Orographic mechanical and surface thermal effects of the Tibetan–Iranian Plateau on extratropical intraseasonal waves in boreal summer: numerical experiments

The intensity and location of boreal summer extratropical intraseasonal oscillations along the subtropical westerly jet (EISO-SJ) are crucial in triggering and distributing extreme events over Eurasia. Based on numerical experiments, this study distinguishes the orographic mechanical and surface thermal forcing of the Tibetan–Iranian Plateau (i.e. TIP-MF and TIP-TF) on EISO-SJ. The TIP-MF primarily modulates the amplitude of EISO-SJ that strengthens over the upstream and weakens over the downstream. Comparatively, the TIP-TF not only reduces/increases the intensity of EISO-SJ over the TIP upstream/downstream, but also significantly migrates the track of EISO-SJ northward. Further analysis demonstrates that the changes of the westerly jet, eddy energy propagation and energy conversion are consistent with the track and amplitude changes of EISO-SJ. This study indicates the variations of the TIP surface sensible heating in interannual variation and global change, as well as the terrain uplift of the TIP in paleoclimate influence on the mid-latitude subseasonal variation.


Introduction
The atmospheric intraseasonal oscillations prevail over extratropical Eurasia in boreal summer (Yang et al 2015, Hannachi et al 2017, Stan et al 2017, Zhu et al 2023, which primarily feature a zonal quasi-biweekly wave train with eastward propagation along the subtropical westerly jet (SJ) (Fujinami and Yasunari 2004, Yang et al 2014. Numerous studies have identified the importance of extratropical intraseasonal oscillations along the SJ (EISO-SJ) on causing regional subseasonal variations and frequently triggering extreme meteorological events, such as heatwaves (Schubert et al 2011, Gao et al 2018 and flooding (Dugam et al 2009, Li et al 2021. Meanwhile, EISO-SJ has been proven to significantly affect the local subseasonal prediction skills (Qi and Yang 2019, Liu et al 2020, Yan et al 2021Yan et al , 2022 and even provides a window of opportunity for East Asian subseasonal prediction (Zhu et al 2023). Therefore, understanding the factors affecting the variations and features of EISO-SJ is crucial for the subseasonal community.
The Tibetan Plateau has remarkable impacts on changing atmospheric circulation through thermal (Yeh 1950, Wu and Liu 2000, 2016, Wu et al 2012 and mechanical effects (Boos andKuang 2010, Park et al 2012). In terms of the thermal forcing of the Tibetan Plateau, numerical studies found that the warming Tibetan Plateau causes the increasing trend of summer frontal rainfall in the East Asian region through exciting two Rossby wave trains over the upper-level westerly jet stream and the low-level southwesterly monsoon (Wang et al 2008). Additionally, the thermal forcing of the Tibetan Plateau modulates the decadal (Duan et al 2011, Wang and Li 2019 and interannual variations (Ueda and Yasunari 1998, Bansod et al 2003, Hsu and Liu 2003, Ullah et al 2021 of the atmospheric circulation and Eurasian temperature/precipitation. In the intraseasonal timescale, Zhu and Guan (1997) found that the anomalous surface sensible heat fluxes of the Tibetan Plateau has a significant effect on the intensity and propagation speeds of extratropical atmospheric intraseasonal oscillations using numerical experiments; however, which only resulted from a single case analysis using a simple two-layer atmospheric circulation model. Liu et al (2007) found that the thermal forcing of the Tibetan Plateau produces the atmospheric quasibiweekly oscillation over the Tibetan Plateau using a global primitive theoretical model, which were also reflected from an individual case. As for the mechanical forcing of the Tibetan Plateau, previous studies mainly focused on its effects on the formation of the Asian summer monsoon systems (Boos and Kuang 2010, Cane 2010, Park et al 2012. Yang et al (2019), (2020) addressed that the mechanical forcing of the Tibetan Plateau facilitates the northward propagation of tropical boreal summer intraseasonal oscillations. The Iranian Plateau, although its area and altitude are smaller than those of the Tibetan Plateau, also has significant impacts on atmospheric circulation through thermal (Zhang et al 2002) and mechanical effects (Zarrin et al 2011). Liu et al (2017 have emphasized that the Tibetan Plateau and Iranian Plateau are not only geographically adjacent but also have mutual influences and feedbacks. Therefore, numerous studies regard the Tibetan-Iranian Plateau (TIP) as a whole in their research (Wu et al 2012, Zhou et al 2016, He et al 2019. TIP is located at the pathway of EISO-SJ, and its thermal and mechanical effects on EISO-SJ have not been clarified yet. Recently, the Global Monsoons Model Intercomparison Project (GMMIP), one of the endorsed MIPs in the Coupled Model Intercomparison Project Phase 6 (CMIP6), was launched, aiming to understand the behavior of monsoon circulations as well as the TIP's thermal and mechanical effects on monsoon variations (Zhou et al 2016, He et al 2019. Based on this research project, the Chinese Academy of Sciences (CAS) Flexible Global Ocean-Atmosphere-Land System (FGOALS-f3-L) (He et al 2019, 2020) and the First Institute of Oceanography-Earth System Model 2.0 (Song et al 2020) climate system models finished a series of experiments with and without TIP's terrain height and surface sensible heating, but only the FGOALS-f3-L provides daily circulation data output. Therefore, this study borrows it to investigate the thermal and mechanical effects of the TIP on EISO-SJ. This paper is organized as follows: section 2 presents the data, numerical experiment and method. Section 3 shows the simulated performance of EISO-SJ in CAS FGOALS-f3-L, and the orographic mechanical and surface thermal effects of TIP on EISO-SJ are displayed in section 4. Section 5 discusses the underlying physical mechanisms. Conclusions are provided in section 6.

Data, numerical experiment and method
Observed daily atmospheric circulation fields are retrieved from ERA-Interim provided by the European Centre for Medium-Range Weather Forecasts (Dee et al 2011), with a 1.5 • × 1.5 • horizontal resolution. The historical record is the period between 1980 and 2014.
The global atmospheric general circulation model used in this study is the CAS FGOALS-f3-L, which was developed at the Institute of Atmospheric Physics/State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics. Three sets of experiments are carried out, named the AMIP, GMMIP amip-TIP (hereafter GMMIP-TIPnoTH) and GMMIP amip-TIPnosh (hereafter GMMIP-TIPnoSH), respectively. Their detailed configurations are introduced in the description paper of the AMIP and GMMIP datasets (He et al 2019, 2020), and a brief summarization is shown in table 1. In general, the GMMIP-TIPnoTH and GMMIP-TIPnoSH have the similar configurations to the AMIP, except that the TIP's terrain height is removed in the former experiment by setting the topography above 500 m to 500 m over the TIP region. The TIP's surface sensible heating is removed in the latter experiment by cutting off the sensible heating over the same TIP region as GMMIP-TIPnoTH. Therefore, the AMIP can be treated as the control run, and the GMMIPs as the sensitive runs. Unless otherwise specified, the TIP's mechanical forcing (TIP-MF) mentioned in this study refers to the orographic blocking of TIP, and the TIP's thermal forcing (TIP-TF) refers to the surface sensible heating of TIP. The selected period of the experimental data is consistent to the ERA-Interim , and the simulated data has the raw horizontal resolution of C96 (about 1.0 • × 1.0 • ). To be more comparable with the ERA-Interim, a bilinear interpolation scheme is used to interpolate the raw data to the regular 1.5 • longitude-latitude grid.
The quasi-biweekly component of a particular variable can be obtained by the following two steps: I) subtracting the climatological mean and the first three harmonics, and II) using the Butterworth bandpass (8-25 d in this study) filter. Empirical orthogonal function (EOF) is used to extract EISO-SJ. In detail, EISO-SJ is retrieved by regressing the quasi-biweekly 250 hPa meridional wind (V250) onto the first principal component, obtained by making the EOF analysis for the boreal summer quasi-biweekly V250 over the SJ region (25-55 • N, 15 • W-130 • E), which refers  Zhu et al (2023) in more detail. Note that compared to Zhu et al (2023), only the EOF1 is analyzed and discussed in this study, because the EOF2 exhibits similar sensitive changes with EOF1 as a response to the numerical experiments (figure ignored). A t-test is used in this study to determine whether the regression coefficient and the differences between AMIP and GMMIPs are significant. It should be noted that the filtered data is often not independent, the effective degrees of freedom for significance tests are estimated referring to Bretherton (1999). Wave activity flux is considered to investigate the energy propagation and dispersion (Takaya and Nakamura 2001), and can be computed as where W represents the horizontal wave-activity flux, U is the wind velocity, u and v are the zonal and meridional winds respectively, ψ denotes the stream function, and a bar and a prime are the summer basic state and quasi-biweekly component. Barotropic energy conversion (CK) and baroclinic energy conversion (CP) are calculated to probe the underlying physical mechanisms and the formulas are as follows (Xu et al 2020): where f is the Coriolis parameter; p is the pressure; and S is the static stability, which is defined as ( , in which C p is the specific heat and R is the gas constant of dry air. An overbar denotes the summer mean state, and a prime is the quasibiweekly component. A positive value of CK/CP represents energy conversion from the mean flow to the quasi-biweekly perturbation by barotropic/baroclinic processes (Kosaka and Nakamura 2006, 2010).

Simulated EISO-SJ in CAS FGOALS-f3-L
Before lending insight into the thermal and mechanical effects of TIP on EISO-SJ, it is necessary to examine the ability of the AMIP experiment in CAS FGOALS-f3-L to reproduce the characteristics of the SJ and EISO-SJ. Figure 1(a) shows the observed zonal wind (U) in the upper troposphere (i.e. U250), which exhibits that the main body of the Eurasian SJ is located along the latitudes of 30-50 • N, in which the SJ axis is about 35-40 • N and two Eurasian SJ cores lie near the Caspian Sea and to the north of the Tibetan Plateau. The primary features of the SJ are well simulated by the AMIP, except that the SJ slightly shifts northward and the intensity of the SJ core around the Caspian Sea is slightly underestimated ( figure 1(b)). Figures 1(c)-(f) display the features of EISO-SJ in observation and AMIP experiment, respectively. Compared with the observation, the AMIP-simulated EISO-SJ well exhibits the location of the positive and negative anomalous centers along the wave train i.e. the centers of the anomalous northerly are over the central North Atlantic, central Mediterranean, Caspian Sea-Lake Balkhash and East Asia, while southerly anomalies are over the south of Great Britain, Black Sea-Caspian Sea and northeast of the Tibetan Plateau (figures 1(c) and (d)). Moreover, the phase velocity speed (∼3.3 m s -1 ), group velocity speed (25.1 m s -1 ) and wavelength (∼4400 km) of EISO-SJ shown in observation can be well simulated in the AMIP experiment (figures 1(e) and (f)). Note that for the AMIP-simulated EISO-SJ, the European part is weaker than the observed one, while the Asian part is stronger than that in observation, which may be related to the biases of the simulated SJ in location and intensity (Lu et al 2002). Overall, the CAS FGOALS-f3-L shows reasonable performance in simulating the observed climatological features of the SJ and EISO-SJ, which is a reliable tool to investigate the thermal and mechanical effects of TIP on EISO-SJ.

Distinguishing the orographic mechanical and surface thermal effects of TIP on EISO-SJ
Figures 2(a)-(c) show the features of EISO-SJ in the AMIP, GMMIP-TIPnoTH and GMMIP-TIPnoSH, respectively. When the TIP-MF is removed, the main body of EISO-SJ exists and the location of EISO-SJ almost remains unchanged compared with the AMIP-simulated one, in which the pathway axes of EISO-SJ are both lied around the latitude of 40 • N. However, the relative intensities of each anomalous center are changed (figure 2(a) vs. (b)), that is, the anomalous amplitudes to the west of 70 • E are weakened, while the anomalous centers to the east of 70 • E are enhanced. In detail, the anomalous centers over the south of Great Britain, central Mediterranean and Black Sea-Caspian Sea weaken, and the anomalous center over the central North Atlantic even disappears. However, the anomalous centers over the Caspian Sea-Lake Balkhash, northeast of the Tibetan Plateau and East Asia strengthen, and a new significant anomaly is generated over Northwest Pacific ( figure 2(b)). Here, we simply divided the upstream and downstream regions of the TIP by using 70 • E as a boundary, i.e. the domain of 25-55 • N, 40 • W-70 • E is defined as the TIP upstream, while the domain of 25-55 • N, 70 • E-180 • E is defined as the TIP downstream. In order to quantify the EISO-SJ's differences over the TIP upstream and downstream between the results with and without the TIP-MF, we calculated the strength of EISO-SJ over the TIP upstream/downstream using the TIPupstream-averaged/TIP-downstream-averaged variance of quasi-biweekly V250, and named it as EISO-SJ-U/EISO-SJ-D. As a result, EISO-SJ-U decreases by 18.6% (|20.1 − 24.7|/24.7 ≈ 18.6%) while EISO-SJ-D increases by 28.7% (|34.1 − 26.5|/26.5 ≈ 28.7%) with the removal of the TIP-MF ( figure 2(d)).
In contrast, EISO-SJ exhibits more significant changes with the removal of the TIP-TF, as shown in figure 2(c). Compared with the AMIP-simulated EISO-SJ, the pathway axis of EISO-SJ simulated in the GMMIP-TIPnoSH experiment is located around the latitudes of 30-35 • N, which shifts southward by about 10 • of latitudes over the Eurasian continent ( figure 2(a) vs. (c)). Meanwhile, the anomalous centers over the TIP upstream strengthen, while the anomalous centers over the TIP downstream weaken, in which EISO-SJ-U are increased by 28.3% (|31.7 − 24.7|/24.7 ≈ 28.3%) while EISO-SJ-D is decreased by 9.4% (|24.0 − 26.5|/26.5 ≈ 9.4%) when the TIP-TF is removed ( figure 2(e)).
The comparison between the AMIP and GMMIPs demonstrates that both the TIP-MF and TIP-TF could significantly modulate EISO-SJ but have different effects. The TIP-MF strengthens the EISO-SJ's amplitudes over the TIP upstream but weakens its TIPdownstream intensities. However, the TIP-TF exhibits the opposite impacts on regulating the amplitudes over the TIP upstream and downstream along EISO-SJ. Meanwhile, the TIP-TF forces the pathway axis of EISO-SJ to significantly migrate northward over the Eurasian continent.

Change of SJ and causes
The SJ is the dominant atmospheric waveguide whose location traps the propagation track of transient waves (Branstator 2002, Wirth et al 2018. Therefore, we first explore the changes in the SJ's location. According to the numerical experiments, the SJ's location remains almost unchanged with the removal of the TIP-MF ( figure 1(b) vs. 3(c)). In contrast, the SJ's location evidently shifts southward after removing the TIP-TF, (figure 1(b) vs. 3(d)). To quantitatively depict the changes in the SJ's location, the westerly jet axis index is calculated, which is defined as the latitude of maximum U250 in the meridional direction between 20 • N and 55 • N (Xiao and Zhang 2013). We restrict the latitudes between 20 • N and 55 • N to exclude the potential identification of the polar front jet at higher latitudes. According to this definition, the zonally averaged (15 • W and 130 • E) westerly jet axis is 40.2 • N in the AMIP experiment, 40.9 • N in the GMMIP-TIPnoTH experiment, and 32.8 • N in the GMMIP-TIPnoSH experiment. The change in the SJ's location are highly consistent with the change in the path of EISO-SJ.
To further explore the reason for the changes in the SJ's location, figure 3 shows the change of the zonal-averaged (30-100 • E) temperature with pressure and upper tropospheric (250 hPa) geopotential height (GHT250) between the GMMIPs and AMIP. On the one hand, as the TIP-MF is removed but TIP-TF remains, the temperature around the TIP obviously increases ( figure 3(a)), which indicates that the meridional thermal gradient is evidently decreased in the subtropics. The reduced meridional temperature gradient causes the northward shift of the SJ (Sha et al 2020). On the other hand, lacking TIP's mechanical blocking means the disappearance of both the topographic bifurcation on SJ (Ding 1994) and the increasing pressure in the windward slope of TIP and over TP (Wu and Liu 2016), just corresponding to the anomalous low pressure over TIP upstream and TIP area (figure 3(c)). The above two aspects work in the opposite way, so that the SJ's location almost maintains similar latitudes. In contrast, when the TIP-TF is removed but the TIP-MF remains, the temperature around the TIP obviously decreases ( figure 3(b)), causing an increased meridional thermal gradient from tropical to subtropical, which forces the SJ to move southward (Nan et al 2021). Meanwhile, the removal of the TIP-TF means the disappearance of the sensible-heat-driven air pump, resulting in anomalous low pressure over the TIP area and its downstream  ( figure 3(d)). The combined effects of the two, lead to a significant southward shift in the SJ's location. Figures 4(a) and (b) display the 250 hPa wave activity fluxes in the GMMIP-TIPnoTH and GMMIP-TIPnoSH experiments, respectively. The wave activity flux is the common method used to describe the energy propagation and dispersion of Rossby waves. As the TIP-MF is removed but TIP-TF is kept, more wave activity fluxes propagate eastward toward TIP downstream without the TIP's blocking effects (Rhines 2007, White et al 2018, which strengthens the TIP-downstream EISO-SJ ( figure 4(a)). However, when only the TIP-TF is removed, the eastward propagating wave activity fluxes are significantly reduced over the TIP downstream due to the disappearance of the TIP's heating (Liu et al 2007), so that the intraseasonal waves are evidently weakened over the TIP downstream ( figure 4(b)). The eddy kinetic energy (EKE) further quantifies this process (figure 4(c)). The EKE increases by 30.0% (|33.4 − 25.7|/25.7 ≈ 30.0%) over the TIP downstream with the removal of the TIP-MF while decreases by 15.2% (|21.8 − 25.7|/25.7 ≈ 15.2%) over the TIP downstream without the TIP-TF.

Change of eddy energy
In addition, previous works have suggested that extratropical intraseasonal Rossby waves can develop efficiently by harvesting perturbed energy from basic flow via baroclinic (Kosaka et al 2009, Chen et al 2013, Xu et al 2020 and barotropic processes (Wang et al 2013, Zhu andYang 2021). Therefore, we calculate the baroclinic energy conversion (denoted as CP) and barotropic energy conversion (denoted as CK) over the TIP upstream and downstream in the GMMIP-TIPnoTH and GMMIP-TIPnoSH experiments, respectively, as shown in figures 4(d) and (e). With the removal of the TIP-MF (figure 4(d)), CP is decreased by 22.8% (|1.42 − 1.84|/1.84 ≈ 22.8%) over the TIP upstream, while increased by 131.9% (|3.85 − 1.66|/1.66 ≈ 131.9%) over the TIP downstream. The increased/decreased positive CP indicates that less/more time-mean available potential energy is converted to perturbed available potential energy over the TIP upstream/downstream to develop the   Though CK is about an order of magnitude smaller than CP in both GMMIPs, similar with CP, with the removal of the TIP-MF, CK is decreased by 56.5% (|0.10 − 0.23|/0.23 ≈ 56.5%) over the TIP upstream, while increased by 50.0% (|0.18 − 0.12|/0.12 ≈ 50.0%) over the TIP downstream. Conversely, with the removal of the TIP-TF, CK is increased by 60.9% (|0.37 − 0.23|/0.23 ≈ 56.5%) over the TIP upstream, while decreased by 25.0% (|0.09 − 0.12|/0.12 ≈ 25.0%) over the TIP downstream. The increased/decreased positive CK indicates that less/more time-mean kinetic energy is converted to EISO-SJ.

Conclusions and outlooks
Using the state-of-the-art CAS FGOALS-f3-L model, this study distinguished the orographic mechanical and surface thermal effects of TIP on EISO-SJ and discussed the underlying physical mechanisms. When the TIP-MF is removed, EISO-SJ weakens over the TIP upstream but strengthens over the TIP downstream. In contrast, EISO-SJ exhibits more significant changes when the TIP-TF is absent. That is, the intensity of EISO-SJ is enhanced over the TIP upstream, while the amplitude is reduced over the TIP downstream. Meanwhile, the core pathway of EISO-SJ shifts southward with the removal of the TIP-TF. Further analysis indicated that the change in the SJ's location migrates the track of EISO-SJ while the change in the eddy energy modulates the intensity of EISO-SJ. When the TIP-MF is removed but TIP-TF is kept, more wave activity fluxes propagate eastward toward TIP downstream without the TIP's blocking effects, accompanied by a weaker/stronger positive energy conversion over the TIP upstream/downstream, which favors the enhancement of EISO-SJ over the TIP downstream and the weakening over the TIP upstream. In contrast, when the TIP-TF is removed, a southward shift in the SJ's location forces the track of EISO-SJ to move southward. Meanwhile, the intraseasonal perturbations are weakened downstream due to the disappearance of the TIP's heating, accompanied by a stronger/weaker positive energy conversion over the TIP upstream/downstream, which favors the enhancement of TIP-upstream EISO-SJ and the attenuation of the TIP-downstream EISO-SJ.
This study has identified that the TIP-TF more significantly affects the intensity and location of EISO-SJ. The evident changes of the TIP-TF often occur on seasonal (Yanai et al 1992), interannual (Hsu and Liu 2003) and decadal (Wang and Li 2019) time scale. In particular, a weak decreasing trend (Wu et al 2015, Wang andZhao 2020) has been reported in recent decades. According to this study, the changes of the TIP-TF may modulate the intensity and location of EISO-SJ, and subsequently change the distribution and probability of EISO-SJ-related extremes.
This study also indicates that the variations of TIP surface sensible heating in interannual variation and global change as well as the terrain uplift of TIP in paleoclimate modulate the mid-latitude atmospheric subseasonal waves.