The Empirical Inuences of Tibetan Plateau Soil Moisture on South Asian Monsoon Onset

The South Asian High (SAH) location and intensity are linked with the Tibetan Plateau (TP) and Yangtze River basin latent heating. The existing feedbacks of SAH variability is rarely linked with TP soil moisture regulated energy uxes. In this study, remotely sensed soil moisture and global atmospheric reanalysis products are used to quantify the relationship between the TP spring (April, May, June) soil moisture with SAH and South Asian (SA) monsoon onset during 1988–2008. The diagnostic analysis infers that the SAH exhibits a signicant correlation (R ≥ 0.90) with TP spring soil moisture and monsoon onset indices (R ≥ -0.56−-0.61). During the early and late monsoon onset, a signicant anomalous soil moisture regime inuenced the surface energy uxes, which affected the vertical diabatic heating prole. The diabatic heating prole affects the TP ascending motion and SAH intensity, which leads to regional monsoon circulation changes and onset. An asymmetric SAH movement in the upper troposphere appears before the early and late monsoon onset composites and drives the lower tropospheric westerlies/easterlies winds towards the continental SA. The wind shear and transition from prevailing easterlies into westerlies during the pre-onset, onset, and post-onset pentad results in strong/weak ascent in the Bay of Bengal and advances into continental regions. The onset- mechanism further suggested intensied/weaker westerlies/easterlies during early/late onset composites. The SAH intensity and movement are linked with TP soil moisture, which exhibits teleconnections with the regional circulation pattern. A detailed model experiment will be conducted to verify the inuence of soil moisture as a driver of energy uxes and SA monsoon onset.


Introduction
The Tibetan Plateau (TP) thermal and mechanical controls are known as regional climate drivers (Anmin The TP thermal and mechanical forcing are studied with empirical approaches and coupled atmosphericocean general circulation models. Using a coupled atmosphere-ocean circulation model, Abe et al. (2013) found that TP's presence leads to strong air-sea interactions and, consequently, variable onset of SA monsoon. He et al. (2019) used a coupled general circulation model with multiple initializations and prescribed forcing's to study the TP in uences. It was suggested that the combined thermal effects of TP and Iranian Plateau (IP) could modify the tropospheric cyclonic circulation, transporting water vapors to continental SA. Using the Weather Research and Forecast (WRF) model, Wang et al. (2019b) found a signi cant negative relationship between southeastern TP thermal forcing and summer precipitation over Pakistan. Wu et al. (2012) found that the TP thermal controls over both components of the Asian Monsoon (SA and EASM) are signi cantly stronger than its mechanical forcing, whereas the EASM component of the Asian Monsoon is more sensitive to TP thermal pro le than SA. Wang et al. (2017a) argued that the TP control both EASM and SA monsoon systems; without TP uplift, the SA monsoon will be much weaker than the present day, whereas the EASM will disappear. Furthermore, the vertical pro le of TP thermal heating in uences the tropospheric anticyclonic patterns, known as South Asian High (SAH); and the SAH exhibits strong feedbacks with TP thermal heating Studies of TP forcing's have indicated that the thermal process of a sensible-heat-driven air pump (SHAP) is dominant over the mechanical forcing (Wu et al. 2007. However, the land surface parameters, especially soil moisture, have not been explicitly studied in affecting the TP thermal conditions. Soil moisture anomalies can persist for months, affecting the albedo and thermal uxes (Seneviratne et al. 2010), and thus they can in uence the coupled interactions between the land, atmosphere, and oceans However, the role of soil moisture in the TP thermal processes is not well understood up to date, and its possible link with SA monsoon remains unclear. In this study, using satellite-based derived soil moisture and reanalysis data, we attempt to understand how the TP spring soil moisture in uences the atmospheric thermal regimes and how these soil moisture-related regimes in uence the SAH and the SA monsoon onset.

Study Area
The dominant topography of SA are the northern mountains that stretch from east to the northwest with an altitude of > 4000 meters (Fig. 1), and the topography in the central and southern regions are quite complex. The SA climate is diverse; the precipitation varies from 100 mm in the western region to > 1000 mm in the eastern region (Jain et  TP serves as the source of perennial water supply into the main rivers of SA, and it can be termed as "water tank of SA" (Vemsani 2015). TP also in uences the summer precipitation by inducing lower-level westerlies towards continental regions and preventing cold Siberian fronts in winter. The TP soil moisture follows a strong freeze/thaw cycle in winter and summer seasons (Ullah et al. 2018b); soil water thawing is initiated in April, and freezing is initiated in November (Su et al. 2011;Yang et al. 2013).

Data
In-situ soil moisture observations over TP are limited, and thus satellite soil moisture from Special Sensor Microwave Imager (SSMI) is used in this study (Van Der Velde et al. 2014). This daily data, spanning  ). The air temperature, geopotential height, and wind components are obtained from ERA-Interim, which is a global reanalysis product from ECMWF (European Centre for Medium-Range Weather Forecasts). ERA-Interim is upgraded from ERA-40 with a four-dimensional variation analysis scheme (4Dvar), an improved atmospheric model, and an assimilation system (Dee et al. 2011). The spatial resolution of the ERA-Interim reanalysis is 0.75 degrees.
The precipitation data is obtained from Climate Hazards Group InfraRed Precipitation with Station (CHIRPS). CHIRPS is a combination of three different sources of precipitation products, including the IR-derived precipitation merged with climatologies (CHPclim) and blended with station observations (Funk et al. 2015). The CHIRPS product has been validated in SA with a good performance during the monsoon season (Ullah et al. 2019b). The Modern-Era Retrospective analysis for Research and Application (MERRA-2) is used to estimate the TP's vertical diabatic heating pro le. The temperature tendencies due to moist processes, radiation, and turbulent processes at multiple pressure levels were used to estimate the diabatic heating following (Chattopadhyay et  The WB index is de ned as the difference between 850hPa zonal wind averaged for (5°N 15°N, 40°E 80°E) and (20°N 30°N, 70°E 90°E), shown in Eq. (1). The monsoon onset date is declared when the wind shear between these two regions changes its sign from negative to positive for ve consecutive days. The de ned onset date represents changes in both circulation and precipitation, which exhibited a signi cant correlation with the subjective Indian monsoon onset index. The de nition can be seen from Eq. (1), The XV index is based on the tropospheric temperature gradient (averaged between 200hPa-600hPa) for the northern hemisphere (40-100°E, 5-35°N) and southern hemisphere (40-100°E, 15°S-5°N). The XV index is de ned as the rst day when the averaged tropospheric temperature (TT) gradient between the northern and southern hemisphere changes its sign from negative to positive. This index has less in uence from the surface and boundary layer induced uncertainty and appeared to be signi cantly correlated with the Indian monsoon onset index. The de nition can be seen from Eq.  Figure 2 shows the climatology of monthly mean precipitation (Fig. 2a) and spatial precipitation pattern for the monsoon season (JJAS) (Fig. 2b) during 1988-2008. Figure 2 suggests that maximum precipitation is observed from June to September with a peak of > 200 mm in July. SA monsoon onset is partly in June; however, the whole SA experiences the monsoon onset in July, which continues till September. The spatial pattern of total precipitation in JJAS shows more precipitation in the northeastern SA (Fig. 2b) with an amount of > 500 mm. It is reduced to about 300 mm in the central SA, and in western SA, it is less than 80mm. The monsoon onset across the region has been linked with several factors, including the sea surface temperature, westerlies progression, convection, and local continentalscale tipping elements. Less attention has been paid to the in uence of the TP soil moisture-induced thermal pro le in uence on the monsoon onset across time (Zhang et al. 2002;Stolbova et al. 2016;Ullah et al. 2020Ullah et al. , 2021). The following section shows a diagnostic relationship of the soil moisture pro le over the TP and its thermodynamics role in the SA monsoon onset.

Monsoon onset and soil moisture
The standardized SAH, the WB, and the XV indices are shown in Fig. 3. The positive WB and XV anomalies indicate late monsoon onset, and their negative anomalies indicate early monsoon onset, respectively. The XV index shows a negative correlation of -0.56 with the SAH index, the correlation between the WB index and the SAH is -0.61; both correlation coe cients have passed the signi cance test with a 95% con dence level. Thus, it is suggested that the SAH and the SA monsoon may have different mechanisms, and the SAH negatively affects the SA monsoon onset. These ndings complement the previous studies and suggest that the SAH intensity may enhance the divergence aloft accompanied by lower-level cyclonic and convective activities, leading to monsoon onset vortex (MOV) in the SA monsoon domain (Liu et al. 2013;Wei et al. 2014). The TP thermal heating and energy uxes are known to link with the SAH; however, the soil moisture relation with SAH remains unknown, although it is known to play an essential role in land surface energy uxes partitioning.
Hence, a pixel-wise correlation analysis is conducted to understand the TP soil moisture relations with the WB, the XV, and the SAH index, respectively, and the results are shown in Fig. 4. The average soil moisture from April to June (AMJ) indicates springtime soil moisture over the TP (Ullah et al. 2020). It appears that both XV (Fig. 4a) and WB (Fig. 4b) indices are negatively correlated with springtime soil moisture across the TP, with correlation coe cients ranging from − 0.50 to -0.70. However, the SAH index is positively correlated with the springtime soil moisture across the plateau, and the correlation coe cients are ≥ 0.90 (α = 5%) in most areas. The TP soil moisture is hypothesized to modulate the SAH intensity through a certain mechanism, affecting the SA monsoon's onset. Several studies have indicated that soil moisture affects the boundary layer structure and the upper troposphere by altering the vertical thermal processes

Spring soil moisture in uence on TP thermal forcing
We conducted the composite analysis to understand the possible thermal processes that link the TP soil moisture to the SAH intensity and the SA monsoon onset. As shown in Fig. 3, the positive and negative values of the z-score indicate the deviation of the monsoon onset date from its climatological mean onset date. In this regard, the years with negative z-score values were referred to as early-onset composite (1990, 1999, 2002, and 2004) and late-onset composite (1992, 1995, 1997, and 2003) for positive z-score values, respectively. The respective average onset date/pentad for the early and late-onset years are further shown in Table 1. The onset timing estimated from the two indices has an obvious difference, implying that the onset inferred from zonal wind maybe earlier than tropospheric temperature. The onset composites have an evident difference of 20-25 days between the early and late-onset composite. The monsoon onset's regionally indicating the reversal mechanism of the wind and tropospheric temperature, preceded by convective activities and abrupt precipitation in the Indian ocean, which moves towards continental regions in the following days (Liu et al. 2013). Based on these two composites, the TP's spring season thermodynamics is expressed as a precursor of the processes leading to the SAH modulation and the monsoon onset. The soil moisture, sensible heat, and latent heat anomalies in the spring season are analyzed in the earlyand late-onset composites, respectively, shown in Fig. 5. In the early-onset composites, the TP soil moisture shows mostly positive anomalies (Fig. 5a) over the plateau; during the late-onset composites, the TP soil moisture shows negative anomalies across the plateau, except for the southeast corner (Fig. 5b). The associated thermal processes of the contrary composites are further studied. In the earlyonset composite, the sensible heat uxes (Fig. 5c) Figure 6 shows the vertical pro le of latent and sensible heat uxes averaged for 27N to 37N and 85E to 103E. The vertical pro le of latent and sensible heat uxes is for the same region as used in the above studies, but consider the soil moisture effect rather than the precipitation-induced heating. In the early-onset composites (Fig. 6a), both sensible and latent heat anomalies are positive (0.8 K day − 1 ) near the surface; however, at midtroposphere from 500 to 300 hPa, the sensible heat anomalies reduce to zero, and the latent heat anomalies are the strongest, reaching 2.5 K day − 1 . In the late-onset composites (Fig. 6b) The in uence of soil moisture anomalies and, consequently, heat uxes are also associated with the TP diabatic heating's vertical column. Figure 7 shows the vertical pro les of diabatic heating averaged from 27N to 38N in spring of the early-onset (Fig. 7a) and late-onset (Fig. 7b) composites. The diabatic heating has shown positive anomalies in the early-onset composites, which are more than > 1.  Figure 8 shows the SAH anomalies at 200 hPa in the spring of the early-onset and the late-onset composites. In the early-onset composites (Fig. 8a), a high-pressure system prevails (> 18 gpm) over southwestern TP, which stretches towards Iranian Plateau (IP) and the Middle East. Another high-pressure system appears over the Eastern ank of TP, with the center over the Mongolian and Siberian regions. In the late-onset composites (Fig. 8b), the geopotential height in the SAH region shows obvious negative anomalies of <-16 gpm. Over the Indian Ocean and Northern Eurasian Continent, the strong high-pressure system is persistent with an increased geopotential height of > 10gpm. The high-pressure system apparent during the early monsoon onset composite is symmetrically replaced by a low-pressure system centered over the southern and northern TP with an intensity of <-18gpm each, respectively. The SAH (Fig. 8) has a consistently high and low-pressure mode in the spring of both early-and late-onset composites, varying in a tripole pattern. In the early-onset composite, the SAH exhibits two distinct highpressure centers located at the TP's southern and northern sides, respectively. There is a low-pressure system over the Eurasian region, and hence a tripole pattern is formed. In the early-onset composite, the spring soil moisture-induced thermal forcing may intensify the SAH intensity. The wet soil moisture in uences the vertical thermal column aloft through latent heat of condensation released, as Koster et al.
(2016) described. In the late-onset composite's spring season, the vertical heating and energy uxes over TP are weaker, which could be associated with surface energy and negative soil moisture anomalies. In response to the reduced thermal heating, the high-pressure centers at southern and northern TP are replaced by the low-pressure systems, whereas a high-pressure system replaces the Eurasian lowpressure system. The tripole pattern mechanism in monsoon onset is further explored with composites of vertical velocity and meridional wind components over TP. Figure 9 shows the vertical velocity and meridional wind in the spring of the early- (Fig .9a) and late-onset (Fig .9b) composite. In the early-onset composite, an obvious ascending motion exists over the eastern plateau, which produces lower tropospheric cyclonic motion and diverges towards the western plateau. In the late-onset composite, an ascending motion is obvious at the western plateau, followed by a sinking motion at the eastern plateau. Figure 9 infers that in the spring season of the early monsoon onset composites, the intensi ed SAH favors wind ascent towards TP as evident in the eastern TP during early monsoon onset composites and vice versa for the delayed onset composites. The ascent/descent is vastly linked with the convergence of winds from the Indian ocean, including a land-atmosphere ocean thermal contrast, leading to monsoon onset and precipitation Pathak et al. 2017a). In the following section, a pentad-scale movement of the SAH and its associated ascent over the TP is linked with the transition of the equatorial zonal easterlies into prevailing westerlies resulting in monsoon onset due to strong convective activities and precipitation in the SA monsoon domain. Figure 10 shows the pentad evolution of the SAH for early and late monsoon onset composites during the SA monsoon onset phase. For both early and late-onset composites, two pentads, including the pre-onset (onset pentad: -1) and onset-pentad (onset pentad: 0), were selected as an indicator of the SAH location and short-term evolution. During the early onset composite, in the pre-onset pentad (Fig. 10a), the SAH center is located over TP stretched zonally with a weak low located westward of the plateau. Such a high(low) combination indicates a strong ascending motion over TP accompanied by an intensi ed sinking west of the plateau. From the onset pentad (Fig. 10b), the SAH centers of the SAH move northeastward with secondary high generated over Iranian Plateau. The northeastward movement indicates the downstream convective activities initiation in the Bay of Bengal (BOB) due to monsoon onset and establishment of the westerlies and easterlies into the region (Liu et al. 2013). During the lateonset composite, in the post-onset (Fig. 10c) the SAH low is centered over the Iranian plateau extending into the SA in the east, indicating a stable high pressure in the lower troposphere. The TP high-pressure system is farther northwest of the plateau, indicating a weaker ascent. During the onset pentad (Fig. 10d), the two low-pressure systems located eastward of the plateau and over the Iranian plateau replace the SAH centers, as evident during the early onset pentad. The northwest high center of the SAH moves into TP, but weaker intensity indicates an ascent over continental regions located westward of the Bay of Bengal. In conclusion, two aspects of the SAH are evident during the early and late-onset composites. The rst aspect is its potential relationship with TP soil moisture induced thermal forcing that maintains its intensity and controls the monsoon onset. The second aspect of the SAH is its expansion into the high latitudes, which intensify the meridional circulations. Hence, it can be the potential reason for early monsoon onset triggered by TP soil moisture-induced thermal forcing. The ndings complement the previous studies reporting similar characteristics of the SAH (Wu 2002;Wei et al. 2014); however, a model study will further be conducted to validate the ndings. Figure 11 shows the TP meridional wind component, and vertical velocity averaged for the latitudinal range of 27N to 38N across the longitude. During the pre-onset pentad (Fig. 11a), the western and central plateau experience an intensi ed ascent accompanied by descent in the western plateau, which during the onset-pentad (Fig. 11b) moves eastward into the eastern plateau and Bay of Bengal region. In the onset-pentad, the ascending motion over the eastern plateau results in the regional westerlies and easterlies movement towards the SA domain that initiates in the Bay of Bengal and pushes the SAH to the north. A similar pattern of the ascending motion and upper-tropospheric divergent motion was also reported by Ullah et al. (2020) and Liu et al. (2013), attributed to the plateau soil moisture thermal pro le and condensation heating release. During late-onset composite, the pre-onset pentad (Fig. 11c) meridional wind and vertical velocity anomalies are weaker with obvious descending motion evident over the plateau, inferring a weaker thermal pro le of the plateau. The onset pentad (Fig. 11d) has experienced a relatively weaker ascent over the western plateau and Yangtze river basin but rather a stronger descent over the eastern plateau. Such weaker ascent and partly descent over the eastern plateau can potentially be attributed to soil moisture negative anomalies that can lead to weaker thermal pro le and westward shift in the ascent and associated precipitation. To explore the regional westerlies and easterlies convergence into the SA monsoon domain forced by TP soil moisture-laden thermal forcing, the lower tropospheric wind anomalies and vertical velocity are shown at 850 hPa. The wind anomalies shown are a representation of the monsoon onset due to wind shear in the pre-de ned monsoon index (WB-index) representing the westerlies evolution replacing the zonal easterlies. Figure 12 shows the wind components (vectors) and vertical velocity (shaded) for the pre-onset, onset, and post-onset pentads for the early and late-onset composites at 850 hPa. The wind anomalies are used to show the transition of the westerlies and establishment of the subtropical jet stream associated with enhanced moisture contents, enhanced convective activities, and ascent showed by vertical velocity anomalies. During early-onset composites, in the pre-onset pentad (Fig. 12a), the prominent features include zonal easterlies from the paci c warm pool and north-easterlies from South China Sea (SCS) enters into Indian ocean and Bay of Bengal (BOB). These two easterlies merge into the continental westerlies and form prevailing continental easterlies gusting towards eastern Africa. In the onset pentad (Fig. 12c), the equatorial zonal easterlies after entering into Indian ocean changes into zonal westerlies, and the SCS currents after entering into BOB, join the zonal westerlies and advances into the continental landmasses. In the onset pentad, a clear transition of the easterlies into westerlies is evident with stronger intensities and ascent (shaded) in the Indian ocean and peninsular India, which is relatively weaker in the continental regions. During the post-onset pentad (Fig. 12e), the south-westerlies from the Indian ocean and zonal easterlies intensify the ascent, evident from vertical velocity in the continental regions. During the pre-onset pentad of late-onset (Fig. 12b), the zonal equatorial easterlies are much weaker, and southwesterlies also replace the SCS north-easterlies with stable atmosphere and descending motion evident in BOB and Arabian sea. In the onset-pentad (Fig. 12d), the zonal easterlies change into westerlies and advance towards SA, but the BOB easterlies are suppressed, and hence a south-westerlies jet into SCS prevails. In the post-onset pentad (Fig. 12f), the westerlies are fully established but rather weaker and limited to the western parts of SA with stable conditions over BOB. In conclusion, the TP soil moisture can in uence the ascent and SAH intensity and modulation in zonal and meridional domains that exhibit teleconnections with downstream zonal and meridional wind and modulate the monsoon onset. The ndings agree with previous studies suggesting that TP thermal pro le has a strong in uence on the Asian monsoon by changing the circulation patterns and thus precipitation magnitude (Wu et  Such thermal conditions can affect the SAH intensity, which results in the high/low-pressure centers across southern/northern TP. The SAH variability can trigger vertical ascending/descending motion, leading to cyclonic/anticyclonic activities, and thus modulate the SA monsoon onset.

Summary And Discussion
The Tibetan Plateau thermodynamics are studied and linked with regional and large-scale climate phenomena using empirical and numerical models He et al. 2019). Soil moisture role is least explored in the thermal effects of TP due to limited observations. The current study has used remotely sensed soil moisture during 1988-2008 to study soil moisture role in the Monsoon onset over South Asia (SA). The Monsoon onset over SA is usually associated with an abrupt change in precipitation, wind shear, and tropospheric heat maxima, on which the monsoon onset indices are generally based. In this study, the monsoon onset indices, which represent large-scale monsoon onset dynamics, are used (Wang et al. 2001;Xavier et al. 2007). The South Asian High (SAH) index is considered for its sensitivity to TP thermal heating and can affect the onset of Asian monsoon (Liu et  A signi cant positive correlation was observed from the results for spring soil moisture (April, May, and June) and SAH index; furthermore, both SAH and soil moisture exhibited a signi cant negative correlation with WB and XV monsoon onset indices. The underlying mechanism between the monsoon onset and SAH and TP soil moisture is explored for early-onset (1990, 1999, 2002, and 2004) and late-onset (1992, 1995, 1997, and 2003) composites, respectively. In the early-onset composite, the spring soil moisture has shown obvious positive anomalies in the eastern TP and vice versa in late-onset years. In response to soil moisture anomalies, a consistent positive/negative latent/sensible heat anomaly is persistent in the early-/late-onset composite, which infers that soil moisture has a substantial energy ux role partitioning.
The ndings are consistent with previous studies that near-surface energy partitioning is sensitive to soil ). In the late-onset years, the diabatic heating is rather weaker and possibly implies that dry surface conditions may favor dry adiabatic processes and hence a decrease in the upper tropospheric heating.
The proposed mechanism may be attributed to multiple processes and may not only depend on soil moisture; however, further studies with model experiments will better delineate the underlying mechanism. In a previous study, similar local feedback between soil moisture precipitation is reported with a strong correlation between eastern plateau spring soil moisture and TP monsoon precipitation involving a similar mechanism observed in the current study (Zhou et al. 2018).
The latent heat of condensation over eastern TP enhances the diabatic heating and leads to strong vertical ascending motion ( onset is associated with changes in wind shears and heat maxima from the southern hemisphere towards the northern hemisphere. During the early onset pentad, the SAH move towards the east, and during the onset pentad moves towards the northeast; such zonal and meridional movement of the SAH are previously reported as an indication of the monsoon onset initiation in BOB proceeded by a continental increase in precipitation (Liu et al. 2013). The plateau's divergent motion (Ullah et al. 2020) appeared to link the soil moisture as thermal heat source that triggers above normal diabatic heating existence, leading to intensi ed ascent and early onset, which is weaker in the later onset composite. The 850 hPa wind components transition from prevailing easterlies during the pre-onset pentad to westerlies in the onset-pentad accompanied by enhanced convection and ascent is showing the teleconnections of the plateau soil moisture-induced thermal pro le with regional circulations in the SCS, western Paci c, and Indian ocean.
In conclusion, this study has shown an empirical linkage between TP soil moisture anomalies and SA monsoon onset through thermal processes. The thermal processes associated with soil moisture anomalies affect the SAH variability and, consequently, the SA monsoon onset. The monsoon onset and physical processes are subjected to annual and decadal variability, including external forcing and internal forcing such as oceanic forcing, including the ENSO and IOD in uence (Wang et al. 2017a). The intraseasonal oscillations of air-land-sea interactions and their coupling with Asian monsoon are still unclear and require further studies (Liu et al. 2013). However, such observed linkages are empirical, and further numerical experiments are expected to improve our understanding of TP soil moisture's role in modulating the SA monsoon onset.