4.1.1. Vertical variation of potential temperature
Potential temperature is an important temperature parameter of dry air. Because it is conserved through dry adiabatic processes, it can be used to compare the thermal differences of air mass under different air pressures. However, potential temperature is not conserved when there is a release of latent heat. We can evaluate atmospheric stability by using vertical variation rate profiles of potential temperature as (Sheng et al., 2003) \begin{equation} \label{eq7} \frac{\partial\theta}{\partial z}=\frac{\theta}{T}(\gamma_{\rm v}-\Gamma)\approx\frac{\theta}{T}(\gamma_{\rm d}-\Gamma),(7) \end{equation} where θ is the potential temperature calculated from Eq. (7), Γ is the temperature lapse rate of a stratified atmosphere, γv is the temperature lapse rate of humid air, and γ d is the temperature lapse rate of dry air. When ∂θ/∂ z>0 and (γ d-Γ)>0, the atmosphere is stable; when ∂θ/∂ z<0 and (γ d-Γ)<0, the atmospheric condition is unstable.
In the following analysis, the vertical variation rate profiles of potential temperature in Wanyuan before and after precipitation are presented; potential temperature was obtained using the Poisson formula in Eq. (7). The AIRS scanning time and footprint coordinates are included in Table 2, and the vertical variation profiles of potential temperature before and after precipitation are depicted in Fig. 5.
As shown in Fig. 5, potential temperature vertical variations for whole layers are negative, and the entire atmosphere is unstable, particularly below 10 000 gpm prior to precipitation. When rainfall occurred, the vertical variation in potential temperature for whole layers became positive. This phenomenon implies that the process of rainfall releases instability energy of the atmosphere; thus, the stability of atmospheric stratification increases after rainfall.
4.1.2. Variation in cloud water content
Cloud water content is a key parameter for describing the characteristics of clouds and moisture and is a primary indicator of the potential for artificial precipitation. Cloud moisture plays a key role in the occurrence of rainfall. Therefore, it can be used to aid the forecasting of rainstorms, and the variation of cloud water content can provide insight into the physical mechanism of precipitation. Many studies have revealed that the precipitation amount depends strongly on cloud water content and that these parameters are correlated (Ding et al., 2011; Li and Niu, 2012).
The variation in cloud water content before and after rainfall in Wanyuan was analyzed by using the cloud water content data from AIRS. The AIRS scanning time and footprint coordinates are shown in Table 2 and in Fig. 6. It is evident that cloud water content significantly increases after precipitation; its unit increased approximately two magnitudes once precipitation began. However, the vertical distribution of cloud water content showed little variation, with large values at low levels and smaller values at high levels.
To verify the reliability of AIRS cloud water content data, we analyzed the relationship between FNL precipitation and cloud water content data (figure not shown). The result shows that the period of maximum precipitation is inconsistent with the period of maximum cloud water content because the magnitude of precipitation is not dependent on cloud water content only; it is also related to cloud particle concentration, water vapor transport conditions, and other factors. At 0700 UTC 11 July, the FNL cloud water content showed a significant increase and reached a peak after heavy rainfall, which can be attributed to the increased evaporation of the underlying surface water after a rainfall event. (Su et al., 2003) concluded that cloud water content increases after precipitation events and that the distribution of cloud water content can be depicted as unimodal in type by detecting the physical characteristics of precipitation clouds.
4.2.3. Variation in OLR
OLR is the energy radiating from Earth as infrared radiation at low energy to space. OLR is a critical component of Earth's energy budget and represents the total radiation emitted to space by the atmosphere (Susskind et al., 2011). The OLR is affected by clouds and dust in the atmosphere, which tend to reduce it to below clear-sky values.
The granule numbers of AIRS OLR data were 188 (9 July), 186 (11 July), and 063 (13 July). The AIRS OLR data in Fig. 7 indicate that the developing phase of the SWV was from 1848 UTC 9 July to 1836 UTC 11 July (Figs. 7a and b). During this phase, the value of OLR was obviously lower in Sichuan, Chongqing and Guizhou compared with other areas. With the eastward movement of the SWV and its shift to the puissant phase, the area with low OLR values moved eastwards, and the OLR value further reduced. Subsequently, the area of low OLR values moved gradually eastwards and finally left Sichuan and Chongqing. At 0618 UTC 13 July (Fig. 7c), the area with low OLR values moved into Hunan and Hubei provinces. From the movement of the low OLR area, we determined that the propagation path of the area with low OLR values is consistent with the path of the SWV.
4.1.4. Change in the temperature of brightness blackbody
Cloud top brightness temperature is the outward radiation of cloud tops obtained from infrared detection channels in meteorological satellites. In cloudless areas, it reflects the outward radiation of Earth's surface. Cloud top brightness temperature is the most primitive measurement for generating infrared images and various enhanced cloud pictures and is equivalent to the temperature of brightness blackbody (TBB). In cloudy regions, TBB is the blackbody radiation temperature at cloud tops; low TBB values correspond to higher cloud tops and more vigorous convection.
Thus, TBB can reveal the existence of clouds and some significant characteristics of the phase of evolution of clouds. In addition, TBB values display the relationship with surface precipitation to some extent (Cao et al., 2013). (Fei et al., 2008) revised the minimum scale criteria for the β-MCS (mesoscale convective system) census according to the scaled classification criteria proposed by (Orlanski, 1975). They revised the criteria so that the TBB values were less than or equal to 241 K and the diameters of continuous cold cloud areas were greater than 20 km. In addition, they mentioned that TBB values less than or equal to 221 K indicate deep convection. The granule numbers for the AIRS TBB data are the same as those for the OLR data.
The time series of TBB depicted in Fig. 8 shows that, at 1848 UTC 9 July, there was no obvious low TBB area. At 1836 UTC 11 July (Fig. 8b), an obvious low TBB area appeared at the junction of Sichuan, Guizhou and southern Chongqing, with a TBB value of approximately 230 K. Then, the low TBB area disappeared at 0618 UTC 13 July. Through a diagnostic analysis of a rainstorm event, (Wang and Cheng, 2013) determined that lower TBB values within a certain range are indicative of more intense precipitation. The time that the obviously low TBB area appeared coincided with the time at which the SWV reached its puissant phase, which also coincided with periods of heavy precipitation. This phenomenon shows that TBB data can reflect the development of MCS and has a strong relationship with precipitation intensity.
4.1.5. Changes in temperature and humidity profiles in different phases of the SWV
According to vortex data at 700 hPa collected using MICAPS, this SWV process can be divided into seven phases: nascent, developing, puissant, receding, redevelopment, new puissant, and perishing. The corresponding times for each phase are shown in Table 3. According to the movement of the SWV (Fig. 2), the maximum vortex center was located at (30°N, 107.5°E) from 1200 UTC 11 July to 1200 UTC 12 July. To examine the changes in temperature and moisture caused by the SWV's movement, we analyzed the characteristics of temperature variation and mixing ratio profiles from the developing phase to the receding phase. The AIRS data used for this analysis are shown in Table 4.
We screened the coordinates of the maximum vorticity based on AIRS temperature and humidity profiles. As indicated by the distribution of temperature (Fig. 9a) and mixing ratio (Fig. 9b) from the developing phase to the receding phase, the change in temperature in the lower layers was small during the SWV's developing phase, and an isothermal layer existed with the temperature maintained at 285 K between 900 hPa and 600 hPa. In addition, the vertical temperature lapse rate was very small. When the evolution of the SWV approached the puissant phase, the temperature of the lower layers was greater than that during the developing phase, and the isothermal layer disappeared. When the SWV approached the receding phase, the temperature of the lower layers was smaller than that during the puissant phase, and the temperature below 500 hPa was slightly greater than that during the developing phase. As depicted in Fig. 9b, the moisture distribution of the whole layer showed a "low layers wet while high layers dry" pattern from the developing phase to the receding phase. During the puissant phase, the vertical distribution of the mixing ratio peaked at 900 hPa in the center of the SWV, and the layers beneath 900 hPa displayed a significant moisture inversion. When the SWV approached the receding phase, the water vapor content was obviously reduced compared with that in the developing and puissant phases. Moreover, the distribution of the mixing ratio fluctuated, and a moisture inversion occurred between 700 hPa and 500 hPa.
To further investigate the reasons for the changes in moisture profiles and to analyze the transport of water vapor, we calculated the divergence of water vapor vertical flux from the surface to 300 hPa. To reduce the influence of moisture content owing to the SWV's movement, we analyzed the vapor transport condition from the developing phase to the receding phase of the SWV. During this period, the SWV was stable at the junction of Sichuan and Chongqing. Torrential rain occurrence is often closely related to water vapor; a continuous supply of water vapor is a necessary condition for the formation, development, and occurrence of rainstorms. The vertical flux of vapor divergence characterizes the vertical transport of vapor flux divergence. Compared with vapor flux divergence, the vertical flux of vapor divergence can better describe the powerful rise, convergence and divergence movement associated with torrential rainfall and its water vapor transport conditions (Li and Deng, 2013). The distribution of the vertical flux of vapor divergence in this SWV process is shown in Fig. 10.
During the nascent and developing phases (Figs. 10a and b), the SWV was located in an area characterized by positive vertical moisture flux divergence, which is evidence of the vertical advection of moisture divergence. When the evolution of the SWV approached the puissant phase (Fig. 10c), the source of the SWV was the convergence center of the vertical flux of vapor divergence, which implies the presence of intense water vapor convergence compared with the intense vertical transport. Thus, as shown in Fig. 9a, the mixing ratio profiles during the puissant phase of the SWV indicate that the moisture inversion occurred below 900 hPa. Furthermore, the mixing ratio was underestimated with respect to observations during the developing phase. When the evolution of the SWV approached the receding phase (Fig. 10d), the value of the vertical flux of vapor divergence had reduced but was still positive. In addition, a rainfall process occurred during this phase, which caused the value of the mixing ratio to be less than that during the puissant phase.