Differences in The Causes of Light And Extreme Precipitation For The Qaidam Basin In Summer


 The Qaidam Basin (QB) locates over the northeast of the Tibetan Plateau (TP), where precipitation especially extreme precipitation possesses obvious local characteristics compared with that over the whole TP. This study tries to investigate cause of light (50% threshold) and extreme (95% threshold) precipitation in boreal summer in the QB, which is helpful to deepen understanding of the mechanism of precipitation formation in different regions of the TP. The extreme (light) precipitation thresholds in the eastern QB are greater than that in the western QB, with a value of 6~16mm (2mm) for most regions. There are two main moisture transport channels for light and extreme precipitation events. One is from the Eurasia and carried by the westerlies, which provides 48.2% and 55.8% of moisture for light and extreme precipitation events, respectively. The other moisture transport channel is from the Arabian Sea and the Bay of Bengal, which is transported toward the QB at the joint role of the South Asian summer monsoon and the plateau monsoon, contributing 51.8% and 44.2% of moisture for light and extreme precipitation events, respectively. The stronger moisture transport to precipitation mostly attributes to the enhanced moisture influxes from the western and southern boundaries. Additionally, the weaker moisture outflux across the eastern boundary is also responsible for the extreme precipitation. The circulation characteristics shows that, the precipitation in the QB has a closely relationship with the weak ridge over the Caspian Sea and Aral Sea, the enhanced South Asian summer monsoon and plateau monsoon, which are conducive to the moisture transport from the Eurasia and low-latitudes toward the QB. The meridional circulation enhances, meantime the westerly jet stream splits into east- and west-branch, and the south Asian high (SAH) strengthens, which are beneficial for the stronger convective motion. Especially, the trough in the northwest of the QB and the more significant east- and west-branch structure of westerly jet are the main circulation characteristics for the extreme precipitation events. Further analysis reveals that the apparent heat source over the QB is contributed to more synchronous moisture transport around the TP and its surrounding areas for light precipitation events, while the apparent heat source enhances 1 day prior to moisture transport from the east part region of the South Asian summer monsoon to around the eastern TP for extreme precipitation events. Meantime, the apparent heat source triggers an abnormal cyclone over the TP which can positively strength the local convective motion. Such abnormal configuration of atmospheric circulation and the influence of apparent heat source can explain the difference in cause of precipitation with different magnitude to a great extent in the QB.


Introduction
The Qaidam Basin (QB), as the largest cold arid region with high elevation in China, is located at the northeast of the Tibetan Plateau (TP) and the northwest China (shown in Fig. 1a), far from the ocean, with mainly arid climate. As with rich Salt Lake mineral resources, the QB is known as "Cornucopia", an important regional energy strategy in China, and is classified as the first batch of national circular economy experimental areas. Under the background of the climate change, the QB has experienced two periods with relatively cold period before 1987 and relatively warm period after 1987 (Shen et al.,2016). The annual mean temperature, maximum temperature, and minimum temperature have increased in the past, with a higher increased trend than that in the TP, and even the global. Associated with global warming, there are more frequent, more intense climate extremes (Huang et al., 2021), so as in the QB the warming trend can also affects the local hydrological recycle. In addition, lag behind warming, it has happened that the climate changes from warm dry to warm wet as at the beginning of 21 st century. The precipitation and precipitation days increase, and the increased magnitude in the eastern QB is larger than that in the western QB (Shen et al., 2016). Zhang et al.,2019). Based on theory and numerical models, it is claimed that the heavy precipitation increases more than total precipitation (Shiu et al., 2012;Trenberth et al., 2003), which implicates that there must be a decrease in relatively weaker precipitation events (Trenberth et al., 2003). That is, the heavy precipitation increases at the cost of light precipitation. The trends in the intensity of light precipitation events are opposite to the frequency of light precipitation events, which is possibly associated with light precipitation shifts toward heavier precipitation, making the frequency of light To summarize, despite some progresses have been made on the changes of precipitation, and corresponding moisture transport and atmospheric circulation over the TP, such work for the QB solely have not been systematically studied before.
Especially, as the regional characteristics of the precipitation in different parts of the TP, it is necessary to focus on the light precipitation and extreme precipitation in the QB, respectively. According to the CN05.1 precipitation data , we investigate the change in the precipitation of the whole TP and the QB. The mean annual precipitation values of the QB and TP are 197.13 mm·yr -1 and 409.80 mm·yr -1 during 1981-2017, respectively. The annual precipitation time series also shows different change trends with a strong increased trend of 9.51 mm·decade -1 significant at the 0.01 level over the QB and an insignificantly upward trend of 1.14 mm·decade -1 over the TP, and the correlation coefficient of precipitation between these two regions is 0.228 (Fig.1b). These results suggest that the QB and TP show completely different precipitation variation characteristics, highlighting the significance to study the regional characteristics for the precipitation in the TB.
Hence, focusing on the characteristics and cause of precipitation in the QB is not only of greatly scientific and practical significance for both natural and managed ecosystems in such arid region, but also beneficial to further understanding of the physical mechanism for regional precipitation over the TP. The annual precipitation in the QB is usually less than 220 mm ( Supplementary Fig. 1a), with summer precipitation accounting for more than 70% of annual precipitation in western QB and less than 60% of annual precipitation in eastern QB (Fig. 1c), which indicates the majority of precipitation of the QB occurs in boreal summer. Thereby, this study focuses on the boreal summer season (June, July and August) for the precipitation. The purpose of this study is to address and discuss the following questions: (1) What are the main causes in the atmospheric circulation inducing the different types of precipitation in the QB?
(2) How does atmospheric heat source influence moisture transport and convective activity in light and extreme precipitation events? Addressing such issues is beneficial for deep understanding of the physical mechanism of summer precipitation as well as their prediction in arid area of China.
The remainder of this paper is organized as follows: The data and methods are introduced in section 2. The characteristic of light and extreme precipitation over the QB and the corresponding moisture transport are described in section 3. The atmospheric circulation cause and response of precipitation to atmospheric heat source are analyzed in detail in section 4 and section 5, respectively. Finally, summary and discussion are illustrated in section 6.

Data
A daily precipitation dataset of ten meteorology stations during 1981-2017 in the QB is used to analyze the characteristics of precipitation, which is obtained from local Weather Bureau. Fig. 1c shows the specific geographic location of these meteorology stations, including Mang Ya, Leng Hu, Xiao Zaohuo, Da Chaidan, Ge Ermu, Nuo Muhong, De Lingha, Du Lan, Wu Lan, and Cha Ka. Ten stations are evenly distributed, indicating the data is representative of weather and climate in the QB. Additionally, to test the difference of precipitation between in the QB and TP, the daily China regional grid observation precipitation data set-CN05.1 is used, with a horizontal resolution of 0.25º latitude by 0.25° longitude. This data is based on the observation data of more than 2,400 stations in China and established by interpolation using the "anomaly approximation" method, which has been widely used in modeling test and precipitation analysis (Hsu et al., 2016).Finally, the reanalysis data, including geopotential height (GPH), relative humidity, and wind

The definition of light and extreme precipitation events
According to the definition of extreme precipitation (Zhai et al., 2003), the day with daily precipitation equal to or greater than 0.1 mm is regarded as a precipitation day.
For each meteorological station, the precipitation of all precipitation days is arranged in ascending order and the 95 th percentile value of the precipitation sequence is then taken as the extreme precipitation threshold. Similarly, the 50 th percentile value of the precipitation sequence is taken as the light precipitation threshold .
When the daily precipitation of at least 1/3 continuous stations of the total meteorological stations exceeds the extreme (light) precipitation threshold of the station, the extreme (light) precipitation event is considered. According to this criteria, we picked out 27 extreme precipitation events which all last only one day and 281 light precipitation events including 231 events, 47 events, 2 events and l events lasting 1day, 2 days, 3 days and 4 days in boreal summer in the QB, respectively.

Moisture transport and contribution
To evaluate the important impacts of moisture transport on the QB precipitation, the vertically integrated horizontal moisture flux was calculated following by Where q , g , p s , V ⃑ ⃑ , and Q represent specific humidity, gravity acceleration, surface pressure, horizontal wind, and moisture flux, respectively. The vertically integrated of the moisture flux was preformed from surface pressure to 300hPa due to the air above 300hPa is highly dry and contributes little to the moisture transport process (Sun et al., 2014). Besides, the horizontal moisture flux divergence integrated from the surface to 300hPa can be expressed as In general, the net moisture flux for target region is most determined by the sum of We also use the HYSPLIT_4 model to quantitatively evaluate the contribution of each moisture transport to precipitation in the QB, and the detailed description of the According to the specific humidity of all trajectory endpoints, we can calculate the moisture contribution of each cluster as Q s is the moisture contribution, q last is the specific humidity of trajectory endpoint, m is the number of trajectories in the cluster (henceforth moisture channel), and n is the total number of trajectories.

Apparent heat source and apparent moisture sink
Following Yanai et al. (1973), the atmospheric apparent heat source ( Q 1 ) and apparent moisture sink (Q 2 ) can be expressed as where c p denotes the specific heat at constant pressure, T is the air temperature, t is the time, V ⃑ ⃑ is the horizontal wind velocity, ω is the vertical velocity, p is the pressure,θ is the potential temperature, p 0 = 1000 hPa, q is the specific humidity, L is the latent heat of condensation, and κ = R C p with R is the gas constant. Here Q 1 denotes the total diabatic heating, including radiation, latent heat and surface heat flux, and subgrid-scale heat flux convergence; Q 2 , on the other hand, represents the latent heat due to condensation or evaporation processes and sub grid-scale moisture flux convergence (Yanai et al., 1973). The Q 1 and Q 2 integrated from 1000hPa to 10hPa are the whole column apparent heat source (< Q 1 >) and apparent moisture sink (< Q 2 >).

Quantitative indices of atmospheric circulation systems affecting precipitation
The evolution of plateau monsoon is essential to the weather and climate processes U850-U200, averaged over south Asian from the equator to 20ºN and from 40ºE to 110ºE (henceforth WYI). The SAHI is defined as the weighted sum of the GPH of all grid points which is greater than 12520gpm in the northern hemisphere at 200hPa . Here the SAHI represents not only the intensity but also the area controlled by the SAH.
We also calculated westerly jet index (WJI) and westerly circulation index (WCI) in order to investigate how do the westerly affect the precipitation in the QB. The WJI is defined as zonal wind averaged over 35ºN-45ºN,50ºE-105ºE at 200hPa (Liao et al., 2018), and the WCI refers to the difference of GPH at 500hPa between 35ºN and 55º N (Rossby, 1939). In addition, the index of low pressure system intensity (henceforth LPSI) in the QB was considered, which is vorticity averaged over 35ºN-40ºN, 90ºE-100ºE at 700hPa. precipitation. It is shown that total light precipitation accounts for 8%~ 22% of summer precipitation while total extreme precipitation accounts for 2%~14% of summer precipitation, which indicates both light precipitation and extreme precipitation are crucial for summer precipitation in the QB. The rainfall decreases from the western to eastern QB for light precipitation events, and decreases from the central QB to the western and eastern QB for extreme precipitation events, which shows the physical processes causing two types precipitation events are different. It is obvious that the values of the ratio of total light precipitation to summer precipitation are greater than that of the ratio of total extreme precipitation to summer precipitation in most region except for the south-central QB, indicating the light precipitation contributes more to summer precipitation in the QB, even arid region of China. As the precipitation occurs infrequently in the QB and total precipitation is greatly less than that in eastern China, the ability to cope with extreme precipitation in the QB is weaker than that in eastern China. Once extreme precipitation happens, more serious damage will be brought to the QB, highlighting significance of more attention to extreme precipitation here. Light precipitation refers to the precipitation with relatively small magnitude, accounting for nearly one-fifth of total summer precipitation in the QB, which could reflect the basic precipitation characteristic in a region to some extent.

Spatial distribution of precipitation
The light and extreme precipitation thresholds for each meteorological station, shown in Fig Fig. 1b), indicating that the region with higher precipitation in summer is also the area where extreme precipitation occurs with higher possibility.

Moisture transport
To investigate the relationship between moisture transport and precipitation in the QB, we show the vertically integrated moisture flux and moisture flux divergence of two types of precipitation events, as shown in Fig. 3. For light precipitation events, there are two main moisture channels. One moisture channel is from the black Sea, Caspian Sea and Aral Sea and then turns southeastward to affect the study region, which is the main moisture transport affecting precipitation in the QB. Besides, part of the moisture originating from these region turns southward and combines with the moisture from the Indian Ocean, which is the south branch moisture channel affecting the QB and carrying moisture from Arabian Sea and the Bay of Bengal northward subsequently.
But this moisture channel is slight and only affects the southern QB (Fig. 3a). Thereby, we can conclude that the moisture transport affecting the QB mostly come from the west which is most influenced by the westerlies, which can also be found in Yao et al. In fact, as the endpoint of eastern channels all could be lie in the Eurasia even if we calculate more than 14-day backward trajectory for these two types event, it is feasible to classify the eastern channels to the Eurasia channel. Thereby, two moisture channels could be concluded for precipitation events in general, i.e., the Eurasia channel (called as the north branch moisture channel before) and the Arabian Sea-the Bay of Bengal channel (called as the south branch moisture channel before  To further evaluate the relative role of low-, middle-and upper-level moisture on the QB precipitation, the vertical structure of moisture transport was computed, as shown in Fig.5. We divided the whole air column into low (from surface pressure to700hPa), middle (700hPa-500hPa), and upper (500hPa-300hPa) layers. Because the altitude of the QB is above 2600 m, as illustrated in Fig. 1a, there is no moisture transport across southern boundary at low layer.
The main moisture influxes to the QB for summer climate state are from the western and northern boundaries, and the main moisture outflux is from the eastern boundary (Fig.5a). The moisture transport across western boundary is mainly concentrated in the upper layer and decreases from upper level to the surface. As the altitude decrease to 700hPa, the eastward moisture transport turns back to the westward. The moisture transport across northern boundary is concentrated in middle-and low-level, with the largest value of southward moisture in low-level. In the middle layer, the southward moisture transport is less than that in the low layer, while in the upper layer, the southward moisture is highly weak due to low humidity and still maintains southward direction. It can be seen that the moisture transport across the southern boundary is little in middle and upper layer, which is much smaller in magnitude than that across the other three boundaries because of the barrier role of the TP, and the northward moisture transport turns back to the southward as the altitude increases. The moisture transport across eastern boundary is eastward and increases from low level to upper level.
For light precipitation events, the main moisture influxes are also from the western and northern boundaries. The moisture transport influx across southern boundary is small, while it is significantly stronger compared with the summer climate state, and the main moisture outflux is still from the eastern boundary (Fig.5b). There are four aspects illustrating the main characteristics about the moisture transport across the western, northern, southern and eastern boundaries. Firstly, the moisture transport across western boundary is concentrated in the upper layer and decreases from upper level to the surface. Secondly, the moisture transport across northern boundary is concentrated in middle-and low-level, and the moisture transport in the upper layer is northward instead of southward which is opposite to that of the middle and low layers.
Thirdly, the moisture transport from the southern boundary increases from middle to upper layer and remains smaller in magnitude than that from the other three boundaries.
Finally, the moisture transport across eastern boundary is eastward and increases from low level to upper level, which is consistent with summer climate state but with larger values.
As is shown in Fig.5c, the main moisture influx (outflux) for extreme precipitation events is from western (eastern) boundary. Compared with that in summer climate state, the moisture transport from northern boundary is obviously weaker, while the moisture transport from southern boundary is significantly stronger. The moisture transport from western boundary still increases from low layer to upper layer and the total moisture transport is 98.04 kg.s -1 , which is larger compared with 85.12 kg.s -1 for light precipitation events. In low and middle layers, the magnitude of moisture transport is larger than that of light precipitation events, while the situation is opposite at upper level. The total moisture influx from northern boundary is 45.14 kg.s -1 , which is less than that light precipitation events with 86.45 kg.s -1 , because the northward (southern) moisture transport in upper (middle) level is far larger (less) than that light precipitation events, and the southward moisture transport in low level is slightly larger that light precipitation events. And the moisture transport from the southern boundary increases in upper layer compared with light precipitation events. Besides, the moisture transport across eastern boundary is westward in middle and low levels which is different from that in summer climate state and light precipitation events, while is eastward at upper level, and the total eastward moisture transport is 52.84 kg.s -1 .
From the above analysis we can conclude that the stronger moisture transport from western and southern boundaries, and may significantly contribute to the occurrence of kg.s -1 for summer climate state, light precipitation events and extreme precipitation events, respectively. As a result, the total net moisture influx of extreme precipitation events is 262.64% (197.23%) of that of summer climate state (light precipitation events).

Convective activity
Besides the strengthened net moisture inputs, proper meteorological conditions are needed to facilitate the moist air to be precipitated out. In this section, we analyze convective activity when precipitation occurs in the QB. The vertical structure of ω and horizontal wind divergence averaged over 35ºN-40ºN, 90ºE-100ºE for summer climate state and two types precipitation events are shown in Fig. 6.
It is observed that in summer the ω is less than zero from low level to upper level in the QB, indicating the upward motion dominates here which is a direct result of the thermal effect over the TP (Ye et al., 1957). Compared with summer climate state, the upward motion is stronger when precipitation especially extreme precipitation happens in the QB. In addition, for two types precipitation events, there is obvious horizontal wind convergence below 550hPa and horizontal wind divergence above 550hPa over the QB, which is not obvious in summer climate state. The difference of the intensity of horizontal wind divergence from surface to upper level between light and extreme precipitation events indicates the magnitude of net moisture influx is different. The stronger of the horizontal wind convergence in low level and divergence in upper level, the more moisture is converged and lifted in the QB.

Atmospheric circulation characteristics
To further understand the causes of the moisture transport and the precipitation of two types precipitation events, we investigated the atmosphere circulation that are associated with the summer precipitation in the QB (Fig. 7). As shown in Fig.7a, the light precipitation is closely related to a positive GPH anomaly over the west Siberian and a negative GPH anomaly over the QB and the adjacent area at 500hPa.
Consequently, there is anomalous pressure gradient over mid-high latitudes inducing anomalous northerly wind southward that can affect the QB. Additionally, a weak ridge lying over the Caspian Sea and Aral Sea, makes the northwesterly flow prevailing at northwest of the QB, which is consistent with the moisture transport from the northwest of the QB (Fig. 3b). There are a positive GPH anomaly which extends northeastward from the eastern of Iranian plateau to the Northeast China Plain and a negative GPH anomaly which extends from the north of the QB to the northern of the Bay of Bengal during the extreme precipitation period (Fig. 7c). The intensity of GPH anomaly is stronger compared with that of light precipitation events, which leads to a stronger pressure gradient anomaly over mid-high latitudes, then induces stronger anomalous northerly wind affecting the QB. The intensity of the weak ridge over the Caspian Sea and Aral Sea enhances, and a new trough appears in the northwest of the QB, inducing more strong northwest flow enters into the study region. Besides, it is observed that the trough over the Bay of Bengal is stronger than that in light precipitation events. As a result, there is more moisture from the Arabian sea and the Bay of Bengal transported northward. From Fig. 7d, it is shown the body of the SAH is similar to that in Fig. 7b,  Fig. 7a and Fig. 7c, we also find the plateau monsoon is important during the precipitation process, i.e., the greater the plateau monsoon, the stronger the precipitation in the QB (Fig. 7e).

Quantitative analysis of major atmospheric circulation system
Based on above results, it is shown that plateau monsoon, SAH, westerly jet and the anomalous low pressure over the QB, are very important for the precipitation in the QB. although the center of the SAH is divided into two parts and the maximum value of the center of SAH is 12540gpm (Fig. 7d), the intensity of the SAH is stronger compared with that in light precipitation. The value of LSPI is 2.75, 2.37 and 3.35 for summer climate state, light precipitation events and extreme precipitation events, respectively, which denotes there is usually a cyclonic circulation exits in the QB, consistent with results in Fig.6. The value of LSPI of light precipitation events is very similar to the climate state, with an insignificantly weaker intensity, indicating that strong LSPI is not the necessary factor for light precipitation in the QB. But for the extreme precipitation events, strong LSPI is needed, revealing this cyclonic circulation is facilitating to the precipitation in the northeastern QB. The value of WJI is 1.66, 1.89 and 0.48 for summer climate state and two types events, respectively, indicating the westerly jet weakens when extreme precipitation occurs. However, from the Fig. 7b(d), it is shown the westerly jet splits into east-and west-branch during the period of precipitation especially extreme precipitation in the QB, indicating that the westerly jet can affect precipitation in the QB by its pattern instead of its intensity.
The structure of east-and west-branch of the westerly jet is the most prominent in precipitation events especially extreme precipitation events. And the center of the SAH divides into two parts during the period of extreme precipitation (Fig.7b(d)). In fig. 7ef, there are positive and negative GPH anomalies ranging from the Iranian plateau to the Baikal Sea and the east of Mediterranean to the eastern Europe plateau, respectively.
The difference of the SAH, westerly jet and GPH indicate the meridional circulation of extreme precipitation events is more obvious than that of light precipitation events. As a result, to simply analysis the intensity of meridional (zonal) circulation, the WCI was computed and normalized. Here, a large and positive value indicates that the westerly wind is strong, corresponding to the zonal circulation, while a negative and large absolute value indicates that the westerly wind is weak, corresponding to the meridional circulation.
As shown in Fig. 8b, the values of WCI show a pattern as extreme precipitation events>light precipitation events>summer climate state between 30ºE and 75ºE. The WCI are below zero for summer climate state and light precipitation events, while is above zero from 50ºE to 60ºE and is below zeros from 35ºE to 50ºE, and from 60ºE to 75ºE for extreme precipitation events. Also, the WCI show another pattern as summer climate state>light precipitation events>extreme precipitation events, and is below zero from 79ºE to 110ºE (presented in green box). Therefore, the meridional change of the values of WCI indicates that the zonal circulation from 30ºE to 75ºE and the meridional circulation from 79ºE to 110ºE of extreme precipitation events are the strongest compared with summer climate state and light precipitation events. The strong zonal circulation over the Europe leads to more moisture eastward. Then the strong meridional circulation lying the QB and its upstream region makes this eastward

The heat source pattern
The diabatic heating over the TP plays an important role in the atmospheric circulation over the TP and its adjacent areas (Zhao and Chen, 2001), as well as the precipitation in China (Xu et al., 2013). The QB is located at the northeastern TP, are there any relationship between summer precipitation here and the heat source? This issue is discussed in this section. with the enhanced precipitation which can be seen from Fig. 9d and Fig. 9f.
It also can be seen that the changes of < Q 1 > and < Q 2 > over the QB are more obvious than that of the whole TP, suggesting the pressure-meridional cross section of apparent heat source (Q 1 ) and apparent moisture sink (Q 2 ) and their anomalies averaged between 35ºN and 40ºN, as shown in Fig. 10. It can be seen that the apparent heat source and apparent moisture sink extends from surface to the tropopause and are more remarkable near the surface than that on other layers, especially in the eastern QB.
Another evident fact is that the intensity of < Q 1 > and < Q 2 > during extreme precipitation is larger than that in summer climate state and light precipitation events Furthermore, the deep convection is closely related with < Q 1 > and < Q 2 > over the eastern TP. Greater < Q 1 > and < Q 2 >, stronger upward motion presented over the eastern TP . It is evident that the heat source over the TP affects the atmospheric circulation associate with the local vertical motion and the moisture source. So, the strengthened atmospheric heat source over the QB may be closely related to the summer precipitation especially extreme precipitation here, which will be discussed in the next section.

Relationship between moisture transport and the apparent heat source
To have a complete description of the processes of atmospheric heating affects precipitation, it is also necessary to know the relationship between the moisture transport and heat source over the QB. Besides, in order to highlight the effect of the local heat source, we calculated the cross correlation between the apparent heat source (< Q 1 >) over the QB and the whole column moisture flux of two types events (Fig.   11). Because < Q 1 > includes < Q 2 >, radiative heating and heat sink due to vertical transport of dry-static energy and latent heat (Yanai and Tomita, 1998), we concentrate only on the effect of apparent heat source (< Q 1 >) on the moisture transport here.
For the light precipitation events, there is only the southward correlation vectors in light-blue region in the north of QB on lag -1 day (Fig. 11a). The above across correlation analysis indicates that the < Q 1 > over the QB can affect the moisture transport. The convergence correlation vectors in the TP and its adjacent areas of light precipitation events represent that the moisture flux is "pumped up" by the heat source over the QB (Fig. 11c) only affects the moisture transport in the eastern TP. Additionally, there is also an abnormal cyclone circulation over the TP on lag 0 day stimulated by heat source which facilitates the convergence of moisture and upward motion over the QB.

Relationship between convective activity and the apparent heat source
The above analysis reveals that the apparent heat source over the QB can trigger an abnormal cyclonic circulation over the TP on lag 0 day for both light and extreme precipitation events, and this abnormal cyclone can affect the vertical motion over the QB. Thereby, we calculated the correlation relationship between apparent heat source (< Q 1 >) and ω in order to further investigate the effect of diabatic heat over the QB on convective motion (Fig .12).
As is shown in Fig .12, there is negative correlation between < Q 1 > and ω over the QB passing the 90% significant level for two types events, which indicates the larger atmospheric apparent heat source is corresponding to stronger the upward motion in the QB. Hence, the abnormal cyclonic circulation over the TP stimulated by the apparent heat source contributes to the upward motion over the QB. The absolute value of negative correlation coefficient of light precipitation events is larger than that of extreme precipitation events, indicating the source heat over the QB has a stronger impact on convective activity for the light precipitation. In conclusion, the heat source over the QB can not only affect the moisture transport but also the upward motion in the QB, which influences further the pattern of precipitation in the QB.

Summary and discussion
This study analyzes the characteristics and formation causes for light and extreme precipitation in boreal summer in the QB, which located over the northeast TP and shows regional features in weather and climate. In the QB, the extreme (light) precipitation thresholds in the eastern region are greater than that in the western region, which is highly consistent with distribution of total rainfall frequency and magnitude.
The thresholds of light precipitation events for most regions are less than 2mm. The maximum of extreme precipitation thresholds is over 16 mm in De Lingha station and Du Lan station, and the minimum of the thresholds is less than 6 mm in the centralwestern Basin. Besides, although as a small probability event, extreme precipitation contributes a lot to the total summer precipitation in the QB.
There are two main transport channels for moisture in the OB. One is the north branch moisture channel, with the moisture transporting along which is mainly from the Eurasia and carried by the westerlies, contributing around 48.2% and 55.8% of moisture for light and extreme precipitation events, respectively. The other is the south branch moisture channel, with moisture originating from the Arabian Sea and the Bay of Bengal, and then transported by the joint role of the South Asian summer monsoon and the plateau monsoon toward the QB, which provides 51.8% and 44.2% of moisture for light and extreme precipitation events, respectively. Moisture transport across boundaries of the QB is characterized the summer precipitation over the QB is largely attributed to the increased moisture transport across the western and southern boundaries, especially for the extreme precipitation. The decreased moisture output from the eastern boundary is additionally responsible for the increased extreme precipitation. In specific, compared with the summer climatological state, the moisture transport increases by 14.45 kg.s -1 for light precipitation events and by 27.37 kg.s -1 for extreme precipitation events from the western boundary, and increases by 21.97 kg.s -1 for light precipitation events and by 44.31 kg.s -1 for extreme precipitation events from the southern boundary. The moisture output across the eastern boundary increases by 17.81 kg.s -1 for light precipitation events and decreased by 53.91 kg.s -1 for extreme precipitation events. The convective activity during precipitation over the QB is significant because there are the upward motion and horizontal wind convergence below 550hPa and horizontal wind divergence above 550hPa. Especially, the intensity of convective activity is stronger for the extreme precipitation compared with that during light precipitation.
The light (extreme) precipitation in the QB is associated with the weak ridge over the Caspian Sea and Aral Sea at mid-high latitudes, which drives more westerly moisture and cold air from the Eurasia into the basin. In addition, the enhanced South Asian summer monsoon and plateau monsoon bring more moisture originating from Arabian Sea and the Bay of Bengal even the Eurasia, resulting in more moisture release to the QB. Finally, the strengthened meridional circulation may be responsible to the increase of precipitation in the QB. The meridional circulation enhances, the westerly jet stream splits into east-and west-branch, and the SAH strengthens, causing a divergence environment in the upper level over the QB. While, for the extreme precipitation events, the trough is obvious in the northwest of the QB, the east-and west-branch structure of westerly jet is more significant, and the center of SAH divides into two parts with the main center located over the Iranian plateau.
The apparent heat source is another important factor influencing the formation of precipitation events, by both affecting the moisture transport and reinforcing the local convective activity. For light precipitation events, the apparent heat source over the QB mainly affects the synchronous moisture transport which is from the south of Aral Sea, the north of Indian subcontinent, the northwest of the QB, and the eastern TP. While, for extreme precipitation events, the apparent heat source ahead of one day affects the moisture transport originating from the east part region of the South Asian summer monsoon and the southern of Indian subcontinent.
In this study, we have studied the impact of the apparent heat source over the QB which indicates the total atmospheric diabatic heat on the moisture transport and convective activity associated with precipitation, while the contribution of different component of the atmospheric heat source is unclear. Especially, previous studies (Wu et al., 2007) revealed that the sensible heat over the TP affects the Asian summer monsoon and the precipitation in the Asian. As a part of the TP, the QB is whether influenced by the local sensible heat or not, should be studied further. As an inland basin, the topography of the QB and its surrounding areas are very complex, such as the TP, the Qilian Mountain, the Kunlun Mountains, and the Altun Mountains. The effects of the topography on the local weather and climate, as well as the role of meso and small-scale weather systems, are planned to study further by using numerical model experiment.