Observational evidence of the relationship between the tropical tropopause and tropical easterly jet streams over the Indian monsoon region

This paper presents the first quantitative relationship between the cold point tropopause (CPT) and tropical easterly jet (TEJ) using radiosonde observations over Gadanki (13.45° N, 79.2° E) during the Indian summer monsoon season 2006–2014. CPT and TEJ peak altitudes ( HCPTandHTEJ ) show amalgams of two categories of variability on the day‐to‐day scale. In category1 HTEJ occurs close to HCPT and they show in‐phase variation. While in Category2 HTEJ occurs far apart from HCPT and they do not show any relationship. For Category1 HCPT and HTEJ are strongly correlated (0.70), as well as HCPT and TCPT (CPT temperature) are moderately anticorrelated (−0.55) significant at a 95% confidence level, indicating the dominance of adiabatic processes. Whereas in Category2 HCPT and TCPT are not significantly anti‐correlated. Thus, when TEJ and CPT are close to each other, it may serve as an indicator for the prevalence of the synoptic‐scale effect.


| INTRODUCTION
The Indian summer monsoon (ISM) is one of the dominant climatological features of the global circulation.It originates due to the differential heating of the land and sea during the Northern Hemisphere summer (Koteswaram, 1960).ISM brings several changes in the meteorological and dynamical features from the surface to the upper troposphere (UT).A noticeable feature in the UT is the tropical easterly jet (TEJ) streams with speeds often exceeding 30 m/s which span around the equator to 20 N and 50-90 E (Krishnamurti & Bhalme, 1976;Roja Raman et al., 2009).It is a thermal wind maintained by the meridional temperature gradient between land and ocean (Hastenrath, 1995;Koteswaram, 1958).During the ISM, the increased frequency of deep convections generally couples with the surface pollutant transports to UT and lower stratosphere (LS; UTLS) (Garny & Randel, 2016;Mehta et al, 2020) modulating its thermal structure directly due to diabatic heating associated with the convection and indirectly due to convectively generated phenomena such as atmospheric waves (Krishna Murthy et al., 2002;Tsuda et al., 1994), the cirrus clouds occurrence (Tseng & Fu, 2017) and surface pollutants transport (Pan et al., 2016).
Cold point tropopause (CPT) plays a vital role in the entry of water vapor into the LS and hence regulating climate variability (Gettelman et al., 2009;Holton et al., 1995).Over the ISM region, 60% of the day-to-day variability of CPT (i.e., out-of-phase variation of the CPT height (H CPT ) and temperature (T CPT )) is driven by an adiabatic process (Jain et al., 2011;Mehta et al., 2010;Mehta et al., 2011), indicating a strong connection with monsoon circulation.Reid (1994) explained the vertical ascent (increase in H CPT ) and adiabatic cooling (decrease in T CPT ) and vice versa as an important mechanism for out-of-phase relations between H CPT and T CPT .The adiabatic process is due to hydrostatic adjustment above the maximum extent of convective heating/cooling (Holloway & Neelin, 2007;Kim et al., 2018).While the random variation of H CPT and T CPT appears due to the diabatic processes such as radiative heating/cooling from cirrus clouds (Hartmann et al., 2001;Boehm & Verlinde, 2000), turbulent mixing of the overshooting air with the environment (Muhsin et al., 2018;Sherwood et al., 2003), a large-scale planetary wave response (Highwood & Hoskins, 1998;Munchak & Pan, 2014).Recently, RavindraBabu et al. (2019) reported a connection between the CPT and onset of ISM.Kulkarni and Verma (1993) observed higher CPT during active monsoon than weak monsoon years (Varikoden & Preethi, 2013).Ramanadham et al. (1969) reported that the TEJ core is situated between the peaks of the frequency distribution of the H CPT over a few stations in the ISM.Jain et al. (2011) observed the coldest CPT occurrences due to the westward propagating wave associated with TEJ.Ratnam et al. (2011) noticed that sometimes TEJ penetrates the LS leading to ozone transport into the UT.
Thus, TEJ has been associated with the cirrus clouds occurrence (Das et al., 2011), gravity wave generation (Ramkumar et al., 2010;Sasi et al., 2000), and horizontal transport of the constituents (Orbe et al., 2015;Ploeger et al., 2017) which in turn may modify the CPT.However, to the best of my knowledge, no systematic study is available about the TEJ and CPT relationships.The main objectives here are to (i) investigate the plausible connection between H CPT and H TEJ , and, (ii) delineate the effect of the TEJ on the H CPT and T CPT relationship.

| DATABASE
High-resolution radiosonde (Väisälä RS-80 &RS-92, and Meisei RS-06G) temperature and zonal wind profiles at $1730 IST over Gadanki (13.5 N, 79.2 E) during June-July-August (JJA) 2006-2014 are used in this study.These profiles having height resolution $25-30 m (sampled at 5 s intervals) uniformly gridded to 100 m.The temperature and wind speed uncertainties are 0.2/0.3K (below/above 100 hPa) and 0.15 m/s for Väisälä (Vömel et al., 2007) and ± 0.5 K and ± 0.2 m/s for Meisei radiosonde, respectively (Kizu et al., 2018).Sounding data available up to 50 hPa (Mehta et al., 2011) are only used.Globally merged infrared brightness temperature (IRBT) data from the National Weather Service Climate Prediction Centre averaged over 0.5 Â 0.5 latitudelongitude centered to Gadanki within a half-hour of 1730 IST are collected to examine the role of convections (refer Mehta et al. (2017) for more details).The gridded (1.5 Â 1.5 ) version of ERA-Interim reanalysis available at 37 standard pressure levels 4 times per day for 1994-2014 is also used.

| RESULTS AND DISCUSSION
3.1 | Typical observations-the relationship between H CPT and H TEJ After examining several hundred temperature and zonal wind profiles, it was found that H CPT and H TEJ frequently occur close to each other and sometimes match.Typical examples of H CPT and H TEJ for the sharp, broad, and multiple tropopause, respectively, are depicted in Figure 1a-f.For the sharp case (Figure 1a, b), both H CPT and H TEJ occur at $17.3 km, T CPT is $194.5 K and TEJ core speed is À49.4 m/s on July 2, 2006.For the broad case (Figure 1c, d), both H CPT and H TEJ also occur at $17.2, temperature between 16.5 and 17.5 km remains almost constant ($190 K), and the zonal wind between 16.8 and 17.7 km is varied a little (À37.8 to À38.7 m/s) on August 10, 2008.An example of the double tropopauses at $17.9 km (H CPT ) and $16.5 km (Lower Tropopause; Mehta et al., 2011) nearly coinciding with double TEJ peaks at $17.6 km (H TEJ ) and $16.4 km (a lower TEJ peak), respectively on July 9, 2010 (Figure 1e, f).These typical zonal wind and temperature profiles show similarity within the tropopause vicinity which is examined by calculating the correlation within the altitudes $12-22 km (Figure S1).The 82% of profiles possess the similarity in the UTLS that strongly indicates the existence of TEJ and CPT relationships.
Such similarity of the temperature and zonal wind profiles in UTLS region (Figure 1) is expected due to thermal wind balance which describes the zonal wind shear ( ∂u ∂z ) and the meridional temperature gradient ( ∂T ∂y ) relationship under hydrostatic equilibrium given as follows: where u, g, f , and y are the zonal wind, the acceleration due to gravity, the Coriolis parameter, and the northward distance, respectively (Andrews, 2010).In the presence of stronger ∂u ∂z , the temperature gradient ( dT dz ) at the sharper tropopause is more substantial (Figure 1a, b) compared to the broad and multiple tropopauses (Figure 1c, f).The tropopause sharpness is defined as the difference between static stability averaged between 1 km above and below the CPT.TEJ is a thermal wind with colder air in the south and warmer air in the north.The ∂T ∂y changes the UT temperature within 300-70 hPa layer within which TEJ occurs.Thus, dT dz would also be changed due to ∂u ∂z resulting the changes in the tropopause sharpness.Note that the broader and multiple tropopauses are sometimes colder than the sharper tropopause which can be explained in terms of the warm (cold) air advection in association with the turning of wind clockwise (counterclockwise) with height (Holton, 2004).Therefore, the relatively warm (cold) sharper (broader and multiple) tropopauses could be related to warm (cold) air advection due to the TEJ streams.
To illustrate the thermal wind balance due to TEJ streams, the temperature profiles from the India Meteorological Department (IMD) station, Chennai (13.0 N, 80.04 E) meridionally separated by about half a degree from Gadanki are used, as shown in Figure 2.Here another set of specific examples is taken because Chennai radiosonde simultaneous to Gadanki were unavailable for the days illustrated in Figure 1.For the sharp case (Figure 2a-c), both H CPT and H TEJ occur at $16.6 km and a strong TEJ stream of speed $53 m/s is located precisely in the tropopause vicinity (T CPT $190 K) on July 10, 2013.The ∂T ∂y between Chennai (13.0 N) and Gadanki (13.48 N) is calculated to obtain the right-hand side term of Equation (1) (hereafter referred to as ∂T ∂y term).The ∂u ∂z and ∂T ∂y terms are roughly the same between the altitudes 15.1-17.7 km indicating that the layer is in the thermal wind balance.Similarly, for the broad case (Figure 2d-f), the zonal wind has a broad (15.9-16.9km) TEJ peak, and the tropopause is also relatively broader.The presence of temperature inversions at 15.9m and 16.9 km on the lower and upper edges of the TEJ broad peak can also be noticed.In this case also, ∂u ∂z and ∂T ∂y terms show a good similarity between 14.4 and 17.4 km indicating the thermal wind balance.The multiple TEJ peaks (Figure 2g-i) at 16 km (speed $39.4 m/s) and 18 km (speed $34 m/s) are associated with multiple tropopauses at 16 km (temperature $193.2K) and 18 km (T CPT $191 K), respectively.In this case, TEJ peak (H TEJ ) lies $2 km below the H CPT coinciding with UT temperature inversion (Fujiwara et al., 2003) which is in thermal wind balance.The frequent convection (vertical motion) with an average outflow level $12-14 km (Gettelman & de F. Forster, 2002;Mehta et al., 2008) occurs during the ISM that would cause the tropospheric layer away from the hydrostatic equilibrium, however, not the layer above the top of the maximum convective heating which will remain in the thermal wind balance.

| Temporal variation of H CPT and H TEJ and an approach to quantify their relationships
The day-to-day variability of H CPT and H TEJ during JJA 2006 is shown in Figure 3a and for JJA 2007-2014 in Figures S2a-S9a, respectively.In general, H CPT (H TEJ ) varies in the range of 15.2-19 km (13-19 km) within the ISM itself (Mehta et al., 2010;Ratnam et al., 2011).Though CPT is generally lower during JJA, it can occasionally occur as high as 19 km (Annamalai & Mehta, 2022;Mehta et al., 2011).From the H CPT tendency (day-to-day difference) as shown in Figure S10a, H CPT remains unchanged for $2% times while increases (decreases) for $52% (46%) times.Similarly, H TEJ remains unchanged for $6% times while it decreases (increases) for $48% (46%) times (Figure S10b).H CPT has larger day-to-day variability compared to H TEJ indicating that the later is governed by a synoptic-scale process while the former is in the balance between radiative and convective processes.Generally, H TEJ fluctuates abruptly especially during early June when synoptic-scale forcing is weak (Figures 2a, S3a-S9a).The H TEJ often occurs closer to H CPT and found to be within 1 km for about 63% (Figure S10c).
As shown in Figures 3a and S2a  mean temperature profile over July-August from individual temperature profiles.The waves with periods 8-12 days are significant (above the cone of influence) during July 20-August 9, 2006, when H CPT coincides with H TEJ .The amplitude and phase of the temperature and zonal wind for a 12-day wave period are obtained as shown in Figure S12.The temperature and zonal wind amplitudes are $2.0 K and $6 m/s in altitudes $15-20 km, respectively (Figure S12a).The downward phase propagation shows that the temperature phase leads to the zonal wind phase indicating the Kelvin wave propagation (Figure S12b) (Mehta et al., 2013;Tsuda et al., 1994).Thus, the Kelvin wave can sometimes modify both the temperature and wind fields, resulting in coincidence and similar variations in the H CPT and H TEJ .However, they also coincide with other timings, irrespective of the wave occurrence.The wavelet analysis for JJA 2008-2014 are shown in Figures S3d-S9d, respectively.The wavelet analysis for JJA 2007 is not performed due to large data gaps.The CPT and TEJ relationships are also analyzed over Nagpur (21.1 N, 79.05 ) at 17:30 IST located in the ISM region during 2006-2013.The CPT and TEJ altitudes over Nagpur depict a similar relationship as observed over Gadanki (Figure S13).

| Statistical analysis of the TEJ and CPT relationship: Plausible link with adiabatic and diabetic processes
In total 707 out of 828 days of observations of H CPT and H TEJ during JJA 2006-2014 are available for analysis.Out of which, 65% and 35% days are observed for the cases when H CPT and H TEJ are "close to each other" and "far apart," respectively.Among the "close to each other" cases, 76% belongs to Category1 and the remaining 24% cases occur isolated.Similarly, among the "far-apart cases," 44% belong to Category2, and the rest 56% cases are isolated.The probability distribution of the difference between H CPT and H TEJ ranges from À2.9 to 4.0 km (overall monsoon), between À0.9 to 0.9 km (Category1), and between À2.9 to À1.0 km and 1 to 4 km (Category2) (Figure 4a-c).The probability distributions of the difference between CPT and TEJ height depict that TEJ occurs close to the CPT most of the time over Nagpur similar to as observed over Gadanki (Figure S14).
The H CPT and H TEJ is randomly related in the overall data (Figure 4d) as well as for Category2 (Figure 4f) while significantly correlated (r = 0.70) for Category1 (Figure 4e).The results remain the same even if all the isolated H CPT and H TEJ cases are included (Figure 4e, f).However, T CPT and H TEJ are significantly anticorrelated (r = À0.32)(Figure 4g) unlike the correlation between H CPT and H TEJ in overall data which could be attributed as T CPT is more sensitive to tropospheric processes, especially infrared warming, and therefore tropospheric temperature profile whereas H CPT is more sensitive to the ozone heating and dynamical warming (Thuburn & Craig, 2000) associated with stratospheric meridional circulation.T CPT and H TEJ are also significantly anticorrelated (r = À0.55) for the Category1 but poorly anticorrelated (r = À0.16) for Category2 (Figure 4h, i).The correlation between H CPT and T CPT (Figure 4j-l) are similar to H CPT and H TEJ .Thus, H CPT and T CPT are significantly anticorrelated indicating the dominance of the adiabatic process under Category1.However, H CPT and T CPT can also be affected by diabatic processes such as dynamical heating, ozone heating, and the occurrence of cirrus clouds (Ali et al., 2020;Mehta et al., 2010;Reid & Gage, 1996;Thuburn & Craig, 2000), resulting in an insignificant correlation for Category2.The above analysis holds for each monsoon season from 2006 to 2014, except for 2011 and 2014, listed in the Tables ST1  and ST2.TEJ peak speed $35-39 m/s irrespective of their cases.Whereas, H CPT (H TEJ ) is at 16.6 km (16.2 km), 16.5 km (16.3 km), and 16.9 km (15.6 km) for overall, Category1 and Category2 cases, respectively.Note that the mean H CPT T CPT ð Þand TEJ peak speed (H TEJ ) obtained by averaging the daily values are relatively higher by 0.1-0.2km (colder by $2.0 K) and stronger by $3-5 m/s (higher by $0.1-0.2 km), respectively compared to those obtained from the corresponding mean profiles.On average, TEJ occurs 0.2 km (1.3 km) below the CPT under Category1 (Category2).Both temperature and zonal wind profiles have sharper peaks for Category1 and broader for Cate-gory2.The TEJ occurring close to CPT is relatively stronger and can enhance the troposphere-stratosphere exchange due to stronger horizontal advection (Das et al., 2011;Holton et al., 1995;Park et al., 2007) <191 K occurs more frequently under Category 1 compared to Category2 (figure not shown), the entry of water vapor with mixing ratio less than 3 ppmv into the stratosphere is likely to occur more frequently in Category1.

| Mean temperature and zonal wind profiles for overall, Category1 and Category2
3.5 | Relationship between the temperature at 100 hPa, zonal wind at 100 hPa, and OLR over the ISM region To further emphasize the TEJ and CPT relationships, the climatological spatial structure of the temperature and zonal wind at 100 hPa (around 16.6 km) during JJA using ERA-Interim data and the OLR data from the NCEP/ NCAR reanalysis are shown in Figure 6.During the ISM, the lowest temperature has a characteristic horseshoe shape in the Indian Ocean and Pacific region as a response to the Kelvin and Rossby modes in the eastern and western parts of the stationary heating (e.g., Highwood & Hoskins, 1998) (Figure 6a).The climatological TEJ is located between 30 E-120 E longitude and 0-25 N latitude (Figure 6b) and the climatological OLR reveals the deepest convection (<220 W/m 2 ) occurring in the northeast part of the Bay of Bengal and India (75 E-110 E) (Figure 6c).The colder region extension in the ISM has been attributed to deep convection and its relationship with low OLR distribution roughly follows the Gill solution (Gill, 1980).The similar temperature and zonal wind distributions at 100 hPa (Figure 6d) indicate a possibility of cold air advection from the South China Sea to the ISM due to TEJ. Figure S15 shows monthly mean contour maps of the temperatures at 100 hPa (Figure, S15a), T CPT (Figure S15b) and H CPT (Figure S15c) (obtained using COSMIC radio occultation data), and zonal wind at 100 hPa from June to September 2009.There is similar structure of the temperature at 100 hPa and CPT.The CPT colder than 191 K occurring at an altitude higher than 17.5 km will be conducive for the troposphere-stratosphere exchange through the freeze-drying mechanism.The colder tropopause structure observed in the ISM region closely resembles the TEJ (zonal wind speed <À30 m/s) structure shown in Figure S15d.

| SUMMARY AND CONCLUSIONS
TEJ shows a large day-to-day variability which is expected because of variation in the meridional temperature gradient due to ISM variability.About 65% times H TEJ and H CPT occures "close to each other" and 35% occur "far apart."Out of these "close to each other" cases, 76% occur continuously for 3 days or more (Category1).H TEJ and H CPT show in-phase variation and are significantly correlated under Category1 while they are not related for Category2.That is H CPT and T CPT under Cate-gory1 are dominated by the adiabatic processes while Category2 are dominated by the diabatic processes.Thus, whenever TEJ and CPT are close to each other, it can serve as an indicator of the dominance of adiabatic processes.Findings from this study have far-reaching implications for understanding the variability and trend of surface energy balance and stratospheric chemistry due to enhanced cross-tropopause transport of the surface pollutants via the Asian summer monsoon anticyclone.
"Data/Experiment" tab of the NARL website (www.narl.gov.in).The IMD radiosonde data for Chennai (Madras) station code (43279 or VOMM) is available from the link http://weather.uwyo.edu/upperair/sounding.html.The supplementary material also provides information on how to download the radiosonde data.IRBT data can be obtained from NASA Goddard Earth Sciences Data and Information Services Center (GES DISC).

CONSENT FOR PUBLICATION
The author certifies that this paper consists of original, unpublished work not considered for publication in Wiley.The original manuscript and figures will be transferred, following the instruction by the Editorial Committee when the paper is accepted.ORCID Sanjay Kumar Mehta https://orcid.org/0000-0003-1439-7910

F
I G U R E 1 Typical temperature and zonal wind profiles showing (a) sharp tropopause, and (b) TEJ observed on July 2, 2006.(c)-(f) and (e)-(f) are the same (a), (b) but observed on June 12, 2010 and July 9, 2010, showing broad and multiple tropopauses and TEJ cases, respectively.Solid dots and open circles denote the H CPT and H TEJ , respectively.
-S9a, H CPT and H TEJ variabilities can be mainly classified into two categories.In Category1, they either coincide or occur close to each other and vary in phase while in Category2, they occur far apart and vary out-of-phase.To quantify H CPT and H TEJ relationships, their absolute difference (ΔH CPTÀTEJ ) and the corresponding climatological mean (ΔH Clim ) are obtained as shown in Figure 3b (for JJA 2006) and Figures S2b-S9b (for JJA 2007-2014).The ΔH Clim is $0.85 km.The days when H CPT and H TEJ are close to each other (far apart), ΔH CPTÀTEJ falls below (above) ΔH Clim .Generally, H CPT and H TEJ "close to each other" and "far apart" occur in a group.Thus, the days when ΔH CPTÀTEJ is lower (greater) than the ΔH Clim for three consecutive days or more are classified as Cate-gory1 (Category2).However, CPT may randomly coincide with TEJ for isolated (one or two) days, 3 days or more are considered to represent a synoptic-scale feature.For JJA 2006, five and three episodes of Category1 and Cate-gory2, respectively are observed (Figure 3a, b).For JJA 2007-2014, Category1 and Category2 episodes are depicted in Figures S2-S9 . The time series of IRBT is shown in Figure 3c and Figures S2c-S9c to examine the role of the convection on CPT and TEJ relationships.Category1 and Cate-gory2 cases were found to occur during clear sky and convective days indicating that H CPT and H TEJ relationships respond to large-scale synoptic conditions (Figure S11a, b).Furthermore, the downward shift in the H CPT during the last week of July 2006 suggests the possible effect of the planetary wave.To examine its role, the continuous time series of the temperature anomalies averaged over 16-17 km during June 26-August 22 2006 is subjected to Morlet wavelet analysis (Figure 3d).The temperature anomalies are obtained by subtracting the F I G U R E 2 Typical profiles of (a) temperature, (b) zonal wind, and (c) zonal wind shear (left-hand side of Equation (1)) and the term involving meridional temperature gradient (righthand side of Equation (1)) for sharp case observed on July 10, 2013.(d)-(f) and (g)-(i) are the same as (a)-(c) but for broad and multiple cases observed on August 18, 2010 and July 23, 2013, respectively.Dashed and dash-dotted lines represent the layer is in the thermal wind balance.

F
I G U R E 3 Time series of (a) H CPT and H TEJ , (b) ΔH CPTÀTEJ (c) IRBT and (d) wavelet spectrum of temperature (in terms of power) at 16-17 km during JJA 2006.The up (green) and down (magenta) hatches indicate the Category1 (Category2) case.The horizontal dashed line in (b) represents ΔH Clim over 2006-2014, and the white curve in (d) represents the cone of influence.

Figure 5
Figure 5 presents the mean and standard deviation of the temperature and zonal wind profiles for the overall, Cate-gory1 and Category2 cases.They show T CPT $193 K and . As T CPT F I G U R E 5 Average profiles of temperature (T; black line) and zonal wind (U; red line) along with their one standard deviation for (a) overall, (b) Category1, and (c) Category2.CPT altitude and temperature and TEJ altitude and TEJ peak value are also shown.F I G U R E 6 Climatology of the (a) temperature at 100 hPa, (b) zonal wind at 100 hPa, and (c) OLR during the ISM months June-August for the period 1994-2014.(d) Climatological mean temperature <194 K at 100 hPa (thin line), zonal wind <À23 m/s at 100 hPa (thick line), and outgoing long wave radiation (OLR) <220 W/m 2 (dashed line).