Interdecadal Variation of Spring Extreme High-Temperature Events in the Western Tianshan Mountains and Its Relationship with the Tropical SST

,


Introduction
Te Intergovernmental Panel on Climate Change's Sixth Assessment Report points out that extreme events, such as heatwaves, droughts, heavy precipitation, and tropical cyclones, have increased [1]. In recent decades, extreme high temperatures have had a major impact on human health, agriculture, ecological environment, and socioeconomic zones [2,3]. For example, European heatwaves in 2003 and 2010 killed tens of thousands of people [4,5], and in 2021, the super heatwave event in western North America killed hundreds of people and caused forest fres [6]. On the one hand, spring's extreme high temperature afects the growth time of spring wheat and reduces the yield, negatively afecting agricultural production [7]. On the other hand, high temperature may intensify snow and ice melting in the source regions of rivers, causing food disasters such as snowmelt (ice) foods and ice dam breaks [8,9].
As changes in extreme climate/weather events have greater impacts on human life than the variation of the climate mean state, it is of great signifcance to investigate the variations of extreme events. According to previous research studies, high temperatures have been reported to rise on an interdecadal basis in eastern China [2,10], northeastern China [11], southern China [12], and northwestern China [13], but the increasing rates are diferent. In northern and southwestern China, the frequency of extreme high-temperature events (EHEs) is increasing quickly, while it increases slowly in southern China [11]. Te Xinjiang Uygur autonomous region, located in northwest China, also shows a noticeable increase in high-temperature events [14]. Te variation trend of the EHEs can change the pattern of snowmelt foods in Xinjiang [8,15]. Terefore, it is of great social signifcance to focus on the spatiotemporal variations of spring extreme high temperatures in Xinjiang and understand their internal mechanisms.
Te occurrence of EHEs is closely related to atmospheric circulations, and atmospheric circulation anomalies are a direct cause of weather and climate changes [16]. However, the circulation characteristics that cause the EHEs are different in diferent regions of China [17]. In Northeast China, positive geopotential height anomaly/anticyclonic circulation anomaly in the middle and upper troposphere is closely associated with the EHEs. Anticyclonic circulation anomalies afect atmospheric descending motion and reduce cloud cover, leading to adiabatic heating, increasing solar radiation, and thus increasing surface temperature [16]. From central to southern China, the lower troposphere's temperature advection has an impact on the EHEs as well [18]. In eastern and southern China, the EHEs are afected by the western Pacifc subtropical high (WPSH) anomaly [19,20]. Te atmospheric circulation anomalies that contribute to the EHEs in China are related to several elements, including the tropical Pacifc sea surface temperature (SST) [18,21], El Niño-southern oscillation [22,23], Pacifc El Niño-like pattern [24], Atlantic SST [25], tropical Indian Ocean SST [26], and sea ice [27,28]. SST anomalies in the tropical Indian Ocean enhance latent heat to excite the Rossby wave train, resulting in the anomalies of the South Asian high and WPSH, and thus afecting the summer EHEs in eastern China [23]. Te East Asia-Pacifc/Pacifc-Japan teleconnection may cause anticyclonic anomalies and atmospheric sinking in southern China as a result of the maritime continent's abnormal warming, favoring EHEs there [12].
Previous studies on the EHEs in Xinjiang, Northwest China, mostly focused on the spatial distribution characteristics and temporal variation trends of the EHEs [8,[29][30][31], but less research has been conducted on the atmospheric circulation anomalies of the EHEs and the related mechanisms. A few studies indicate that the anomalous variations of the Iran high and South Asia high are of guiding signifcance for the evolution of high temperature processes in Xinjiang [32,33].
Te western Tianshan Mountains in Xinjiang, China, namely, the confuence of the Yili River in central Xinjiang and the Tarim River in southern Xinjiang, are an essential agricultural and animal husbandry production area in Xinjiang [34,35]. Terefore, paying attention to the spring EHEs in the western Tianshan area is of great signifcance for food prevention and mitigation, as well as for agricultural production and life. In recent decades, the interdecadal variability of the summer climate in the western Tianshan Mountains has been widely identifed. In contrast, little is known about the interdecadal variability and related mechanisms of the spring EHEs. Terefore, this research is mainly concerned with the following two questions: What is the interdecadal variability of the EHEs and the associated atmospheric circulation characteristics in the western Tianshan Mountains? What are the mechanisms of the interdecadal variability of the EHEs in the western Tianshan Mountains and the possible infuencing factors?.

Data.
Te data used in this study consist of three parts. Te daily average temperature at 824 stations provided by the National Climate Center, China Meteorological Administration, is used to analyze the spatial distribution and interdecadal variation of the EHEs, widely used to investigate climate characteristics in China. Te monthly (daily) reanalysis data from the National Centers for Environmental Prediction/National Center for Atmospheric Research include the cloud-forcing net solar radiation fux and net surface heat fux, net surface shortwave and longwave radiation, net surface sensible heat fux and surface latent heat net fux, with a horizontal resolution (Gaussian grid) of 92 × 192 (94 × 192). In addition, for variables, i.e., geopotential height, sea level pressure, air temperature, horizontal wind, vertical velocity, relative humidity, and specifc humidity, the horizontal resolution of these data is 2.5°× 2.5°, and the vertical levels from 1000 hPa to 100 hPa are 11, namely, 1000 hPa, 925 hPa, 850 hPa, 700 hPa, 600 hPa, 500 hPa, 400 hPa, 300 hPa, 250 hPa, 200 hPa, 150 hPa, and 100 hPa. Analysis of the surface heat fux anomalies and atmospheric circulation anomalies related to the EHEs is done using the aforementioned variables from the reanalysis data. Te monthly average SSTprovided by the Hadley Center of the UK Meteorological Ofce, with a horizontal resolution of 1.0°× 1.0°, is used to analyze the possible relationship between the SST anomalies and the EHEs. Te spring period in this study is from March to May, and all data span the period 1979-2019.

Defnition of Extreme High-Temperature Events (EHEs).
In this study, the relative threshold method is adopted to defne the EHEs at single stations, i.e., using a percentage threshold method to determine the standard of the EHEs. Tis method can take into account the diferent climatic conditions in diferent regions and has been used in many previous studies [12,25,36]. Specifcally, the temperature series from 1979 to 2019 at each station are extracted, and the 90th percentile is selected as the threshold. If the temperature values at a station exceed this threshold, they are recorded as the EHEs at this station. Te western Tianshan Mountains are defned within the area of 37°N-44°N, 74°E-82°E. A regional EHE is defned based on the regional average of the EHEs at stations in the western Tianshan Mountains.

Statistical Methods.
Te comprehensive analysis is used in this research to investigate the anomalies of circulations and the SST in diferent periods, and the Student's t-test at the 95%/99% confdence level is performed to estimate the signifcance of these anomalies. Te sliding t-test method is used to examine interdecadal variability. Te vertical integration of whole-layer water vapor fux can be calculated by the following equation: where g denotes the acceleration of gravity, q the specifc humidity, V the horizontal wind vector, and P s the surface pressure. P s is set to 300 hPa. Te atmospheric water vapor content can be expressed by Q � p s p qdp.

Interdecadal Variations of the Frequency of the Spring
EHEs. Figure 1 Figure 1(e) demonstrates that the variation trend of EHE frequency at almost all stations had a mutation around 1996. Hence, in order to more thoroughly analyze the interdecadal variation of the EHEs and explore the associated mechanisms, the period of 1979-1995 is considered the inactive period (P1), 2000-2015 as the active period (P2), and 1995-2000 as the transition period.

Circulation Anomalies Related to the Interdecadal
Variation of the Spring EHEs

Surface Heat Flux.
Surface thermodynamic efciency is impacted by the surface heat fux, which has the greatest infuence on surface temperature [37]. Both the net surface latent heat fux ( Figure 2(d)) and the net surface long-wave radiation ( Figure 2(b)) exhibit negative anomalies that do not contribute to the net surface heat fux's positive anomaly (Figure 2(a)). Te above results indicate that the net surface heat fux in the western Tianshan Mountains is mainly afected by the surface shortwave radiation and sensible heat fux. Te reasons for the increase in short-wave radiation and sensible heat fux are discussed further in the following sections.

Water Vapor Conditions and Circulations.
Sufcient water vapor and atmospheric ascending motion facilitate the increase of clouds, which can reduce the solar radiation reaching the surface and afect the surface temperature. Figure 3 presents the composite felds of the 850 hPa vertical velocity, 850 hPa relative humidity, and net solar fux forced by clouds between the P2 and P1 periods. Te descending motion ( Figure 3(a)) and the decreased water vapor (Figure 3(b)) reduce the cloud cover, resulting in the positive anomaly of cloud-forced net solar radiation (Figure 3(c)). Terefore, the surface short-wave radiation and sensible heat increase (Figures 2(c) and 2(e)), and the surface heat fux increases (Figure 2(a)), which is conducive to the increase in the surface temperature and the occurrence of the EHEs. Te causes of atmospheric dryness and ascending motion in the western Tianshan Mountains are analyzed in the following sections. Te net surface latent heat fux in the western Tianshan Mountains shows a negative anomaly (Figure 2(d)), indicating that there is more upward surface latent heat fux, i.e., more water vapor evaporation. Terefore, the dry atmosphere in this area (Figure 3(b)) may not be caused by the local water vapor evaporation but by the water vapor transport. Vertically integrated water vapor fux and water vapor fux divergence between the P2 and P1 periods are shown as composite felds in Figure 4, respectively. Te results indicate a westward path of water vapor transport (Figure 4(a)). However, water vapor is not transported to the western Tianshan Mountains, and this region is in a water vapor divergence area, which may lead to a decrease in atmospheric water vapor content in the western Tianshan Mountains (Figure 4(b)).
Te anomalous water vapor transport over the western Tianshan Mountains is closely associated with large-scale circulation anomalies. Figure 5(a) depicts the composite feld of the 500 hPa geopotential height diference between P2 and P1 periods, which is characterized by an anomalous high pressure over the Iranian Plateau to western China and resembles the distribution pattern of vertically integrated water vapor transport. At 100 hPa ( Figure 5(b)), there is also an anomalous anticyclone over the Iranian plateau. Te water vapor and atmospheric circulations in the western Tianshan Mountains are afected by this deep anomalous high pressure system ( Figure 5(b)).

Advances in Meteorology
Te anomalous Iranian high is located in the subtropical high belt (Figures 5(a) and 5(b)), with a strong descending motion in the troposphere (Figure 6(a)). Te diference in meridional circulation over this region between the P2 and P1 periods shows an anomalous ascending motion near the equator and an anomalous descending motion near the subtropics (Figure 6(b)), which implies a strengthening of the Hadley circulation. Te Iranian Plateau is located in the descending branch of the Hadley circulation, which means that the abnormal Iranian high is mainly dynamic. Te warming efect of upper-layer air sinking is the main reason for maintaining the high pressure center [38].

Relationship between the EHEs and Anomalous SST.
Air-sea interactions signifcantly impact atmospheric circulation [39,40]. Te discussion continues on whether the tropical and temperate oceans afect the anomalous Iranian high system. Te regression of the SST onto the EHE index was calculated to investigate the potential infuence of anomalous SST (Figure 7). Tese fndings indicate that EHEs are highly connected to SST in the Atlantic, Indian, and Pacifc oceans. In addition, the key oceanic areas related to the Iranian high are similar, except that the Indian Ocean shows signifcant positive anomalies. However, the warming of the Indian Ocean is weaker than that of the Pacifc Ocean, and there is still a zonal temperature gradient ( Figure S1). In particular, the Pacifc Ocean is characterized by negative anomalies over the tropical eastern and northern Pacifc and positive anomalies over the tropical western Pacifc, corresponding to the PDO's negative phase (Figure 7). Additionally, positive anomalies over the eastern and northern Atlantic Oceans correspond to the positive phase of the AMO (Figure 7). Terefore, we determined the correlation between the ECEs and ocean indices to confrm their       Figure 4: Diference in the spring average (a) and water vapor transport (kg·m −1 ·s −1 ) and water vapor transport divergence (10 −5 kg·m −2 ·s −1 ) (b) whole-layer water vapor content (kg·m −2 ) between P2 and P1 periods.  Figure 5: Diference in the spring average (a) 500 hPa and (b) 100 hPa geopotential height (train the) between P2 and P1 periods. Te dots represent the values passing the signifcance test at the 95% confdence level. 6 Advances in Meteorology

Te Infuence of Indo-Pacifc Anomalous SST.
Te diferences in SST between P2 and P1 were calculated to investigate the interdecadal variation in the Indo-Pacifc SST. Consistent with the frequent occurrence of EHEs, there was anomalous warming over the western Pacifc to the northern Pacifc and anomalous cooling over the tropical eastern Pacifc, both of which corresponded with the negative phase of the PDO (Figure 8(c)), except in the Indo-Pacifc sector. Te annual mean SST composite (Figure 8(b)) and the PDO annual mean SST composite (Figure 8(d)) over the Pacifc were similar to those of the spring SST (Figures 8(a) and 8(c)). Accompanied by anomalous warming of SST over the western Pacifc (Figure 9(a)), there was a signifcant increase in upward long-wave radiation (Figure 9(b)); thus, wind felds converged at 700 hPa in the lower troposphere (Figure 9(c)) and diverged at 200 hPa in the upper troposphere (Figure 9(d)). In contrast, the anomalous warming in the western Indian Ocean was weaker than that in the western Pacifc (Figure 9(a)), resulting in weaker upward long-wave radiation (Figure 9(b)). Terefore, the wind      Advances in Meteorology converged at 200 hPa in the upper troposphere (Figure 9(d)) and diverged at 700 hPa in the lower troposphere (Figure 9(c)). Based on mass compensation, it can be considered that the troposphere over the tropical western Pacifc is controlled by anomalous ascending motion and the troposphere over the tropical western Indian Ocean is infuenced by anomalous descending motion (Figure 9(e)), suggesting a strengthened Walker circulation. Moreover, the lower branch of the Walker circulation at 700 hPa was characterized by westerly anomalies from the western Indian Ocean to the western Pacifc (Figure 9(f )), and the upper branch (200 hPa) was characterized by strong easterly anomalies (Figure 9(g)). Te upper branch of the Walker circulation strengthened the easterly wind south of the anomalous Iranian high, resulting in pressure strengthening. In summary, the zonal thermal gradient between the tropical western Pacifc and the western Indian Ocean directly afected the zonal vertical circulation and strengthened the anomalous Iranian high [41].

Te Infuence of Atlantic Anomalous SST.
We further focused on the accompanying variations in Atlantic SST to study other factors causing interdecadal variations in anomalous Iranian highs. Figure 10 shows the composite SST in the Atlantic between P2 and P1. Consistent with the frequent occurrence of EHEs, there was anomalous warming over the North Atlantic (Figure 10(a)), corresponding to the positive phase of the AMO (Figure 10(c)). Te annual mean SST composite (Figure 10(b)) and AMO's annual mean SST Accompanied by the anomalous warming of SST over the Atlantic (Figure 11(a)), there was a signifcant increase in upward long-wave radiation (Figure 11(b)), inducing anomalous ascension in the overlying atmosphere ( Figure 11(c)). Owing to the anomalous ascent, precipitation increased ( Figure 11(d)), mainly convective precipitation (Figure 11(e)). Terefore, the latent heat increased (Figure 11(f)) and heated the atmosphere, favoring the persistence of a negative potential height anomaly over the North Atlantic (Figure 11(g)). Te interaction between air and sea over the North Atlantic triggered a mid-latitude wave train (Figure 11), strengthening the anomalous Iranian high.
In addition, the oceanic front may also afect the negative geopotential height anomaly over the Atlantic Ocean. Te change in the oceanic front caused by SST anomalies afects the atmospheric transient eddy [40]. Te anomalous warming pattern over the Atlantic (Figure S2(a)) resulted in positive anomalies in meridional SST gradients and sensible heat fuxes over the northern Atlantic (Figures S2(b) and S2(c)), thus enhancing the temperate oceanic front. Te lowlevel atmospheric meridional temperature gradient and, thus, the low-level atmospheric baroclinicity tended to be reinforced when the atmospheric boundary layer adjusted to the strengthened oceanic front (Figures S2(d) and S2(e)), favoring the persistence of a potential negative height anomaly over the Atlantic (fgure 11(g)). , and (g) 500 hPa geopotential height (shading; gpm) and the corresponding wave activity fux (vectors; unit: m 2 ·s −2 ) between P2 and P1. Te grids and dots represent the values passing the signifcance test at the 99% confdence level and the 95% confdence level based on the Student's t-test, respectively.

Conclusions
In this study, the spatial distribution characteristics and interdecadal variation of the EHEs in the western Tianshan Mountain from 1979 to 2019 are analyzed. Te possible causes afecting the interdecadal variation of the EHEs are investigated, and the possible relationship between the tropical SST anomaly and the EHEs is discussed. Te main conclusions are as follows: (1) Te spring EHEs in the western Tianshan Mountains showed interdecadal variability in the mid-1990s, characterized by less in 1983-1996 and more in 2000-2015. In addition, the spatial distribution of the EHEs has local characteristics, with a high occurrence frequency in the area along the Tarim River. Under the control of the anomalous high pressure system, the atmospheric water vapor content is lower over the western Tianshan area, and this region is controlled by atmospheric descending motion. On the one hand, with the descending motion, the adiabatic heating heats the atmosphere in the lower layers. On the other hand, dry atmospheric conditions with descending motion are not conducive to the increase of cloud cover, and thus, the net solar radiation forced by clouds increases correspondingly. In addition, the short-wave radiation fux and sensible heat fux reaching the surface increase, and the net surface heat fux increases cumulatively. Tese conditions are conducive to the increase of the surface temperature and the occurrence of EHEs. (3) Te comprehensive analysis of the SST suggests that the zonal SST diference between the tropical western Pacifc and the western Indian Ocean directly strengthens Walker circulation. Te western tropical Pacifc Ocean is controlled by atmospheric ascending motion, and the western tropical Indian Ocean is controlled by descending motion. Te circulations in the lower troposphere are dominated by westerly anomalies, and those in the upper troposphere are dominated by easterly anomalies. Te upper troposphere's easterly anomalies strengthen the easterly airfow south of the anomalous Iranian high, thereby strengthening the high pressure. (4) Te anomalous Iranian high is also associated with the atmospheric wave trains at midlatitude, which are related to diferent air-sea interaction in the Atlantic. Te anomalous warming of Atlantic SST inducing convective precipitation was increased and the oceanic front was enhanced. Te anomalous convective precipitation induces more latent heat release, which are led to negative potential height anomaly over Atlantic and trigger a midlatitude wave train. In addition, the enhanced oceanic front induces the atmospheric baroclinicity stronger and synoptic transient eddy activities frequently, which favors the persistence of a negative potential height anomaly over the North Atlantic ( Figure 12).
In this study, we discuss the possible efects of SST anomalies in the tropical Pacifc and Indian Oceans on the local high pressure over the Iranian Plateau. Previous studies have also mentioned that Indo-West Pacifc warming was one of the factors afecting the interdecadal variations of atmospheric circulations in the late 1970s [42]. Additionally, the changes in circulations and water vapor distribution caused by the phase transition of the Pacifc Decadal Oscillation further afect the EHEs in Asia. Te Atlantic SST anomalies excite the mid-latitude wave train through air-sea igure 12: A schematic diagram illustrating the mechanism through which key oceanic areas infuence the EHEs in the western Tianshan mountains by a mid-latitude wave train and walker circulation.
interactions, thereby infuencing the Eurasian atmospheric circulations and extreme temperature and precipitation events [43]. In addition, recent studies have increasingly suggested Pacifc-Atlantic interaction since the early 1990s [44]. Te spring Atlantic Ocean actively infuences the winter Pacifc climate and ENSO variability by infuencing subtropical teleconnections on the intertropical convergence zone (ITCZ), similar to a discharging capacitor [45]. Te possibility of a global warming trend interfering with the prior positive phase of the AMO (during the 1930s-1960s) could explain why the previous positive phase of the AMO (during the 1930s-1960s) was not accompanied by heightened Atlantic-Pacifc interactions [45]. Further investigating the relationship between these factors and EHEs can contribute to a comprehensive understanding of the formation mechanisms of the EHEs over the western Tianshan Mountains.
In terms of the diagnostic analysis in this study, the seasonal mean data are used to determine the relationship between oceanic and atmospheric anomalies. Tis relationship can be understood as an equilibrium state due to the interactions among these felds. Comparative analyses of multi-source data and model simulation would make the results more robust, which will be carried out in the future.

Data Availability
Te NCEP/NCAR reanalysis data is accessible at https://psl. noaa.gov/data/gridded/data.ncep.reanalysis.html. Te monthly SST data derived from the Met Ofce Hadley Center can be downloaded online at https://www.metofce. gov.uk/hadobs/hadisst/. Other data used to support the fndings of this study are available from the corresponding author upon request.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this article.