. ^[15]) and the TP AHS (Jiang et al. ^[12]; Ji et al. ^[16]). Meanwhile, the interannual variation of summer AHS over the central and eastern TP is significantly affected by convection around the western Maritime Continent (Jiang et al. ^[12]). Model results indicate that the TP heating regulates the Asian monsoon circulation through air-sea interactions (He et al. ^[17]; Wang et al. ^[18]).

The thermodynamic equations show that the local temperature changes are the result of horizontal advection, vertical transportation, and AHS. In most previous studies, the atmospheric temperature was chosen as the index to study the land-sea thermal contrast (Dai et al. ^[19]), yet the distribution of temperature is inconsistent with the heat sources. Hence, the thermal contrast between the TP and the TIO is an important factor for studying the SASM. The greater the land-sea thermal contrast, the stronger the SASM circulation.

Despite these works, our understanding of the thermal effect of the TP-TIO region on the SASM remains insufficient, since the AHS in this region has been discussed widely but often separately. The seasonal variability of the TP AHS is out of phase with the TIO AHS (Zhang et al. ^[20]), and the summer AHS shows significant meridional gradient in this region (Zhang et al. ^[21]). The purpose of the present study is to investigate temporal and spatial distributions of the AHS in the TP and the TIO, and theirs link to the SASM on an interannual scale. In the rest of this paper, the data and methods are described in sections 2 and 3, respectively. Section 4 gives the results, and the summary and discussion are provided in section 5.

The data used in this study are as follows.

(1) Station data. We used data from 133 observation stations (Fig. 1) provided by the China Meteorology Administration (CMA). Variables include 2-m air temperature, surface temperature, 10-m wind speed, and precipitation from 1980 to 2012. All these data went through quality-control procedures to eliminate erroneous data and ensure homogeneity. Then the station data were interpolated to the grid points to calculate the AHS.

Figure 1.  The 133 observation stations. The black spots represent stations, the black curve lines and green curve lines represent rivers and altitude≥2500m, respectively, and the shading indicates the topography (m).

(2) Satellite radiation data from the Global Energy and Water Exchanges-Surface Radiation Budget (GEWEX-SRB) with a resolution of 1.0° × 1.0°. Radiation fluxes included the downward and upward shortwave and longwave fluxes at both the top of the atmosphere and the surface. The monthly net atmospheric radiation was calculated from 1984 to 2008. According to the direct method (see section 3), we can use station data and satellite data to obtain the AHS on the TP (75°E-103°E, 28°N-38°N) from 1984 to 2008.

(3) Four sets of monthly reanalysis data: the ERA-Interim (ERA-I) data from the European Centre for Medium-Range Weather Forecasts with a resolution of 0.75° × 0.75° (Dee et al. ^[22]), the Japanese 55-year Reanalysis (JRA-55) data from the Japan Meteorological Agency (JMA) with a resolution of 1.25° × 1.25° (Kobayashi et al. ^[23]), the MERRA2 data from the NASA Goddard Space Flight Center with a resolution of 0.67° × 0.5° (Gelaro et al. ^[24]), and the National Centers for Environmental Prediction/Department of Energy (NCEP/DOE) data with a resolution of 2.5° × 2.5° (Kanamitsu et al. ^[25]). The present study calculated AHS and atmospheric moisture sources (AMS) by using horizontal wind, vertical wind, temperature, specific humidity on pressure levels and the surface pressure in the domain 30°E-120°E, 20°S-45°N from 1980 to 2016 with four reanalysis datasets.

(4) Global Precipitation Climatology Project (GPCP) precipitation data. This is a monthly precipitation dataset from 1979 to the present that combines data from rain gauge stations, satellites, and sounding observations into 2.5° × 2.5° global grids.

(5) The monthly outgoing longwave radiation data from the National Oceanic and Atmospheric Administration with a resolution of 2.5° × 2.5°, and the geopotential height data from NCEP/DOE with a resolution of 2.5° × 2.5°.

The AHS is the sum of surface sensible heating, condensation heating, and radiative heating based on direct method (formula (1)), and <Q₁> is the vertically integrated AHS. Similarly, the AMS is defined by formula (5) and <Q₂> is the vertically integrated AMS. The components of <Q₁> are obtained from formulas (2)-(4), respectively. Q_L, Q_SH, and <Q_R> are condensation heating, surface sensible heating, and vertically integrated radiative heating, respectively.

In the above formula, angle brackets represent the integration of the whole atmosphere. S and F are the shortwave and longwave radiative fluxes. The subscripts "∞" and "0" represent the fluxes at the top of the atmosphere and at the surface of the Earth. Upward arrows denote upward radiative fluxes, and vice versa. c_p =1004J kg⁻¹ K⁻¹ and it is the drag coefficient for heat, and C_D=0.004 and it is the specific heat of dry air at constant pressure. ρ_air and ρ_w are air density and water density, respectively. T_s, T_a, V₁₀, P_r and E represent the surface temperature, 2-m air temperature, 10-m wind speed, precipitation, and evaporation, respectively. L= 2.5 × 10⁶J kg⁻¹ and it is the condensation heating coefficient.

The indirect method is also called the"residual diagnosis method". <Q₁> and <Q₂> are defined by formulas (6) and (7), which are the sum of the local variation term, the advection variation term, and the vertical variation term. According to formulas (6) and (7), we can obtain the AHS and AMS at each layer, also known as the heating rate (K d⁻¹). The overbar in the formulae represents the regional average value. θ and q are potential temperature (K) and specific humidity (g kg⁻¹), respectively. $\vec{V}$ and ω are horizontal wind speed (m·s⁻¹) and vertical velocity (Pa s⁻¹).

For simplicity, we used Q₁ to represent <Q₁> and Q₂ to represent <Q₂> hereafter. A value of Q₁ greater (less) than zero indicates that the atmosphere has a net heat gain (loss). In general, a positive value of Q₁ is called a heat source and a negative value is called a heat sink. A value of Q₂ greater (less) than zero means that there is a net moisture loss (gain). If the spatial distributions of Q₁ and Q₂ are consistent and both are positive, this indicates that the AHS is mainly dominated by precipitation latent heating. Section 4 shows the AHS as Q₁ and the AMS as Q₂ for simplicity. According to formula (1), the present study used station data, radiation data and four reanalysis data to calculate the AHS of the TP in the 75°E-103°E, 28°N-38°N area. Meanwhile, reanalysis data based on formula (6) and formula (7) were used to calculate the AHS and AMS in the 20°S-45°N, 30°E-120°E area. The TIO is defined in the 15°S-5°N, 60°E-100°E area.

The Pearson correlation coefficient (r_kl) defined by formula (8) is used to measure the correlation between the observed AHS values and those from the four reanalysis data. In formula (8), x is the AHS, and r_kl is the correlation coefficient calculated from observational data (k) and reanalysis data (l), respectively. Temporal correlation coefficients in Fig. 2 and Table 1 were calculated after averaging the AHS over the TP (75°E-103°E, 28°N-38°N) which maskouts the altitude ≥2500m for observational and reanalysis data from 1984-2008.

Figure 2.  The monthly variations of correlation coefficients for Q₁ between the observational data and those from four reanalysis datasets from 1984 to 2008 on the TP (The average area is in the 75°E-103°E, 28°N-38°N which maskouts the altitude ≥2500m; I and D represent the indirect method and the direct method, respectively).

Figure 2 shows that the correlation coefficient is greater in July-September while the JRA-55 dataset is the greatest among the four reanalysis datasets. In summer, the correlation coefficients all have passed the 95% significance test except for the MERRA2 with indirect method (Table 1). The JRA-55 dataset not only assimilates the upper tropospheric atmospheric observations but also uses a 4D assimilation technique for the upper tropospheric observations since 1958 (Fujiwara et al. ^[26]). Hence, we can see that no matter which method is used, the JRA-55 data shows the best correlation with the observational data (Table 1). Hu and Duanalso concluded that the JRA-55 data can best describe the TP thermal heating ^[14].

On the basis of the above analysis, and in view of the scarcity of observational data in the TP, we used the JRA-55 data to further study the temporal and spatial characteristics of the AHS in the TP-TIO region, and theirs link to the SASM.

The TP (75°E-103°E, 28°N-38°N) and TIO (15°S-5°N, 60°E-100°E) regions are located to the south and the north of the SASM respectively. The Asian monsoon regions are heat sources in summer, while the TP and TIO are strong heat source centers along their latitudes (Fig. 3). The TP has a strong heat source center, with the maximum heating (Q₁≥300 W m⁻²) located on the southern slope of the TP, and the TIO maximum also reaches 300 W m⁻² in the south of the equator. The AHS in the TP is accompanied by lower wind convergence, while the AHS in the TIO is located in the area of southeast trade winds and equatorial westerly winds (Fig. 3(a) and (b)). Studies have shown that the TP thermal forcing and the zonal land-sea thermal contrast between East Asia and the Pacific region have a profound influence on the East Asian monsoon system (He et al. ^[27]; Zhang et al. ^[28]); therefore, it is necessary to study the meridional thermal contrast between the TP and the TIO.

Figure 3.  The zonal anomaly Q₁ (shaded, W m^-2) and horizontal wind (vector, m s^-1) in summer during 1980-2016. (a) The TP and winds at 650 hPa, and (b) the TIO and winds at 850 hPa. The green curve represents altitude ≥2500 m.

The seasonal variations of AHS in the TP and the TIO are almost reversed (Fig. 4). The TP is a heat source from April to September and that the strongest intensity is in July (Q₁ = 113 W m^-2), while the TIO AHS gets the minimum in June (Q₁=46W m^-2) and the maximum in January (Q₁ = 163 W m^-2). In summer, the TP AHS is mainly dominated by precipitation latent heat, which coincides with the strongest precipitation in July (Feng and Zhou ^[29]). Therefore, the TP AHS reaches the maximum in July. The TP is a cold source in winter and the strongest intensity is in January (Q₁ = -113 W m^-2), which is consistent with the monthly AHS variation in Duan et al. ^[30]. However, quantitative estimations of the monthly AHS in previous studies (Yanai et al. ^[1]; Zhao and Chen ^[13]; Ji et al. ^[16]) are considerably different, which may be due to differences in regions, calculation methods, data, study periods, and so on (Liu and Li ^[31]). For the TIO, it's a heat source all year round due to the convection precipitation latent heat. The standard errors of AHS in the TP and TIO are 78.97 and 38.56 respectively, which indicate that the seasonal variability in the TP is more obvious.

Figure 4.  The monthly variations of Q₁ from 1980 to 2016 in the TP (The area is in the 75°E-103°E, 28°N-38°N which maskouts the altitude ≥2500m) and the TIO (60°E-100°E, 15°S-5°N) (Black bars indicate the standard error; units: W m^-2).

In Fig. 5(a1), the heat sources center in the east and west of the TP from March to September and are strongest in summer. The intensity is stronger, and the extent is wider for the eastern heat sources center than those for the western heat source center, and Q₁ ≥ 100 W m^-2 occurs in the east from May to July but only occurs in July in the west. The AMS intensity in the western part of the TP is about twice as strong as that in the eastern TP (Fig. 5(a2)). The AMS is the bias of precipitation and evaporation which represents the precipitation latent heat (formula (5)). By comparing the Fig. 5(a1) and (a2), we can infer that although the AHS is dominated by the precipitation latent heat in the TP, the contribution of AMS to AHS shows large zonal inhomogeneity. The AHS is stronger in the east than in the west, and vice versa for the AMS. This is probably because the surface evapotranspiration is much weaker in the west than in the east (Su ^[32]), which results in the AMS being smaller in the east than that in the west over the TP. The seasonal variations of the AHS and AMS at high altitude (Fig. 5(b1) and (b2)) are the same as the mean distribution and the value of the AMS and AHS at high altitude are similar because the less is affected by the underlying surface.

Figure 5.  The temporal-zonal distributions of Q₁ (1) and Q₂ (2) from 1980 to 2016 over (a) the TP, (b) the TP altitude ≥2500 m, and (c) the TIO. The black dash lines represent the TP (75°E-103°E) in a1, a2, b1 and b2, and represent the TIO (60°E-100°E) in c1 and c2 (W m^-2). The averaged latitudes are 28°N-38°N for the TP and 15°S-5°N for the TIO.

The temporal-zonal distribution of the AHS and the AMS in the TIO are basically the same (Fig. 5(c1) and (c2)), which indicates that the AHS is mainly dominated by the precipitation latent heat in this region. In summer, the intensity of the heat sources and moisture sinks are weaker than that in the TP. In addition, the summer atmospheric thermodynamic properties on both sides of 100° E are very different, and the gradient of latitudinal variation is also large. The region of 100° E-120° E, corresponding to Indonesia, is a cold source and moisture source, and the area between 60° E and 100° E is a heat source and moisture sink. The reason for this phenomenon may be related to the distribution of land and sea, and ocean currents.

The BOB, the southern slope of the TP, and the west of the India Peninsula are the strongest heat sources in summer (Fig. 6(a1)). The strong heat sources in the central and eastern TP in summer is due to latent heating (Fig. 6(a2)). Jiang et al. divided the TP into western, central and eastern domains and reached similar conclusions ^[12]. For the TIO, there is a strong AHS and AMS in the northwest-southeast direction, which indicates that the AHS is dominated by the precipitation latent heat, and the heat sources and moisture sink centers are located at 10° S near 90° E (Fig. 6(a2)). The heat sources in the TIO, BOB and the TP constitute the meridional distribution, while the AHS in the Arabian Sea, Indian Peninsula, BOB, and Indochina Peninsula constitute the zonal distribution. There are cold source centers located to the southeast of the TIO in summer. Zhang et al. also found that there is an obvious cold source in summer in the southeast of the TIO ^[20].

Figure 6.  The zonal anomalies of Q₁ (a1) and Q₂ (a2) and their linear trends (b1 and b2, W m^-2 yr^-1) in summer from 1980 to 2016. The TIO is indicated by a green box and the green curve indicates the TP altitude ≥2500 m. The dotted areas passed the 95% significance test.

In general, the land is a strong AHS in summer, such as the"elevated heat source"of the TP, and the ocean is relatively a weak heat source or a strong heat sink. The intensity of Q₁ in the TP and TIO in July can reach 113 W m^-2 and 58.4 W m^-2, respectively. The thermal contrast between the sea and land in summer can be measured by the difference in Q₁ between the TP and the TIO. The larger the difference in Q₁, the stronger the meridional thermal contrast.

According to Fig. 6(b1), the TIO Q₁ shows a significant increasing trend (the maximum is about 40 W m^-210yr^-1), which indicates that the atmospheric heating is significantly enhanced in the area. In the TP, the Q₁ increased on hinterland with the trend about 10 W m^-210yr^-1, while the southeast of the TP shows a significant decreasing trend (the maximum is about -40 W m^-210yr^-1). The increasing trend of Q₁ in the TIO is greater than that in the TP, which indicates that the thermal contrast between the TP and the TIO in summer is decreasing. The linear trend distribution of Q₁ has an obvious meridional feature, in that Q₁ at 30°N, 10°-20° N, and 10° S-0° along 90° E has an increasing trend, while Q₂ does not (Fig. 6(b1) and (b2)). The spatial distributions of the linear trends for Q₁ and Q₂ in the TP are different. In addition, both Q₁ and Q₂ are increasing significantly in the BOB.

The Hadley circulation is a fundamental regulator of the earth's energy budget due to the redistribution of energy from tropics to higher latitudes (Liu et al. ^[33]), and its location and intensity are closely related to the AHS. The TIO Hadley circulation in the southern hemisphere extends across the equator to around 30° N, and the northward mass transport center lies between 0° and 15°S at 600-400 hPa, with an intensity of about -3.5 × 10⁹ kg s^-1 (Fig. 7(a)). The strong ascending branch of the Hadley circulation is located at the south of the TP, while the descend branch lies in 30° S-15° S. The southern hemisphere Hadley circulation shows a weakening trend driven by the strengthened AHS in the TIO (Fig. 6(b1), Fig. 7(b)), while the northern hemisphere Hadley circulation has an increasing trend due to the enhancement of the vertical wind near 15° N (Fig. 7(b)). The decreased winter northern Hadley circulation in the TIO indicates that the meridional energy transported from the tropics in southern hemisphere to the northern hemisphere is reduced.

Figure 7.  (a) The mass stream function (shaded, units: 10¹⁰ kg s^-1) and vertical velocity (black curve, units: Pa s^-1) of zonal averaged along 30°E-100°E in summer from 1980 to 2016. (b) The linear trends (Shaded is mass stream function trend (10¹⁰ kg s^-1 yr^-1), and black curve is vertical velocity trend (0.01 Pa s^-1 yr^-1).

Due to the decrease of AHS in the southeast and the surrounding of the TP (Fig. 8, Fig. 6 (b1)), the AHS has a weakening trend (Fig. 8). It implies that the ability of the TP heating environment decreases; at the same time, the ascending motion, the lower layer convergence and the upper divergence, and the suction effect of the TP are all weakened (Liu et al. ^[34]). The southwesterlies over the northern Indian ocean also weakened when the TP heating decreased (Wang et al. ^[18]).

Figure 8.  The interannual variations of Q₁ in the TP (The area is in the 75° E-103° E, 28° N-38° N which maskouts the altitude ≥2500m) and the TIO, the IQ, and the SASM_-v in summer from 1980 to 2016. The IQ linear trend is shown with a black dotted line and the linear regression coefficient is -0.075. All data are standardized.

The TIO AHS shows a significant increasing trend before 1998, and then a weaker decreasing trend (Fig. 8). The IQ (formula (9), where the overbar indicates the regional average) is used to measure the thermal contrast between the TP and the TIO. It shows a significant weakening trend, which indicates that the thermal difference between land and sea decreased quickly. The summer heat source on the TP plays an important role in the formation and evolution of the SASM (Wu et al. ^[35]). We defined an SASM index (SASM_-v) by meridional wind shear between 850Pa and 200hPa (formula (10)). Fig. 8 shows that the SASM-v and the QI are similar in interannual variation. The SASM_-v also shows a weakening trend during 1980-2016 and the correlation coefficient between the IQ and SASM_-v is 0.801, passing the 95% significance test.

The TP heating and the Hadley circulation decreased, while the TIO heating increased during 1980-2016, especially during 1980-1998. The IQ shows a significant weakening trend during the period of 1980-1998, whereas during 1999-2016 it maintains a stable low level with no obvious trend (Fig. 8). It indicated that the thermal contrast between the TP and the TIO has changed in the two periods, and during 1999-2016, the thermal contrasts is weaker. In Fig. 9, the southwest airflow at the lower troposphere and the easterly airflow at the upper troposphere are both weak, which implies that the SASM circulation has weakened during 2007-2016 compared with that during 1980-1989. Hence, as the meridional thermal contrast in the TP-TIO region weakens, the SASM circulation also has weakened.

Figure 9.  The horizontal wind speed in summer at 15°N ((a): 850hPa, (b): 200hPa, units: m s^-1).

Comparing the period of 2007-2016 with 1980-1989, we find that there is much more precipitation in the Arabian Sea, the BOB, eastern China and the northwest Indian ocean (Fig. 10(a)). The anomaly precipitation with strong convection is in the Arabian Sea and eastern China where an anticyclone and a cyclone appear, respectively (Fig. 10). The two places are the strong water vapor convergence regions (figure omitted). The AHS is an important factor in determining the South Asian High. The position of South Asia High moves westward because the TP AHS decreased and the AHS in the Iran Plateau and the northwest of the Indian Peninsula increased (Fig. 6(b1), Fig. 9(b), Fig. 10(b)). The south slope of the TP, the Indian Peninsula, the central and southern Indochina Peninsula and southeast Indian Ocean have less precipitation associated with anomalous anticyclones.

Figure 10.  (a) The bias of summer precipitation (shaded, mm d^-1) and 850 hPa horizontal wind (vector, m s^-1) between 2007-2016 and 1980-1989 in the TP-TIO region. (b) The bias of summer outgoing longwave radiation (shaded, W m^-2), and 12560gpm geopotential height (black curve represents 1980-1989, and red curve represents 2007-2016, gpm).

The equatorial Indian Ocean is controlled by abnormal west wind (Fig. 10(a)), and then the sea temperature rises (figure omitted) though the wind-evaporation-SST feedback (Wang et al. ^[18]). The increased SST further increases the TIO AHS, resulting in the abnormal ascent flow near 5° S (Fig. 11(a)). Therefore, the TIO southern hemisphere Hadley circulation weakened.

Figure 11.  (a) The profile of v-w wind bias between 2007-2016 and 1980-1989 (along 90°E; units: m s^-1). (b) The same as (a), but for u-w wind (along 15°N; units: m s^-1).

The meridional section of 90° E and the zonal section of 15° N were used to analyze the vertical circulation in the TIO and the BOB (Fig. 11). It can be seen that there is obvious ascending motion in the whole atmosphere in these two places because the AHS increased (Fig. 6(b1)). The enhanced heat sources can trigger two abnormal meridional circulations in the northern 15°S (Fig. 11(a)). The decreased heat source in the south slope of the TP (Fig. 6(b1)) triggers abnormal downdraft (Fig. 11(a)), resulting in less precipitation. The abnormal zonal circulations appear in the SASM region, with an abnormal descending branch located to the Indian Peninsula and the Indochina Peninsula, and a strong ascending in the BOB (Fig. 11(b)).

The AHS is a comprehensive thermal index of the atmosphere, and its spatial and temporal variations have a significant influence on monsoon circulation. However, quantitative estimation of AHS remains a difficult problem. In this paper, observational data and four sets of reanalysis data were used to calculate the AHS in the TP-TIO region. The statistical correlation analysis shows that the results from the JRA-55 data are closest to the observations. In view of the lack of observations over the TP, the JRA-55 data (1980-2016) were used to study the temporal and spatial distributions of the AHS over the TP-TIO region, and the interannual variation of the land-sea thermal contrast. The major findings are summarized below.

(1) The seasonal variation of the AHS in the TP and the TIO shows great differences. The AHS from April to September is greater than zero in the TP. It reaches the strongest in July and the coldest in January. The TIO is a heat source throughout the year and the monthly variation is"V"shaped. Summer precipitation latent heat is a major contribution to the AHS in the TP-TIO region, and both the AHS and the AMS over the TIO are weaker than those over the TP.

(2) From March to September, there is a heat source center in the east and another one in the west of the TP, and they reach a peak in summer. Both the intensity and extent of the AHS are greater in the east than in the west of the TP, while the AMS shows the opposite pattern. In the TIO, a strong northwest-southeast direction AHS heat center is located near 10° S, 90° E. The thermal properties vary greatly between the east and west sides of 100° E in the TIO region. The linear trend of the summer AHS shows an apparent zonal distribution in the TP-TIO region, while the AMS does not. The AHS increases slightly on the hinterland of the TP, while most of the surrounding areas show a significant decreasing trend.

(3) The IQ is constructed to measure the thermal contrast between the TP and the TIO. It shows a significant decreasing trend, especially before 1998. Comparing the period of 2007-2016 with 1980-1989, we find the thermal contrast between the TP and TIO is weaker. The equatorial Indian Ocean is controlled by abnormal west wind, and with the wind-evaporation-SST feedback, the increased SST further increases the TIO AHS, resulting in weakened Hadley circulation and SASM circulation. The abnormal meridional circulations appear in the northern TIO and the abnormal zonal circulations in the monsoon region. Finally, there are much more precipitation in the BOB during the weakening thermal contrast period.

This study has discussed the temporal and spatial distributions of the AHS in the TP-TIO region and the interannual variation of the land-sea thermal contrast. Some important conclusions have been drawn, but there are still some problems that require further exploration. There are uncertainties in the indicators of thermal contrast between land and sea (Dai et al. ^[19]); using height of the troposphere to define the index has an impact on the results. In addition, the average vertical drop between the TP and the TIO is about 4000m, and the maximum heating height in summer is not consistent (figure omitted). Through the thermal wind relationship, it is necessary to further discuss the relationship between the meridional land-sea thermal contrasts and the SASM with its climatic effects.