Advanced Search
Article Contents

Diagnosis of Moist Vorticity and Moist Divergence for a Heavy Precipitation Event in Southwestern China


doi: 10.1007/s00376-016-6124-9

Get Citation+

Export:  

Share Article

Manuscript History

Manuscript received: 03 May 2016
Manuscript revised: 27 June 2016
Manuscript accepted: 01 August 2016
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Diagnosis of Moist Vorticity and Moist Divergence for a Heavy Precipitation Event in Southwestern China

  • 1. Xichang Satellite Launch Center, Xichang 615000, China
  • 2. State key Laboratory at Numerical Modelling for Atmospheric Sciences and Geophysical Fluid Dynamics, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China
  • 3. College of Meteorology and Oceanography, PLA University of Science and Technology, Nanjing 211101, China

Abstract: A regional heavy precipitation event that occurred over Sichuan Province on 8-9 September 2015 is analyzed based on hourly observed precipitation data obtained from weather stations and NCEP FNL data. Two moist dynamic parameters, i.e., moist vorticity (mζ) and moist divergence (mδ), are used to diagnose this heavy precipitation event. Results show that the topography over southwestern China has a significant impact on the ability of these two parameters to diagnose precipitation. When the impact of topography is weak (i.e., low altitude), mζ cannot exactly depict the location of precipitation in the initial stage of the event. Then, as the precipitation develops, its ability to depict the location improves significantly. In particular, mζ coincides best with the location of precipitation during the peak stage of the event. Besides, the evolution of the mζ center shows high consistency with the evolution of the precipitation center. For mδ, although some false-alarm regions are apparent, it reflects the location of precipitation almost entirely during the precipitation event. However, the mδ center shows inconsistency with the precipitation center. These results suggest that both mζ and mδ have a significant ability to predict the location of precipitation. Moreover, mζ has a stronger ability than mδ in terms of predicting the variability of the precipitation center. However, when the impact of topography is strong (i.e., high altitude), both of these two moist dynamic parameters are unable to depict the location and center of precipitation during the entire precipitation event, suggesting their weak ability to predict precipitation over complex topography.

1. Introduction
  • Southwestern China is in the southeastern Tibetan Plateau, which is characterized by very complex topography (Fig. 1) (Cai, 1996). (Djebou et al., 2014) reported that complex topography has a stabilizing effect on precipitation variability in time and space. For southwestern China, when summer arrives, abundant moisture transport from the Indian Ocean carried by southwesterly flow usually converges over southwestern China (Li et al., 2013). The moisture supply is favorable for heavy precipitation over this region, especially over Sichuan Province. However, under the influence of complex topography, heavy precipitation can usually induce serious natural hazards, such as flash floods and mudslides (Mu, 2006; Fauna, 2010; Jiang, 2011). These natural hazards can result in a variety of societal and economic losses. Therefore, improving the prediction of heavy precipitation is very important to reduce these adverse impacts over southwestern China.

    Actually, many efforts have been made toward improving the understanding and prediction of heavy precipitation over southwestern China in recent years. (Li et al., 2016) pointed out that many previous studies on heavy precipitation over southwestern China have mainly paid attention to the diurnal and interannual variability of heavy precipitation (Yu et al., 2007, 2010; Bai et al., 2011; Shen and Zhang, 2011; Yan et al., 2013), the associated large-scale atmospheric circulation and synoptic system (Xu et al., 2012; Li et al., 2014), and the effect of topography (Ge et al., 2008). The results show that heavy precipitation over southwestern China mainly occurs during July and August, and nocturnal precipitation is prominent over Sichuan Province. Moreover, the western Pacific subtropical high, southwest vortex and plateau vortex have significant impacts on the occurrence and development of heavy precipitation over Sichuan Province. Besides, the steep and complex topography favors the formation of mesoscale vortices over this region, which directly affects the occurrence and development of heavy precipitation. These studies mainly used statistics and synoptic methods to analyze the mechanism of heavy precipitation. On the other hand, it should be mentioned that numerical simulation is also an important method to study heavy precipitation over southwestern China (Wang et al., 2013; Li et al., 2014, 2016; Gao et al., 2016). The above studies have enhanced understanding on heavy precipitation and improved the ability to predict heavy precipitation over southwestern China, to some extent. However, the accuracy of heavy precipitation prediction under the influence of complex topography still needs to be improved, especially regarding the location of heavy precipitation over southwestern China.

    In fact, predicting the location of heavy precipitation has received increased research interest in recent years. Many dynamic parameters have been developed in meteorology to improve the prediction of the location of heavy precipitation, such as helicity (Yang et al., 2010), the Q-vector (Yue and Shou, 2008) and the generalized Ertel-Rossby Invariant (Gao et al., 2012; Zhou et al., 2014). However, a single dynamic parameter cannot depict the location of heavy precipitation exactly, because heavy precipitation is a combination of favorable moisture transport with dynamic conditions; and as such, (Doswell et al., 1996) suggested that combining dynamic and moisture factors together in a forecast method could greatly improve the prediction of heavy precipitation. On this basis, (Gao et al., 2004) developed a generalized potential temperature parameter by introducing the effect of moisture into the conventional potential temperature parameter. A number of thermodynamic parameters have been further developed based on this work, such as the moist thermodynamic parameter, the convective vorticity vector (Gao et al., 2004), the thermodynamic wave-activity density (Gao and Ran, 2009), the thermodynamic potential vorticity wave-activity density (Ran et al., 2009), an advection parameter (Wu et al., 2011), and baroclinic vorticity (Ran et al., 2013). Most existing studies demonstrate that the above dynamic parameters have the ability to depict the location of heavy precipitation to a certain extent, and could be used to forecast heavy precipitation (Ran et al., 2014).

    To improve the correlation between areal coverage of a particular dynamic parameter and the observed location of precipitation, (Qian et al., 2015) proposed a feasible way to introduce the effect of moisture into two conventional dynamic parameters (i.e., vorticity and divergence), following the approach used by (Gao et al., 2004). They demonstrated that the incorporation of the effect of moisture in dynamic parameters can improve the ability of dynamic parameters to depict the location of heavy precipitation. However, they mainly used these two dynamic parameters to diagnose a heavy precipitation event over eastern China, where the land is flat compared with the complex topography over southwestern China. The complexity of topography is known to influence the distribution of atmospheric variables, subsequently affecting the ability of dynamic parameters to diagnose the location of heavy precipitation to some extent (Djebou et al., 2014). Therefore, the ability of these two dynamic parameters to depict the location of heavy precipitation over southwestern China needs to be further explored. Besides, we investigate whether these two dynamic parameters can depict the center of heavy precipitation, to improve their ability to be applied in weather forecasting in the future.

    Following this introduction, section 2 describes the methodology and datasets used in the present study. Section 3 articulates the favorable atmospheric circulation background during this precipitation event. Section 4 examines the ability of two moist dynamic parameters (moist vorticity and moist divergence) to diagnose this precipitation event. And lastly, section 5 summarizes the study and its findings.

    Figure 1.  (a) Distribution of 145 weather stations in Sichuan Province (black dots). (b) 24-h accumulated precipitation in Sichuan Province from 1200 UTC 8 to 1200 UTC 9 September 2015 (contours; units: mm). Contours are shown from 10 to 110 mm, with an interval of 10 mm. Note that the shaded areas in (a) and (b) indicate the topography (unit: m).

2. Dataset and method
  • Two datasets are used in this study. The first one is hourly observed precipitation data obtained from 145 weather stations distributed in Sichuan Province (Fig. 1a), downloaded from the China Meteorological Administration (CMA) (http://data.cma.cn). These stations are relatively densely distributed over eastern Sichuan Province and sparsely distributed over western Sichuan Province, because of the complex topography over western Sichuan Province. The observed precipitation data are interpolated onto a 0.25° latitude-longitude grid using the Cressman method (Cressman, 1959). Note that the actual duration of the precipitation over eastern Sichuan Province is from 1200 UTC 7 to 0000 UTC 11 September 2015. However, the heavy precipitation mainly occurred from 1200 UTC 8 to 1200 UTC 9 September 2015. Figure 1b shows the 24-h accumulated precipitation during this period. The center of precipitation is located over eastern Sichuan Province (30.5°N, 105°E). The second dataset is the NCEP FNL (Final) Operational Global Analysis data, which is used to calculate various diagnostic parameters. This dataset is available online (http://rda.ucar.edu/ datasets/ds083.2), with a horizontal resolution of 1°× 1° and a vertical resolution of 26 levels.

    As we know, relative vorticity [Eq. (2)] and horizontal divergence [Eq. (3)] are two important dynamic parameters in diagnosing atmospheric variability associated with heavy precipitation: \begin{eqnarray} \label{eq1} \zeta&=&\frac{\partial v}{\partial x}-\frac{\partial u}{\partial y} ,(1) \\[-0.5mm] \label{eq2} \delta&=&\frac{\partial u}{\partial x}+\frac{\partial v}{\partial y} , (2)\end{eqnarray} where u is the zonal wind, v is the meridional wind, ζ is the vorticity, and δ is the divergence. Many studies have shown that these two parameters can greatly improve our understanding of precipitation events (Gao et al., 2015; Qian et al., 2015, 2016; Liu et al., 2016). However, a heavy precipitation event requires not only suitable dynamic conditions but also abundant moisture transport, i.e., the dynamic parameters alone are unable to depict such an event accurately. (Qian et al., 2015) pointed out that dynamic parameters cannot always accurately depict the location of precipitation because of a lack of moisture conditions. Therefore, parameters that combine both dynamic and moist effects are needed to depict heavy precipitation more accurately.

    How to properly introduce the effect of moisture into dynamic parameters is an important question. (Gao et al., 2004) first defined a generalized equivalent potential temperature parameter (θ*) by multiplying a condensation probability function (q/qs)k in a formula for equivalent potential temperature: \begin{eqnarray} \label{eq3} &&\theta^\ast=\theta\exp\left[\frac{Lq_s}{c_pT}\left(\frac{q}{q_s}\right)^k\right] ,\ (3) \\ \label{eq4} &&{\rm MPV}=\rho^{-1}\zeta_{\rm a}\nabla\theta , (4) \\ \label{eq5} &&{\rm GMPV}=\rho^{-1}\zeta_{\rm a}\nabla\theta^\ast . (5) \end{eqnarray} Equation (4) shows the generalized equivalent potential temperature. θ and T are the potential temperature and temperature, respectively; L is the latent heat of vaporization; cp is the specific heat of dry air at a constant pressure; q is air specific humidity; qs is air saturated specific humidity; and k is an empirical constant; ρ is the air density; ζ a is the absolute vorticity. When q=qs, air is in a saturated moist situation; whereas when q=0, air is completely dry. They then defined a generalized moist potential vorticity (GMPV) parameter [Eq. (6)] by replacing θ with θ* in the moist potential vorticity [MPV; Eq. (5)]. Many studies have demonstrated that GMPV is superior to MPV with regard to depicting the location of heavy precipitation, because of its more realistic treatment of moisture (Gao et al., 2004; Deng and Gao, 2009; Zhou et al., 2010; Qian et al., 2015).

    Following the method used by (Gao et al., 2004), (Qian et al., 2015) extended conventional vorticity [Eq. (2)] and divergence [Eq. (3)] to moist vorticity [Eq. (7)] and moist divergence [Eq. (8)] by multiplying the condensation probability function: \begin{eqnarray} \label{eq6} m\zeta&=&\left(\frac{\partial v}{\partial x}-\frac{\partial u}{\partial y}\right)\left(\frac{q}{q_s}\right)^k ,(6) \\ \label{eq7} m\delta&=&\left(\frac{\partial u}{\partial x}+\frac{\partial v}{\partial y}\right)\left(\frac{q}{q_s}\right)^k , (7)\end{eqnarray} where k=0,1,2,…, indicating the relative contribution of the effect of moisture. k is an empirical constant. Its choice is arbitrary (Wang et al., 2013; Qian et al., 2015). When k=0, moisture has no impact on these two dynamic parameters, and then and become the original vorticity (ζ) and divergence (δ), respectively. With the increase of k, the impact of moisture will be gradually concentrated in a smaller area, given q/qs≤ 0. (Qian et al., 2015) showed that after the effect of moisture is properly introduced, the ability of dynamic parameters to capture the location of heavy precipitation is significantly improved.

    In general, both ζ and δ are two conventional dynamic parameters that are used to diagnose precipitation. However, precipitation is closely related with both dynamic and moist conditions. Therefore, either dynamic or moisture factors alone cannot accurately depict the area of a heavy rain event (Qian et al., 2015). To depict precipitation more accurately, the moisture effects are included within these two conventional dynamic parameters, forming and . The and reflect both dynamic and moist effects. After the effects of moisture are properly incorporated into ζ and δ, the ability of and to capture the area of a heavy precipitation event is significant improved, compared with ζ and δ (Qian et al., 2015).

    Figure 2.  Distribution of 500 hPa geopotential height (blue line; units: dagpm) and wind vectors (units: m s-1) at (a) 0000 and (b) 1200 UTC 8 September, and (c) 0000 and (d) 1200 UTC 9 September. The bold blue line indicates the geopotential height of 588 dagpm at 500 hPa. The rectangular area represents the location of eastern Sichuan Province.

    In this study, k=20 is used throughout. Besides, we calculate the vertical integration of these two moist dynamic parameters from the surface to 600 hPa, to reveal the relationship between moist dynamic parameters and precipitation more clearly. int [Eq. (9)] and int [Eq. (10)] are used to diagnose a heavy precipitation event over Sichuan Province: \begin{eqnarray} \label{eq8} m\zeta_{\rm int}&=&-\int_{\rm sfc}^{600}{m\zeta dp} ,(8) \\ \label{eq9} m\delta_{\rm int}&=&-\int_{\rm sfc}^{600}{m\delta dp} . (9) \end{eqnarray}

3. The background situation
  • The occurrence and development of precipitation is closely linked to a favorable atmospheric circulation background. Figure 2 shows the evolution of geopotential height and wind vectors at 500 hPa during the precipitation event adopted in the present work. At 0000 UTC 8 September (Fig. 2a), a "two trough and one bridge" situation appears over the mid-high latitudes: the two troughs are mainly located to the east of Lake Barr Kersh and to the southeast of Lake Baikal, and a weak bridge controls the region extending from Sinkiang to Mongolia. Besides, a low pressure system is apparent to the east of Lake Baikal. In the low latitudes, the subtropical high (indicated by thick blue lines) is located over southeastern China between 20° and 28°N. The above distribution of atmospheric circulation forms an "east-high-west-low" situation, which favors the occurrence of precipitation. At 1200 UTC 8 September (Fig. 2b), in the mid-high latitudes, the trough to the east of Lake Barr Kersh and the bridge strengthen to some extent, and the low pressure system to the east of Lake Baikal is stable and motionless. This situation is favorable for cold and dry northwesterly transport from western Lake Baikal to northern China. In the low latitudes, the subtropical high strengthens significantly, favoring warm and moist southeasterly transport from the South China Sea to Sichuan Province. Between then and the next time of observation (0000 UTC 9 September), the above synoptic systems strengthen persistently, leading to increased rainfall. At 1200 UTC 9 September (Fig. 2d), the synoptic systems over the mid-high latitudes weaken and move eastward. Meanwhile, the subtropical high controls the whole of southern China. These distributions weaken the convergence between northwesterly and southeasterly flow over eastern Sichuan Province, decreasing the amount of precipitation.

    Figure 3.  Distribution of 850 hPa wind vectors (units: m s-1) at (a) 0000 and (b) 1200 UTC 8 September, and (c) 0000 and (d) 1200 UTC 9 September. The rectangular area represents the location of eastern Sichuan Province.

    At 850 hPa, before the occurrence of heavy precipitation (Fig. 3a), significant warm and moist southeasterly flow extends from South China to eastern Sichuan Province under the influence of the subtropical high. Besides, an anticyclone is apparent over central Mongolia, causing cold and dry northerly transport from Mongolia to Gansu Province. However, it should be mentioned that both the southeasterly and northerly flow are weak and do not converge over Sichuan Province. With the increase in the intensity of the synoptic system over the mid-high levels, the warm and moist southeasterly flow, and cold and dry northerly flow, strengthen significantly and converge over eastern Sichuan Province (Fig. 3b), resulting in the occurrence of precipitation. At the next observation time (Fig. 3c), with the convergence of southeasterly and northerly flow, wind shear appears over eastern Sichuan Province, which is able to increase the amount of precipitation significantly. At 1200 UTC 9 September (Fig. 3d), both the southeasterly and northerly flows weaken significantly, corresponding to the variability of the synoptic systems in the mid-high levels, subsequently resulting in decreased precipitation over eastern Sichuan Province.

    Figure 4 shows the evolution of vertically integrated (from the surface to 300 hPa) moisture flux and its divergence during this precipitation event. The vertically integrated moisture flux is calculated according to the equation \begin{equation} \label{eq10} Q=-\frac{1}{g}\int_{\rm sfc}^{300}(qV)dp , (10) \end{equation} where g is gravity, q is specific humidity, V is the wind field, sfc is surface pressure, and p is pressure. Before the occurrence of this heavy precipitation event (Fig. 4a), the vector of the moisture flux shows moisture transport from the Bay of Bengal to Sichuan Province. Besides, eastern Sichuan Province is associated with moisture convergence, which provides favorable moisture conditions for this precipitation event. When the precipitation occurs (Fig. 4b), the moisture transport is mainly from the South China Sea, and the convergence of moisture becomes significant, which favors the formation of precipitation. At the next observation time (Fig. 4c), both the transport and convergence of moisture strengthen remarkably, corresponding to increased precipitation. With the weakening of moisture transport and convergence (Fig. 4d), the precipitation also weakens.

    Based on the above analysis, it should be mentioned that the strengthened subtropical high and motion of the synoptic system over the mid-high latitudes at 500 hPa provide a favorable background for this precipitation event. Further, the convergence of cold and dry northerly flow, and warm and moist southeasterly flow, at 850 hPa, over eastern Sichuan Province, is the important trigger mechanism for this precipitation event. Besides, significant moisture transport from the Bay of Bengal and South China Sea and its convergence provide favorable moisture conditions for this heavy precipitation event.

    Figure 4.  Distribution of vertically integrated (from the surface to 300 hPa) moisture flux (vectors; g m-1 s-1) and its divergence (shading; 10-6 g m-2 s-1) at (a) 0000 and (b) 1200 UTC 8 September, and (c) 0000 and (d) 1200 UTC 9 September. The rectangular area represents the location of eastern Sichuan Province.

4. Diagnosis of moist dynamic parameters
  • In this section, we analyze the ability of two moist dynamic parameters (moist vorticity and moist divergence) to diagnose a case of precipitation. We focus on their ability for the whole precipitation event.

  • Figure 5 shows the evolution of 6-h accumulated precipitation and its corresponding int during this precipitation event. It should be mentioned that 6-h accumulated int is the averaged int between two adjacent times. For example, the 6-h accumulated int at 0600 UTC is the averaged int between 0000 UTC and 0600 UTC, considering that the FNL data are produced every 6 h (i.e., 0000, 0600, 1200 and 1800 UTC).

    Figure 5.  The 6-h accumulated int (contours; units: 10-3 kg m-1 s-3) and precipitation (shading; units: mm) at (a) 1800 UTC 8 September, and (b) 0000 UTC, (c) 0600 UTC and (d) 1200 UTC 9 September. Positive (negative) contours corresponding to solid (dashed) lines are drawn at (…-7,-3,3,7,11 kg m-1 s-3).

    From 1200 UTC to 1800 UTC 8 September (Fig. 5a), the 6-h accumulated precipitation is weak. The precipitation band is mainly located to the east of the complex terrain shown in Fig. 1, orienting from northeast to southwest. The centers of precipitation are mainly located over southern and northeastern Sichuan Province, respectively. For the int, its intensity is also weak, with a center over (32°N, 105°E). Unlike the precipitation band, positive int shows a north-south orientation between 104°E and 106°E, but it shows some false-alarm regions over southeastern Sichuan Province. Generally, positive int does not coincide well with precipitation, except over northeastern Sichuan Province. Besides, it is noted that there is no positive int over southern Sichuan Province, where complex terrain is located, indicating the negative influence of terrain on the ability of int to depict precipitation over the complex terrain. With the development and motion of the synoptic system, from 1800 UTC 8 September to 0000 UTC 9 September (Fig. 5b), the precipitation band moves eastward to some extent, and two precipitation centers strengthen significantly. In particular, the intensity of precipitation over southern Sichuan Province reaches its maximum. At the same time, although the positive int center is still located east of the precipitation center, it becomes stronger than the int from 1200 UTC to 1800 UTC 8 September, and moves southward to some extent. The orientation of positive int begins to change from north-south to northeast-southwest, seeming to coincide well with precipitation compared with the int from 1200 UTC to 1800 UTC 8 September. It should also be pointed out that the false-alarm regions over southeastern Sichuan Province reduce to some extent. On the other hand, positive int still does not appear over southern Sichuan Province under the negative influence of complex terrain, as at 1800 UTC 8 September. From 0000 UTC to 0600 UTC 9 September (Fig. 5c), the precipitation band continues to move eastward. The intensity of precipitation over northeastern Sichuan Province reaches its maximum, while the intensity of precipitation over southern Sichuan Province decreases remarkably. The intensity of the corresponding positive int over northeastern Sichuan Province also reaches its maximum, and its distribution shows a clear northeast-southwest orientation. Moreover, the positive int center overlaps the precipitation center exactly. At this time, positive int coincides very well with precipitation over northeastern Sichuan Province. From 0600 UTC to 1200 UTC 9 September (Fig. 5d), both precipitation and positive int weaken significantly, but they still coincide well with each other. However, their centers do not coincide well with each other, as the int from 0000 UTC to 0600 UTC 9 September, i.e., the positive int center is located west of the precipitation center.

    Figure 6.  Latitudinal-vertical cross sections of (shading; units: 10-5 s-1) and 6-h accumulated precipitation (green bars; units: mm) between 104°E and 106°E, at (a) 1800 UTC 8 September, and (b) 0000 UTC, (c) 0600 UTC and (d) 1200 UTC 9 September.

    Figure 6 shows the latitudinal-vertical cross section of and 6-h accumulated precipitation between 104°E and 106°E during this heavy precipitation event. It should be mentioned that positive is mainly located in the mid-lower troposphere below 600 hPa. At the beginning of this precipitation event (Fig. 6a), although positive covers almost all of the precipitation area, its distribution is too wide, resulting in some false-alarm regions. Moreover, their centers do not coincide with each other. With the development of this precipitation event (Fig. 6b), both positive and precipitation become strengthened significantly, and they coincide well with each other compared with the from 1800 UTC 8 September to 0000 UTC 9 September. However, the positive center is still located south of the precipitation center. During the peak of this precipitation event (Fig. 6c), both the location and center of positive coincide well with precipitation. During the decaying period (Fig. 6d), positive weakens significantly, corresponding to the decreased precipitation, but they still coincide well with each other.

    Based on the above analysis, it can be concluded that when the influence of terrain is weak, may have a strong ability to diagnose the location and center of heavy precipitation, except in the initial period. In particular, best matches precipitation during the peak of the event. However, when the influence of terrain is strong, is unable to exactly depict the location and center of the heavy precipitation during the whole event.

  • Figure 7 shows the evolution of 6-h accumulated precipitation and its corresponding int during this precipitation event. The 6-h accumulated int is calculated as per the method used in section 4.1 for the 6-h accumulated moist vorticity.

    At the beginning of the precipitation event (Fig. 7a), the location of negative int moist divergence coincides well with the location of precipitation over northeastern Sichuan Province, but the negative int moist divergence center is located northeast of the precipitation center. On the other hand, there is positive int over southern Sichuan Province, where the influence of terrain is very strong, indicating that int is unable to diagnose precipitation over complex terrain exactly. From 1800 UTC 8 to 0000 UTC 9 September (Fig. 7b), with the increase in precipitation, the intensity of negative int over northeastern Sichuan Province reaches its maximum, but their centers do not coincide with each other. Besides, although int can depict the location of the precipitation, it also exhibits some false-alarm regions over the southwest of Shanxi Province, like the int from 1800 UTC 8 to 0000 UTC 9 September. On the other hand, int is unable to depict the precipitation over southern Sichuan Province. From 0000 UTC to 0600 UTC 9 September (Fig. 7c), negative int covers almost all of the precipitation area over northeastern Sichuan Province, but there are still some false-alarm regions. Besides, although the intensity of precipitation reaches its maximum, the intensity of int begins to weaken compared with the int from 1800 UTC 8 September to 0000 UTC 9 September. In the decaying stage of the precipitation event (Fig. 7d), both int and precipitation weaken significantly. Their locations still coincide well with each other to some extent, but the negative int center is located northeast of the precipitation center, indicating that the negative int center does not match the precipitation center. Moreover, false-alarm regions are still apparent over the southwest of Shanxi Province.

    Figure 8 shows the latitudinal-vertical cross section of between 104°E and 106°E during this heavy precipitation event. In the initial period of the event (Fig. 8a), negative covers all of the precipitation area, but it also exhibits some false-alarm regions. A significant negative center is observed at (32°N, 800 hPa), located north to the precipitation center. In the developing period of precipitation (Fig. 8b), negative tends to strengthen and begins to tilt northward under the influence of the terrain. It coincides well with precipitation, with a decrease in the number of false-alarm regions. In the peak stage of the event (Fig. 8c), negative weakens to some extent, but still tilts northward. It coincides well with precipitation, except for a mismatch in their centers. During the decaying stage of the event (Fig. 8d), coincides with most of the precipitation, except for a few false-alarm regions and a mismatch in their centers.

    The above analysis indicates that can depict the location of the entire heavy precipitation event when the influence of terrain is weak. However, when the influence of terrain is strong, is unable to depict the location of heavy precipitation, like . Besides, the center does not coincide well with the precipitation center during this event, i.e., their variability of intensity shows significant inconsistency, unlike .

    To further compare the abilities of and to depict the precipitation event, Fig. 9 shows time-latitude sections of int and int, and 6-h accumulated precipitation, along 105°E. For the int (Fig. 9a), before 1800 UTC 8 September, although its distribution covers all parts of the precipitation, its distribution is too wide, resulting in some false-alarm regions. After 1800 UTC 8 September, the number of false-alarm regions begins to decrease. In particular, the positive int center coincides very well with the precipitation center at 0600 UTC 9 September. For the int (Fig. 9b), its location coincides well with the location of precipitation, except for a few false-alarm regions. However, the int center does not match the precipitation center well, indicating inconsistency in their intensity variabilities.

    Figure 7.  As in Fig. 4 but for int. Positive (negative) contours corresponding to solid (dashed) lines are drawn at (…-5,-3,-1,1,3 kg m-1 s-3).

    Figure 8.  As in Fig. 7 but for .

    Figure 9.  Time-latitude sections of (a) int (contours; units: 10-3 kg m-1 s-3) and (b) int (contours; units: 10-3 kg m-1 s-3) and 6-h accumulated precipitation (shading; units: mm), along 105°E, during the precipitation event. Positive (negative) contours corresponding to solid (dashed) lines for (a) are drawn at (…-8,-4,4,8,12 kg m-1 s-3). Positive (negative) contours corresponding to solid (dashed) lines for (b) are drawn at (…-7,-5,-3,3,5 kg m-1 s-3).

    Figure 10 shows the evolution of precipitation and the two dynamic moist parameters over the precipitation center (31°N, 105°E) during the precipitation event. From Fig. 10a, the center of precipitation and the corresponding int show significant consistency during the precipitation event. Both develop at 1200 UTC 8 September, and reach their maximum at 0600 UTC 9 September, and then decay simultaneously. Their simultaneous variability indicates the strong ability of moist vorticity to depict the precipitation center. However, for the int (Fig. 10b), it reaches its peak at 0000 UTC 9 September, which leads the maximum precipitation by 6 h, indicating inconsistency between them. This suggests that int is unable to depict the precipitation center exactly.

    Figure 10.  Temporal evolution of (a) int (red line; units: 10-3 kg m-1 s-3) and (b) int (red line; units: 10-3 kg m-1 s-3) and 6-h accumulated precipitation (blue line), at (31°N, 105°E), from 1200 UTC 8 September to 1800 UTC 9 September.

5. Discussion and conclusions
  • Precipitation variability is important for agricultural activities and vegetation growth. Therefore, it is very important to improve the prediction of precipitation. In this study, a regional heavy precipitation event that occurred over Sichuan Province on 8-9 September 2015 is used as a case study. We begin by analyzing the background circulation of this precipitation event, pointing out that the subtropical high and synoptic system over the mid-high latitudes has an important impact. Then, we use two new vertically integrated moist dynamic parameters (i.e., moist vorticity and moist divergence), proposed by (Qian et al., 2015), to diagnose the event.

    When the influence of topography is weak (i.e., low altitude), both and have the ability to depict the location of precipitation to some extent. For , it is unable to diagnose the location of precipitation exactly during the initial stage of precipitation, exhibiting some false-alarm regions. However, as the precipitation develops, its ability to depict the location of precipitation improves significantly. In particular, best matches the location of precipitation during the peak of the event. Moreover, it can depict the precipitation center exactly when the precipitation reaches its maximum. For , although it exhibits some false-alarm regions, it can cover all parts of the precipitation during the whole precipitation event. However, the variability of the precipitation center and center show significant inconsistency, indicating does not depict the precipitation center exactly during the precipitation event. These results suggest that both and have significant ability to predict the location of precipitation. Moreover, has a stronger ability than in terms of predicting the variability of the precipitation center.

    When the influence of topography is strong (i.e., high altitude), both of these two moist dynamic parameters cannot exactly depict the location and center of heavy precipitation, during the whole event, over southern Sichuan Province, which is characterized by the complex topography of the southeastern part of the Tibetan Plateau. This suggests they possess weak ability to predict precipitation over complex terrain. Note that the surface topography acts as a boundary for the atmospheric circulation. It plays a fundamental role in determining climate gradients on Earth (Roe, 2005). The typical effect of topography on precipitation is the well-known mechanism of orographic lift, consisting of the local enhancement of precipitation by means of moist air uplift operated by versants (Bonacina, 1945; Alpert, 1986; Roe, 2005; Smith, 2006). Therefore, the first possible reason for the weak ability of two moist dynamic parameters to predict precipitation over complex terrain is that the ability of models to depict the complex terrain over southwestern China needs to be improved. On the other hand, data paucity is always a big problem over the Tibetan Plateau and its surrounding regions, which may be another possible reason. If these aspects can be improved, it is likely that the ability of moist dynamic parameters to diagnose precipitation over complex terrain will also improve significantly.

    It should be acknowledged that only one case is investigated in this study. To further demonstrate the ability of these two moist dynamic parameters to depict and predict precipitation, more case studies are needed. Besides, as pointed out by (Qian et al., 2015), although these two moist dynamic parameters show significant ability to depict precipitation, they cannot replace a complete, multiscale forecasting methodology. However, they could be used to post-process a model forecast to improve precipitation prediction. On the other hand, considering the impact of climate change on precipitation (Djebou and Singh, 2016), we should introduce the effect of climate into dynamic parameters to diagnose precipitation more precisely.

Reference

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return