Statistical Characteristics of Summer Season Raindrop Size Distribution in the Western and Central Tianshan Mountains in China

Abstract

The characteristics of the raindrop size distribution (DSD) in summer in the western (Nilek) and central (Urumqi) regions in the Tianshan Mountains of China were studied based on three years of second-generation Particle Size Velocity (Parsivel²) disdrometer data. The FengYun-2G satellite remote sensing data and the ERA5 reanalysis product were used to reveal the possible thermo-dynamical and microphysical processes that caused the dissimilarities in DSD between Urumqi and Nilek. The DSD in Nilek is significantly different from that in Urumqi. The concentration of mid- and large-size drops is higher in Nilek than in Urumqi. The DSD characteristics for six rain rate classes and two rain types (convective and stratiform) are studied. It is found that the raindrops in Nilek have overall higher mass-weighted mean diameters (D_m) and lower logarithm of normalized intercept parameters (log₁₀ N_w) than those in Urumqi, which is true for different rain rates and rain types. Convective clusters in Urumqi are similar to maritime clusters, whereas those in Nilek are more similar to continental clusters, according to a classification standard of convective clusters. The radar reflectivity, rain rate relations, and the shape and slope relations for rainfall in Urumqi and Nilek are also different. The DSD variability in the two regions may be attributed to differences in convective intensity that are closely related to the specific terrain of the Tianshan Mountains.

1. Introduction

Raindrop size distribution (DSD) information is essential for understanding the cloud microphysical processes (Rosenfeld and Ulbrich 2003) and improving the algorithms of radar quantitative precipitation estimation (Seliga and Bringi 1976; Ryzhkov and Zrnić 1995; Chapon et al. 2008). Knowing the DSD variability is of great importance for improving the microphysical parameterization schemes of numerical weather prediction models (Milbrandt and Yau 2005; Zhang et al. 2006). In addition, the DSD also plays an important role in soil erosion studies (Rosewell 1986; Nanko et al. 2016; Janapati et al. 2019).

DSD varies with climate regime, geographical location, and rain type (Ulbrich 1983; Tokay and Short 1996; Testud et al. 2001; Bringi et al. 2003; Rosenfeld and Ulbrich 2003; Zhang et al. 2001, 2003). Bringi et al. (2003) classified the convective clusters into maritime- and continental-like clusters based on the normalized intercept parameter (log₁₀ N_w) and massweighted mean diameter (D_m) derived from DSDs over different regimes, and found that the D_m of the maritime-like cluster is smaller than that of the continental-like cluster. Seela et al. (2017) indicated that the DSD of summer rainfall shows a higher D_m and a lower log₁₀ N_w in Taiwan than in Palau, although both Taiwan and Palau are islands located in western Pacific. Thompson et al. (2015) studied the DSDs from two equatorial Indian (Gan) and West Pacific Ocean (Manus) islands, and found that the two sites have similar DSD spectra of liquid water content, median diameter, log₁₀ N_w, and other integral rain parameters. Wu et al. (2019) investigated characteristics of summer raindrop size distributions in three typical regions of the western Pacific. They found the largest log₁₀ N_w values in the western West Pacific and the largest D_m values in the southern West Pacific. Rainfall structures in regions of different topographic features (mountains, transitional zones, plains, etc.) in southern France were studied by Zwiebel et al. (2016), who revealed the dependency of DSD on orography. Wu and Liu (2017) studied the characteristics of DSD over the Tibetan Plateau and southern China, and pointed out that the number concentration of raindrops of all sizes over southern China is much higher than that in the Tibetan Plateau for convective rainfall. Comparison of the DSD characteristics at five sites located at five districts of Nanjing during the East Asian rainy season (Pu et al. 2020) indicates that the percentage of total rainfall accounted for by extreme rainfall is significantly higher at Luhe (industrial zone) than at other sites by up to 38 %, and the largest log₁₀ N_w value also occurred at Luhe.

The above studies are mainly focused on DSD and rainfall in humid areas, while the research on DSD in arid areas is far less than sufficient. The Tianshan Mountains are about 2500 km long and 300 km wide and composed of a series of tall mountains, intermountain basins, and valleys. Located in the arid area in the hinterland of Eurasia with the main body in Xinjiang, China, the Tianshan Mountains are the farthest mountains from the ocean in the world (Sorg et al. 2012). Compared with other regions in central Asia, the Tianshan Mountains have more precipitation and water resources, which makes them known as “the water tower of central Asia” (Chen et al. 2016). Zeng et al. (2020) studied the diurnal variation characteristics of spring DSD in the Tianshan Mountains and found that they are related to the precipitation system, valley winds, and solar radiation. A recent study of the characteristics of rainy season DSD in the Tianshan Mountains (Zeng et al. 2021) indicated a clear difference between the DSDs over the Tianshan Mountains and in humid regions of China. However, the DSD during the main precipitation period in summer over the Tianshan Mountains has not been well studied. Moreover, the east–west span of the Tianshan Mountains is quite large, and it is still unclear whether there are differences in DSD in different areas of the Tianshan Mountains. Therefore, the present study attempts to illustrate the DSD differences in summer between the western (Nilek) and central (Urumqi) regions of the Tianshan Mountains based on three years of disdrometer data. The results of the present study will shed light on the microphysical processes in arid areas. The data and methods are briefly introduced in Section 2. The observational results are illustrated in Section 3. The possible reasons for the difference in DSD characteristics between the two regions are discussed in Section 4. The summary and conclusions are presented in Section 5.

2. Data and methods

2.1 Observational sites and instruments

The disdrometer data used in the present study were collected at two different regions (Urumqi and Nilek) in the Tianshan Mountains in Northwest China using a second-generation Particle Size Velocity (Parsivel²) disdrometer manufactured by OTT Hydromet (Kempten, Germany). Both Urumqi and Nilek are located near the Tianshan Mountains in China, and they represent large urban areas with abundant precipitation, good ecological environment characterized by high vegetation coverage and suitable measurement conditions, and dense human habitation near the Tianshan Mountains. In addition, considering the huge size and complex terrain of the Tianshan Mountains as well as the obvious differences in terrain between the western Tianshan Mountains and the central area of the Tianshan Mountains in China, whether there are differences in the DSD characteristics needs to be further explored. Therefore, we comparatively analyzed DSDs of Urumqi and Nilek in this study. Specifically, Urumqi (935 m a.s.l; 43.78°N, 87.65°E) is located in the central area of the Tianshan Mountains in China with east–west terrain, and Nilek (1105 m a.s.l; 43.80°N, 82.52°E) is located in the western Tianshan Mountains with a trumpet-shaped topography that opens to the west. The two sites are situated at almost the same latitude and have similar altitudes, which facilitates this comparative study. Figure 1 shows the locations of the two observational sites and the topography of the Tianshan Mountains, on the north and south sides of which are the extremely arid Taklimakan and Gurbantungut deserts. Chen, Y. et al. (2017) proposed that the Tianshan Mountains are important for water resource and ecological environment maintenance as well as social and economic development in the arid regions of central Asia. In order to further reveal the DSD variability over the Tianshan Mountains in China, two different regions, i.e., Urumqi and Nilek, were selected for comparative study in this paper.

Fig. 1.

Locations of the observation sites (the blue and red dot) and the topography (m) of the Tianshan Mountains.

The Parsivel² disdrometer can simultaneously measure the size and fall speed of precipitation particles (Löffler-Mang 2000; Tokay et al. 2014); detailed size and fall speed classification information is described in Yuter et al. (2006). In the past ten years, the Parsivel² disdrometer has been widely used in the measurement of DSD around the world (e.g., Jaffrain and Berne 2011; Thurai et al. 2011; Chen et al. 2013; Chen, B. et al. 2017; Marzuki et al. 2013; Tokay et al. 2013; Konwar et al. 2014; Wu and Liu 2017; Wu et al. 2019; Pu et al. 2020; Fu et al. 2020; Zeng et al. 2021; Wang et al. 2021). As proposed by Yuter et al. (2006), when a particle is partially within the measuring area, it may be misidentified as a small particle falling faster than other particles with the observed size, and these spurious particles are called margin fallers. Additionally, strong winds and splashing from raindrops hitting instrument surfaces during heavy rainfall may produce unrealistically large numbers of slow-falling particles (Friedrich et al. 2013). Thus, raindrops with diameters above 8 mm or fall speeds 60 % above or below the empirical fall velocity–diameter relation proposed by Atlas et al. (1973) are eliminated following the approach of Jaffrain and Berne (2011) and Friedrich et al. (2013). Considering the terrain height, before quality control, air-density adjustments are made to the fall speed-diameter relationship of Atlas et al. (1973) by multiplying the correction factor (Chen, B. et al. 2017; Wang et al. 2021), and the correction factors are 1.036 and 1.043 in Urumqi and Nilek, respectively. At the same time, the first two size classes are discarded because of the low signal-to-noise ratio. Furthermore, 1-min samples with fewer than 10 drops or a rainfall rate of less than 0.1 mm h⁻¹ are also excluded (Tokay et al. 2013). Figure 2 presents the number of drops in different diameter and velocity classes before and after quality control in Urumqi and Nilek, respectively. The fall speeds of raindrops of all sizes that are filtered out are mainly below 4 m s⁻¹, and they are most likely caused by strong winds and splashing, especially in Nilek. Eventually, there are 5219 and 9045 1-min effective DSD samples for Urumqi and Nilek, respectively, during the summers from 2018 to 2020.

Fig. 2.

Raindrop numbers in different diameter and fall speed classes before (a) and (b), and after (c) and (d) quality control in Urumqi and Nilek. The black solid line indicates the empirical fall speed–diameter relationship from Atlas et al. (1973), and the black dashed line indicates the ±60 % range of the relationship.

In addition to the data observed by the Parsivel² disdrometer, satellite data from FengYun-2G (FY-2G) (Hui et al. 2016) and reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) Fifth Reanalysis (ERA5) (Hersbach et al. 2019) are also collected. FY-2G is a geostationary meteorological satellite launched by the China Meteorological Administration. It is located above the equator at the longitude of 105°E, and its black body temperature (TBB) data from the IR window (about 11 micrometers) with 0.1° × 0.1° spatial and 1-hr temporal resolutions are used in the present study. Meanwhile, convective available potential energy (CAPE), vertical integral water vapor, horizontal wind field near the ground, vertical profiles of temperature, and relative humidity data with 0.125° × 0.125° spatial and 1-hr temporal resolutions are extracted from ERA5 and analyzed.

2.2 DSD parameters

Based on Parsivel² disdrometer data, the DSD is calculated by:   

where N (D_i) (m⁻³ mm⁻¹) is the number concentration of raindrops per unit volume per unit diameter interval for raindrop diameter D_i (mm); n_ij is the number of raindrops within the size bin i and raindrop terminal velocity bin j; V_j (m s⁻¹) is the fall velocity of class j that ranges from 0.05 m s⁻¹ to 20.8 m s⁻¹ [Table A2 in Yuter et al. (2006)]; Δt (s) is the sampling time, here 60 s; and ΔD_i (mm) is the interval of the i-th bin. S_eff (D_i) (m²) is the effective sampling area (Tokay et al. 2014) and is expressed as follows:   

The integral rainfall parameters include rain intensity R (mm h⁻¹), rainwater content W (g m⁻³), and radar reflectivity factor Z (mm⁶ m⁻³). The equations for their calculations are as follows:   

  

  

where ρ_w (1 g cm⁻³) is the density of water.

The three-parameter gamma function is widely used to represent the measured raindrop spectra proposed by Ulbrich (1983), and it is written as follows:   

where N₀ (mm^−1−µ m⁻³), µ, and Λ (mm⁻¹) represent the intercept, shape, and slope parameters, respectively. The nth-order moment of the drop size distribution is expressed as follows:   

The truncated moment method (Ulbrich and Atlas 1998; Zhang et al. 2003) is implemented to calculate the aforementioned three parameters with the third, fourth, and sixth moments as follows:   

  

  

where G is   

To solve the nonindependence problem associated with the parameters in the gamma function of DSD, the normalized gamma distribution that can better represent the raindrop spectrum has been proposed (Willis 1984; Sempere-Torres et al. 1994, 1998; Testud et al. 2001), which is expressed by:   

where N_w (mm⁻¹ m⁻³) is the normalized intercept parameter and D_m (mm) is the mass-weighted mean diameter. N_w, D_m, and f (µ) are calculated as follows:   

  

  

3. Results and discussion

The variation of the mean raindrop concentration, N (D_i) (m⁻³ mm⁻¹) with raindrop size D (mm) during summer in Urumqi and Nilek are displayed in red and blue, respectively, in Fig. 3. Throughout this study, raindrops with diameters of 1 mm to 3 mm are considered as mid-size drops, and drops below and above this range are considered small and large drops, respectively, according to previous studies (Tokay et al. 2008; Murali Krishna et al. 2016; Seela et al. 2017; Janapati et al. 2020). Figure 3 demonstrates that the concentration of mid-size and large raindrops is higher in Nilek compared to that in Urumqi, while the opposite is true for the concentration of small raindrops. Rainfall in Nilek has higher mean R, D_m, and lower log₁₀ N_w than rainfall in Urumqi. A lower concentration of small drops and higher concentration of mid-size and large drops in Nilek results in higher D_m values in Nilek than in Urumqi.

Fig. 3.

Mean raindrop concentrations in Urumqi and Nilek during summer. The total numbers of 1-min raindrop size distributions samples in Urumqi and Nilek are given by legends in parenthesis. The mean values (enclosed in angle brackets < >) of mass-weighted mean diameter (D_m, mm), rainfall rate (R, mm h⁻¹) and the normalized intercept parameter (log₁₀ N_w, mm⁻¹ m⁻³) for rainfall in Urumqi and Nilek are also shown.

3.1 Raindrop size distribution for different rain rate classes

In order to further determine the differences in DSD between Urumqi and Nilek, DSD observations collected at the two regions are classified into six rain rate classes on the basis of R. The six classes are defined as follows: C1: 0.1–0.5 mm h⁻¹, C2: 0.5–1 mm h⁻¹, C3: 1–2 mm h⁻¹, C4: 2–5 mm h⁻¹, C5: 5–10 mm h⁻¹, C6: ≥10 mm h⁻¹. The mean raindrop spectra in the two regions for the six rain rates are shown in Fig. 4. Statistics of rainfall corresponding to the six rain rate classes in Urumqi and Nilek are provided in Table 1. For the first two rain rate classes (Figs. 4a, b; C1: 0.1–0.5 mm h⁻¹, C2: 0.5–1 mm h⁻¹), concentrations of mid-size and large drops are higher in Nilek than in Urumqi. For the middle two rain rate classes (Figs. 4c, d; C3: 1–2 mm h⁻¹, C4: 2–5 mm h⁻¹), the concentrations of raindrops with diameters greater than 1.3 mm for C3 and 1.6 mm for C4 are higher in Nilek than in Urumqi. Additionally, raindrops with diameters larger than 2.1 mm and 2.3 mm also have higher concentration in Nilek than in Urumqi for the last two rain rate classes (Figs. 4e, f; C5: 5–10 mm h⁻¹, C6: ≥ 10 mm h⁻¹). Figure 4 clearly shows that even after classifying DSDs into different rain rate classes, mid-size and large drops are more common in Nilek than in Urumqi.

Fig. 4.

Average raindrop spectra in Urumqi (red color) and Nilek (blue color) corresponding to six rain rate classes (C1: 0.1–0.5, C2: 0.5–1, C3: 1–2, C4: 2–5, C5: 5–10, C6: ≥ 10 mm h⁻¹).

For the convenience to compare the six rain rate classes at a given location, average size spectra for C1 to C6 in Urumqi and Nilek are respectively superimposed on the same plot and the results are displayed in Fig. 5, which shows that all DSDs in Urumqi and Nilek have a distinct peak structure, and the spectral width and concentration of mid-size and large rain-drops both increase with increasing rain rate.

Fig. 5.

Average raindrop spectra in Urumqi (a) and Nilek (b) corresponding to six rain rate classes.

The box and whisker plot of the variations in D_m and log₁₀ N_w corresponding to different rain rate classes are shown in Fig. 6. In both regions, D_m increases with increasing rain rate class (Fig. 6a), which is caused by the increase in mid-size and large drops accompanied by larger rain rate. The D_m values in Nilek are higher than those in Urumqi due to higher concentrations of mid-size and large drops. The mean D_m value varies between 0.88 mm and 1.61 mm in Urumqi and 1.10 mm and 2.38 mm in Nilek. Contrary to the D_m, the log₁₀ N_w values are higher in Urumqi than in Nilek (Fig. 6b). The mean log₁₀ N_w value varies from 3.46 m⁻³ mm⁻¹ to 4.02 m⁻³ mm⁻¹ in Urumqi and from 2.98 m⁻³ mm⁻¹ to 3.26 m⁻³ mm⁻¹ in Nilek. The mean values of Z, W, D_m, and log₁₀ N_w in Urumqi and Nilek corresponding to the six rain rate classes are provided in Table 2.

Fig. 6.

Variations of the mass-weighted mean diameter (D_m, mm) (a) and the normalized intercept parameter (log₁₀ N_w, mm⁻¹ m⁻³) (b) in Urumqi (red) and Nilek (blue) corresponding to six rain rate classes. The central line of the box indicates the median, and the bottom and top lines of the box indicate the 25th and 75th percentiles, respectively. The bottom and top lines of the vertical lines out of the box indicate the 5th and 95th percentiles, respectively.

3.2 DSD variations for stratiform and convective precipitation

Stratiform and convective precipitation are two fundamental types of rainfall in nature with different physical mechanisms for precipitation formation. The DSD characteristics of the two rainfall types are significantly different (Tokay and Short 1996; Testud et al. 2001; Bringi et al. 2003; Sharma et al. 2009; Niu et al. 2010). In order to classify precipitation into stratiform and convective types, many researchers have developed different classification schemes based on disdrometer data (Tokay and Short 1996; Testud et al. 2001; Bringi et al. 2003; Ulbrich and Atlas 2007; Chen et al. 2013; Murali Krishna et al. 2016; Wen et al. 2016). In the present study, the classification criteria proposed by Bringi et al. (2003) and Chen et al. (2013) are used. Specifically, for at least 10 consecutive 1-min rain samples, the rainfall is determined to be stratiform if R is > 0.5 mm h⁻¹ and the standard deviation of R is £ 1.5 mm h⁻¹, and the rainfall is determined to be convective if R is ≥ 5 mm h⁻¹ and the standard deviation of R is > 1.5 mm h⁻¹. Samples that do not meet the above classification criteria are excluded.

The DSD variations for stratiform and convective precipitation in Urumqi and Nilek are shown in Fig. 7. For both regions, a relatively high raindrop concentration can be found for convective precipitation compared to that for stratiform precipitation, which is true for the raindrops of all sizes (Figs. 7a, b). In both Urumqi and Nilek, the stratiform regimes have nearly exponential distributions, whereas the convective regimes show a broad distribution, which might be at least partly attributable to the collisional breakup of large drops in convective rainfall (Hu and Srivastava 1995). To further compare the raindrop concentrations in Urumqi and Nilek for a given rain type, DSDs in both regions for stratiform and convective regimes are respectively presented in Figs. 7c and 7d. The two plots show that there are more raindrops with a diameter larger than 1.4 mm in Nilek than in Urumqi for the stratiform regimes, and raindrops with a diameter greater than 2.4 mm have a higher concentration in Nilek compared to that in Urumqi for convective regimes.

Fig. 7.

Variations in raindrop concentration with drop diameter for different precipitation types in Urumqi (red) and Nilek (blue).

To further explore the DSD characteristics for stratiform and convective precipitation in Urumqi and Nilek and compare the results with previous studies, the distributions of the mean D_m and log₁₀ N_w values are displayed in Fig. 8. The gray rectangles in Fig. 8 are for the continental and maritime convective rainfall clusters and the gray dashed line is the stratiform rainfall line proposed by Bringi et al. (2003). For the rainfall in both regions, convective regimes have higher mean D_m and log₁₀ N_w values than stratiform regimes. In contrast, both stratiform and convective precipitation in Nilek have higher D_m and lower log₁₀ N_w values than those in Urumqi. Comparing results of the present study with those of Bringi et al. (2003) for the convective cluster, it is found that convective DSDs in Urumqi are more similar to those of the maritime-like cluster, while convective DSDs in Nilek are somewhat similar to those of continental-like cluster. The above results suggest that the approach to classify continental and maritime convective rainfall clusters proposed by Bringi et al. (2003) may not be always appropriate for classification of convective precipitation in the region of Tianshan Mountains in China, considering that their classification method is mainly applied to rainfall in North America, Australia, and the Pacific region.

Fig. 8.

Scatterplots of the mean value of normalized intercept parameter (log₁₀ N_w, mm⁻¹ m⁻³) versus the mass-weighted mean diameter (D_m, mm) for convective rainfall and stratiform rainfall, where the hollow (solid) symbol corresponds to stratiform (convective) rainfall. The red squares and blue circles represent the mean values in Urumqi and Nilek, respectively. The two gray rectangles correspond to the maritime and continental convective clusters, and the gray dashed line is the stratiform rain line reported by Bringi et al. (2003). The green triangles, black stars, and purple diamonds represent the mean values obtained in previous studies by Chen et al. (2013), Ma et al. (2019), and Zhang et al. (2019), respectively.

Additionally, for stratiform rainfall in both regions, the mean D_m and log₁₀ N_w values appear on the left side of the stratiform rainfall line proposed by Bringi et al. (2003). Bringi et al. (2003) proposed two different microphysical processes that can lead to large D_m and log₁₀ N_w variations in stratiform precipitation, i.e., the melting of large snowflakes that is responsible for larger D_m and smaller log₁₀ N_w values and the melting of tiny graupel or smaller rimed ice particles responsible for smaller D_m and larger log₁₀ N_w values. As shown in Fig. 8, stratiform precipitation in Urumqi has smaller D_m and larger log₁₀ N_w values than those in Nilek, implying that stratiform precipitation in Urumqi is associated with tiny graupel or smaller rimed ice particles, whereas in Nilek it is related to melting of large snowflakes. In addition, the mean values of Z, W, D_m, and log₁₀ N_w for the stratiform and convective regimes in Urumqi and Nilek are listed in Table 3.

Figure 8 also provides observational results from other regions of China with the purpose of revealing the differences in DSD parameters between monsoon and arid regions of China. Compared with rainfall in Nanjing in eastern China (Chen et al. 2013), Zhuhai in southern China (Zhang et al. 2019), and Beijing in northern China (Ma et al. 2019), rainfall in Nilek shows a smaller log₁₀ N_w value, while the D_m value in Nilek is larger than that in Beijing and Nanjing and smaller than that in Zhuhai. Rainfall in Urumqi shows a smaller D_m value than those in the other four regions mentioned above, while the log₁₀ N_w value in Urumqi is larger than those in the other four regions, except for convective rainfall in Zhuhai. The above results apply to both stratiform and convective rainfall, indicating that the DSD characteristics are different between the monsoon and arid regions of China.

3.3 The D_m–R and log₁₀ N_w–R relations

Figure 9 shows the scatterplots of D_m and log₁₀ N_w corresponding to different rainfall rates in Urumqi and Nilek. The D_m values in Urumqi are scattered within the range from 0.4 mm to 2 mm with a few points that have values around 2–2.4 mm (Fig. 9a), whereas the D_m values in Nilek are distributed between 0.4 mm and 3.5 mm, with a few points that have values between 3.5 mm and 4 mm (Fig. 9b). In addition, the D_m distribution narrows down and its changes tend to be gentle as the rainfall rate increases for both regions, which could be attributed to the fact that DSDs reach an equilibrium state when the raindrop coalescence and breakup balance each other out at higher rainfall rates (Hu and Srivastava 1995). The values of log₁₀ N_w in Urumqi are mainly distributed between 2 m⁻³ mm⁻¹ and 5.1 m⁻³ mm⁻¹, whereas they are mainly scattered from 1.4 m⁻³ mm⁻¹ to 4.5 m⁻³ mm⁻¹ in Nilek. The power–law fitting algorithms derived for D_m–R and log₁₀ N_w–R are also presented in Fig. 9. Comparing the D_m–R relations at the two places, the coefficient and exponent values are higher in Nilek than in Urumqi. For the log₁₀ N_w–R relations, however, the coefficient value in Nilek is lower than in Urumqi. This indicates that for a given rainfall rate, precipitation in Nilek has higher D_m and lower log₁₀ N_w values than that in Urumqi. In addition, by comparing with the research results of DSDs of typhoon and non-typhoon rainfall observed in Taiwan obtained by Janapati et al. (2021), larger coefficient and exponent values appear in the D_m–R relation of Nilek but not in the two types of rainfall in Taiwan. For the log₁₀ N_w–R relations, the coefficient value of Urumqi is larger and that of Nilek is smaller than the two types of rainfall in Taiwan. From the comparison of the two relations between the area of this study and that between the Yangtze and Huaihe rivers in eastern China during summer reported by Jin et al. (2015), it can be seen that Nilek is significantly different from the area between the Yangtze and Huaihe rivers, while Urumqi is closed to the area between the same rivers.

Fig. 9.

Scatterplots of the mass-weighted mean diameter (D_m, mm) and the normalized intercept parameter (log₁₀ N_w, mm⁻¹ m⁻³) with rainfall rate in Urumqi (red) and Nilek (blue). The fitted power-law relationships are shown by black solid lines. (a) D_m-R relation and (c) log₁₀ N_w-R relation for rainfall in Urumqi; (b) D_m–R relation and (d) log₁₀ N_w–R relation for rainfall in Nilek.

3.4 Radar reflectivity (Z)–rain rate (R) relations

The Z–R relationship (Z = A · R^b) plays an important role in quantitative precipitation estimation by single-polarization radar, which heavily depends on the variability of DSD that is related to climate, topography, season, and rainfall type (Tokay and Short 1996; Atlas et al. 1999; Ulbrich and Atlas 2007; Chapon et al. 2008; Marzuki et al. 2013). In the Z = A · R^b relationship, the coefficient A is related to the presence of large or small drops, and the exponent b is related to the microphysical process. The collision–coalescence mechanism plays a dominant role in rainfall when b is greater than 1, while the collision–coalescence and breakup processes reach an equilibrium in homogeneous rainfall when b approaches 1 (Atlas et al. 1999; Atlas and Williams 2003; Steiner et al. 2004; Sharma et al. 2009; Seela et al. 2017; Janapati et al. 2020; Pu et al. 2020). The samples of convective rainfall collected in Urumqi and Nilek are insufficient in number and are also highly scattered, which makes it hard to fit an appropriate power–law relationship. Therefore, here we mainly focus on the Z–R relationship of stratiform rainfall, which prevails in the two regions. The Z–R relationship (Z = 200R^1.6) proposed by Marshall and Palmer (1948) has been commonly used in the mid-latitude continental region for stratiform rainfall (hereafter referred to as the MP-Stratiform relationship). Figure 10 shows the scatterplots of Z versus R and the fitted relationships for stratiform precipitation over the two regions. As shown in Fig. 10, the Z–R relationship of stratiform rainfall in Nilek has higher coefficient and exponent values than in Urumqi. The MP-Stratiform relationship is also presented in Fig. 10 for comparison. With a low radar reflectivity value [Z < 352.18 mm⁶ m⁻³ (i.e., 25.47 dBZ)], the MP-Stratiform relationship would overestimate the precipitation in Urumqi. The opposite is true when the radar reflectivity is high. For the stratiform rainfall in Nilek, the MP-Stratiform relationship would cause overestimation of rainfall. In other words, there are obvious differences in stratiform rainfall between Urumqi and Nilek, and the MP-Stratiform relationship should be used with caution in the two regions.

Fig. 10.

Scatterplots of radar reflectivity (Z, mm⁶ m⁻³) and rain intensity (R, mm h⁻¹) for stratiform rainfall in Urumqi (purple solid circles) and Nilek (green hollow circles). The fitted power–law relationships are shown by the red (for Urumqi) and blue (for Nilek) lines, respectively, and the black line indicates the empirical relationship (Z = 200R^1.6) proposed by Marshall and Palmer (1948).

3.5 µ–Λ relations

The µ–Λ relationship provides valuable information for in-depth understanding of DSD characteristics and variability (Chen et al. 2013; Zhang et al. 2003). The relationship varies with different climate regimes, rainfall types, and terrains (Cao et al. 2008; Vivekanandan et al. 2004; Zhang et al. 2003; Seela et al. 2018; Tang et al. 2014). Therefore, to enhance our understanding of DSD in arid regions, the two regions of the Tianshan Mountains located in the typical arid area of China are selected for the present study. Figure 11 shows the scatterplots of µ and Λ values in Urumqi and Nilek. To estimate the µ–Λ relationship for rainfall in Urumqi and Nilek, the criteria proposed by Chen, B. et al. (2017) are adopted in the present study. DSD data are filtered first and only those with total drop counts above 300 are used for further analysis. Zhang et al. (2003) and Cao et al. (2008) highlighted that when µ and Λ are greater than 20 mm⁻¹ and 20 mm⁻¹, respectively, the results are more likely to be attributed to measurement errors instead of rainfall physics. Therefore, these results are removed from this study. The second-degree polynomial µ–Λ relationships for rainfalls in Urumqi and Nilek are derived respectively and expressed as:   

  

Fig. 11.

Scatterplots of µ–Λ relationships and fitting curves for rainfall in Urumqi (a) and Nilek (b) with drop counts > 300. The gray lines correspond to the relationship Λ · D_m = 4 + µ given the values of D_m = 1.0, 1.5, and 2.0 mm.

Figure 11 clearly shows that the samples of rainfall are more evenly distributed over the entire range in Urumqi, while which are mainly concentrated in the region when µ < 10 and Λ < 10 mm⁻¹ in Nilek, meanwhile, it can be seen from the fitted µ–Λ relationship that, given a Λ, Nilek has a larger µ than Urumqi.

The relationship can also be expressed as Λ · D_m = 4 + µ (Ulbrich 1983). As shown in Fig. 11, compared with those in Urumqi, more samples of rainfall in Nilek are located in the higher D_m region, indicating that higher values of D_m are observed in Nilek than in Urumqi; this is consistent with the findings presented in Sections 3.1 and 3.2.

4. Discussion

To explore the possible reasons for the above results in the two regions, the CAPE, vertical integral of water vapor, horizontal wind field near the ground, vertical profiles of temperature, and relative humidity from ERA5 and the TBB from FY-2G in rainy days are calculated for the summers of 2018–2020, in which the statistics are made from data only in rainy days. Figure 12a displays the box and whisker plot of CAPE. It is apparent that the CAPE in Nilek is overall relatively higher than that in Urumqi in rainy days, suggesting that precipitation in Nilek is more convective than in Urumqi. Maddox (1980) proposed that TBB can be used as an indicator of convection intensity, and TBB £ ⁻³2°C is generally considered to be evidence of convective development. Figure 12b shows that the TBB values are lower in Nilek than in Urumqi in rainy days, indicating that the intensity of convection is stronger in Nilek than in Urumqi, which is closely related to the trumpet-shaped topography of Nilek that opens to the west of the Tianshan Mountains in China (Fig. 1). Affected by the topography, prevailing westerly winds in the westerly belt (Zhang and Deng 1987; Yang et al. 2011; Huang et al. 2017), and the valley winds with diurnal heating (Zeng et al. 2020), airflow is more likely to converge and rise than Urumqi. Furthermore, we compared the DSD characteristics of Nilek in this study with those of Yining (Zeng et al. 2021), located about 100 km west of Nilek, and Xinyuan (Zeng et al. 2020), located about 100 km east of Nilek, and the results showed that summer season DSD in Nilek has the largest mean D_m of 1.37 mm, and the mean D_m of the rainy season DSD in Yining and the spring season DSD in Xinyuan are 1.11 mm and 0.92 mm, respectively. These differences are closely related to the study periods of these three studies; however, the effect of different locations on these differences is unknown. Recently, Pu et al. (2020) and Han et al. (2021) used multiple disdrometers to study the DSD characteristics of Nanjing and Beijing, respectively, and found differences at the regional scale. Therefore, future research needs to investigate the regional variability of the DSDs in Nilek, Yining, and Xinyuan based on observation data of the same time period. Meanwhile, it is clear that Nilek has relatively higher vertical integral of water vapor than Urumqi, as seen in Fig. 13. Combining CAPE, TBB, and vertical integral of water vapor in the two regions it can be inferred that relatively higher water vapor with more active vertical movement leads to the growth of solid and liquid cloud particles to a sufficiently larger size by aggregation, riming, and collision–coalescence processes in Nilek than in Urumqi. The average vertical profiles of temperature and relative humidity and the horizontal wind field near the ground for Urumqi and Nilek are also shown in Figs. 14 and 15, respectively. It is clear that Nilek has relatively higher temperature at all pressure levels (Fig. 14a) and lower relative humidity below 500 hPa (Fig. 14b) than Urumqi. In addition, compared with the more consistent northwesterly wind in Urumqi, the wind direction is more dispersed and the wind speed is stronger in Nilek (Fig. 15). The relative humidity, temperature (Janapati et al. 2020; Seela et al. 2021), and wind (Zeng et al. 2019; Wu et al. 2019) are primary meteorological variables that contribute to the evaporation of raindrops. Combining vertical profiles of temperature and relative humidity and the horizontal wind field near the ground in the two regions, it can be inferred that the evaporation processes in Nilek are stronger than in Urumqi, which is possibly why Nilek has fewer small drops. The above explanation provides possible reasons for the occurrence of more small drops in Urumqi and more large drops in Nilek. In addition, we also noticed that Nilek has more slow-falling particles before quality control, as shown in Fig. 2. As Friedrich et al. (2013) proposed, large numbers of slow-falling particles may be produced unrealistically during heavy rainfall. For this study, there are significantly more samples for Nilek (125) than Urumqi (36) for the last rain rate classes (C6: ≥ 10 mm h⁻¹), as shown in Table 1. Moreover, through the above analysis of CAPE, TBB, and horizontal wind field near the ground, it can be seen that there are stronger convection processes and more scattered strong winds in Nilek, which may explain the production of a large number of slow-falling particles during the process of heavy rainfall in Nilek. Additionally, there are very few margin fallers for both stations before quality control, as shown in Fig. 2, which may be related to the improvement of Parsivel² disdrometer compared with Parsivel disdrometer (Tokay et al. 2014), the selection of fall speeds 60 % above or below the empirical fall velocity–diameter relation proposed by Atlas et al. (1973) after airdensity adjustments, or better homogeneity of the laser sheet of Parsivel² disdrometer based on more expensive lasers compared to Parsivel disdrometer resulting in improved measurement accuracy (Wen et al. 2017). Other possible reasons will be further studied in the future.

Fig. 12.

Box and whisker plots of (a) convective available potential energy (CAPE, J Kg⁻¹) and (b) black body temperature (TBB, °C) in Urumqi (red color box) and Nilek (blue color box). The center line of each box indicates the median value, and the bottom and top lines of the box indicate the 25th and 75th percentiles, respectively. The bottom and top of the dashed vertical lines indicate the 5th and 95th percentiles, respectively.

Fig. 13.

Box and whisker plots of vertical integral of water vapor (kg m⁻²) in Urumqi (red color box) and Nilek (blue color box).

Fig. 14.

Mean air temperature (°C) and (b) relative humidity (%) profiles in Urumqi (red) and Nilek (blue).

Fig. 15.

Wind rose map of the ground in (a) Urumqi and (b) Nilek.

5. Summary and conclusion

In this research, we investigate the characteristics of DSD over the Tianshan Mountains in China using disdrometer data measured using an OTT Parsivel² during the summers of 2018–2020. For the first time, the characteristics of summer DSD in two different regions (Urumqi and Nilek) of the Tianshan Mountains in China are analyzed. The main conclusions of this study are as follows.

1. Rainfall in Nilek has a higher concentration of mid-size and large raindrops compared with that in Urumqi. The opposite is true for the concentration of small raindrops. This might be attributed to the fact that convective intensity in Nilek is relatively stronger than that in Urumqi.

2. DSDs are classified into six rain rate classes as well as for stratiform and convective precipitation. Results show a higher concentration of large raindrops and a lower concentration of small raindrops in Nilek than in Urumqi. For all rain rate classes and precipitation types, rainfall in Nilek has a higher mass-weighted mean diameter (D_m) and a lower normalized intercept parameter (log₁₀ N_w) than that in Urumqi.

3. Compared with the convective cluster proposed by Bringi et al. (2003), the convective DSDs in Urumqi are more similar to those of the maritime-like cluster, whereas the convective DSDs in Nilek can be classified as continental-like DSDs (Bringi et al. 2009; Thurai et al. 2011). In addition, D_m and log₁₀ N_w in different regions of China (eastern, southern, and northern China) are compared with those in the arid region. We found that the DSD variability is closely related to climate regimes, rainfall types, and terrains.

4. Compared with that in Urumqi, the Z–R relationship for stratiform rainfall in Nilek has higher values of coefficient A and exponent b. The standard Z–R model (MP-Stratiform relationship) tends to overestimate stratiform precipitation in Nilek. For stratiform precipitation in Urumqi, however, the MP-Stratiform relationship would overestimate the precipitation at low radar reflectivity value and underestimate the precipitation at high radar reflectivity value. In addition, the µ–Λ relations are found to be different between Urumqi and Nilek.

The present study discusses the DSD characteristics and the possible factors affecting these characteristics in the western and central regions over the Tianshan Mountains in China. The results of this study are conducive to the improvement of local quantitative precipitation estimation and a deeper understanding of the DSD characteristic under the background of complex terrain in arid regions. Note that the findings of the present study may be affected by the limitations of the Parsivel² disdrometer in measuring small raindrops (Tokay et al. 2013; Zhang et al. 2019). The two-dimensional video disdrometer combined with other observation instruments should be used to further study the microphysical processes that involve cloud droplets and raindrops over the Tianshan Mountains in China in the future.

Data Availability Statement

The ERA5 reanalysis data provided by ECMWF are available at https://www.ecmwf.int/en/forecasts/datasets/, and the TBB data provided by the China National Satellite Meteorological Center can be downloaded from http://satellite.nsmc.org.cn. The DSD data generated and analyzed in this study are available from the corresponding author on reasonable request.

Acknowledgments

This work was supported by the Tianshan Mountains Talent Project (grant 2021-32), the National Key Research and Development Program of China (grant 2018YFC1507102 and 2018YFC1507104), and Uygur Autonomous Region Tianchi Project for Introducing High-Level Talents (2019). We acknowledge China National Satellite Meteorological Center and ECMWF for providing the data.

References

•  Atlas,  D., and  C. R.  Williams, 2003: The anatomy of a continental tropical convective storm. J. Atmos. Sci., 60, 3-15.
•  Atlas,  D.,  R. C.  Srivastava, and  R. S.  Sekhon, 1973: Doppler radar characteristics of precipitation at vertical incidence. Rev. Geophys., 11, 1-35.
•  Atlas,  D.,  C. W.  Ulbrich,  F. D.  Marks, Jr.,  E.  Amitai, and  C. R.  Williams, 1999: Systematic variation of drop size and radar-rainfall relations. J. Geophys. Res., 104, 6155-6169.
•  Bringi,  V. N.,  V.  Chandrasekar,  J.  Hubbert,  E.  Gorgucci,  W. L.  Randeu, and  M.  Schoenhuber, 2003: Raindrop size distribution in different climatic regimes from disdrometer and dual-polarized radar analysis. J. Atmos. Sci., 60, 354-365.
•  Bringi,  V. N.,  C. R.  Williams,  M.  Thurai, and  P. T.  May, 2009: Using dualpolarized radar and dualfrequency profiler for DSD characterization: A case study from Darwin, Australia. J. Atmos. Oceanic Technol., 26, 2107-2122.
•  Cao,  Q.,  G.  Zhang,  E.  Brandes,  T.  Schuur,  A.  Ryzhkov, and  K.  Ikeda, 2008: Analysis of video disdrometer and polarimetric radar data to characterize rain microphysics in Oklahoma. J. Appl. Meteor. Climatol., 47, 2238-2255.
•  Chapon,  B.,  G.  Delrieu,  M.  Gosset, and  B.  Boudevillain, 2008: Variability of raindrop size distribution and its effect on the Z–R relationship: A case study for intense Mediterranean rainfall. Atmos. Res., 87, 52-65.
•  Chen,  B.,  J.  Yang, and  J.  Pu, 2013: Statistical characteristics of raindrop size distribution in the Meiyu season observed in eastern China. J. Meteor. Soc. Japan, 91, 215-227.
•  Chen,  B.,  Z.  Hu,  L.  Liu, and  G.  Zhang, 2017: Raindrop size distribution measurements at 4,500 m on the Tibetan Plateau during TIPEX-III. J. Geophys. Res.: Atmos., 122, 11092-11106.
•  Chen,  Y.,  W.  Li,  H.  Deng,  G.  Fang, and  Z.  Li, 2016: Changes in Central Asia's water tower: Past, present and future. Sci. Rep., 6, 35458, doi:10.1038/srep35458.
•  Chen,  Y.,  Z.  Li,  G.  Fang, and  H.  Deng, 2017: Impact climate change on water resources in the Tianshan Mountains, Central Asia. Acta Geographica Sin., 72, 18-26 (in Chinese with English abstract).
•  Friedrich,  K.,  S.  Higgins,  F. J.  Masters, and  C. R.  Lopez, 2013: Articulating and stationary PARSIVEL disdrometer measurements in conditions with strong winds and heavy rainfall. J. Atmos. Oceanic Technol., 30, 2063-2080.
•  Fu,  Z.,  X.  Dong,  L.  Zhou,  W.  Cui,  J.  Wang,  R.  Wan,  L.  Leng, and  B.  Xi, 2020: Statistical characteristics of raindrop size distributions and parameters in Central China during the Meiyu seasons. J. Geophys. Res.: Atmos., 125, e2019JD031954, doi:10.1029/2019JD031954.
•  Han,  Y.,  J.  Guo,  Y.  Yun,  J.  Li,  X.  Guo,  Y.  Lv,  D.  Wang,  L.  Li, and  Y.  Zhang, 2021: Regional variability of summertime raindrop size distribution from a network of disdrometers in Beijing. Atmos. Res., 257, 105591, doi:10.1016/j.atmosres.2021.105591.
•  Hersbach,  H.,  B.  Bell,  P.  Berrisford,  A.  Horányi,  J. M.  Sabater,  J.  Nicolas,  R.  Radu,  D.  Schepers,  A.  Simmons,  C.  Soci, and  D.  Dee, 2019: Global reanalysis: Goodbye Era-Interim, hello ERA5. ECMWF Newsl., 159, 17-24.
•  Hu,  Z., and  R. C.  Srivastava, 1995: Evolution of raindrop size distribution by coalescence, breakup, and evaporation: Theory and observation. J. Atmos. Sci., 52, 1761-1783.
•  Huang,  W.,  S.-Q.  Chang,  C.-L.  Xie, and  Z.-P.  Zhang, 2017: Moisture sources of extreme summer precipitation events in North Xinjiang and their relationship with atmospheric circulation. Adv. Climate Change Res., 8, 12-17.
•  Hui,  W.,  Q.  Guo,  Q.  Wu, and  L.  Yang, 2016: Image quality evaluation of FY-2G meteorological satellite. Prog. Geophys., 31, 1902-1910.
•  Jaffrain,  J., and  A.  Berne, 2011: Experimental quantification of the sampling uncertainty associated with measurements from PARSIVEL disdrometers. J. Hydrometeor., 12, 352-370.
•  Janapati,  J.,  B. K.  Seela,  P.-L.  Lin,  P. K.  Wang, and  U.  Kumar, 2019: An assessment of tropical cyclones rainfall erosivity for Taiwan. Sci. Rep., 9, 15862, doi: 10.1038/s41598-019-52028-5.
•  Janapati,  J.,  B. K.  Seela,  P.-L.  Lin,  P. K.  Wang,  C.-H.  Tseng,  K. K.  Reddy,  H.  Hashiguchi,  L.  Feng,  S. K.  Das, and  C. K.  Unnikrishnan, 2020: Raindrop size distribution characteristics of Indian and Pacific Ocean tropical cyclones observed at India and Taiwan sites. J. Meteor. Soc. Japan, 98, 299-317.
•  Janapati,  J.,  B. K.  Seela,  P.-L.  Lin,  M.-T.  Lee, and  E.  Joseph, 2021: Microphysical features of typhoon and non-typhoon rainfall observed in Taiwan, an island in the northwestern Pacific. Hydrol. Earth Syst. Sci., 25, 4025-4040.
•  Jin,  Q.,  Y.  Yuan,  H.  Liu,  C.  Shi, and  J.  Li, 2015: Analysis of microphysical characteristics of the raindrop spectrum over the area between the Yangtze River and the Huaihe River during summer. Acta Meteor. Sin., 73, 778-788 (in Chinese with English abstract).
•  Konwar,  M.,  S. K.  Das,  S. M.  Deshpande,  K.  Chakravarty, and  B. N.  Goswami, 2014: Microphysics of clouds and rain over the Western Ghat. J. Geophys. Res.: Atmos., 119, 6140-6159.
•  Löffler-Mang,  M., and  J.  Joss, 2000: An optical disdrometer for measuring size and velocity of hydrometeors. J. Atmos. Oceanic Technol., 17, 130-139.
•  Ma,  Y.,  G.  Ni,  V.  Chandrasekar,  F.  Tian, and  H.  Chen, 2019: Statistical characteristics of raindrop size distribution during rainy seasons in Beijing urban area and implications for radar rainfall estimation. Hydrol. Earth Syst. Sci., 23, 4153-4170.
•  Maddox,  R. A., 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc., 61, 1374-1387.
•  Marshall,  J. S., and  W. M. K.  Palmer, 1948: The distribution of raindrops with size. J. Meteor., 5, 165-166.
•  Marzuki,  M.,  H.  Hashiguchi,  M. K.  Yamamoto,  S.  Mori, and  M. D.  Yamanaka, 2013: Regional variability of raindrop size distribution over Indonesia. Ann. Geophys., 31, 1941-1948.
•  Milbrandt,  J. A., and  M. K.  Yau, 2005: A multimoment bulk microphysics parameterization. Part I: Analysis of the role of the spectral shape parameter. J. Atmos. Sci., 62, 3051-3064.
•  Murali Krishna,  U. V.,  K.  Krishna Reddy,  B. K.  Seela,  R.  Shirooka,  P.-L.  Lin, and  C.-J.  Pan, 2016: Raindrop size distribution of easterly and westerly monsoon precipitation observed over Palau islands in the Western Pacific Ocean. Atmos. Res., 174–175, 41-51.
•  Nanko,  K.,  S. M.  Moskalski, and  R.  Torres, 2016: Rainfall erosivity-intensity relationships for normal rainfall events and a tropical cyclone on the US southeast coast. J. Hydrol., 534, 440-450.
•  Niu,  S.,  X.  Jia,  J.  Sang,  X.  Liu,  C.  Lu, and  Y.  Liu, 2010: Distributions of raindrop sizes and fall velocities in a semiarid plateau climate: Convective versus stratiform rains. J. Appl. Meteor. Climatol., 49, 632-645.
•  Pu,  K.,  X.  Liu,  Y.  Wu,  S.  Hu,  L.  Liu, and  T.  Gao, 2020: A comparison study of raindrop size distribution among five sites at the urban scale during the East Asian rainy season. J. Hydrol., 590, 125500, doi:10.1016/j.jhydrol.2020.125500.
•  Rosenfeld,  D., and  C. W.  Ulbrich, 2003: Cloud microphysical properties, processes, and rainfall estimation opportunities. Meteor. Monogr., 30, 237-258.
•  Rosewell,  C. J., 1986: Rainfall kinetic energy in eastern Australia. J. Climate Appl. Meteor., 25, 1695-1701.
•  Ryzhkov,  A. V., and  D. S.  Zrnić, 1995: Comparison of dual-polarization radar estimators of rain. J. Atmos. Oceanic Technol., 12, 249-256.
•  Seela,  B. K.,  J.  Janapati,  P.-L.  Lin,  K.  Krishna. Reddy,  R.  Shirooka, and  P. K.  Wang, 2017: A comparison study of summer season raindrop size distribution between Palau and Taiwan, two islands in western Pacific. J. Geophys. Res.: Atmos., 122, 11787-11805.
•  Seela,  B. K.,  J.  Janapati,  P. L.  Lin,  P. K.  Wang, and  M.-T.  Lee, 2018: Raindrop size distribution characteristics of summer and winter season rainfall over north Taiwan. J. Geophys. Res.: Atmos., 123, 11602-11624.
•  Seela,  B. K.,  J.  Janapati,  C. K.  Unnikrishnan,  P.-L.  Lin,  J. L.  Loh,  W.-Y.  Chang,  U.  Kumar,  K.  Krishna. Reddy,  D.-I.  Lee, and  M. V.  Reddy, 2021: Raindrop size distributions of North Indian Ocean tropical cyclones observed at the coastal and inland stations in South India. Remote Sens., 13, 3178, doi:10.3390/rs13163178.
•  Seliga,  T. A., and  V. N.  Bringi, 1976: Potential use of the radar differential reflectivity measurements at orthogonal polarizations for measuring precipitation. J. Appl. Meteor., 15, 69-76.
•  Sempere-Torres,  D.,  J. M.  Porrà, and  J.-D.  Creutin, 1994: A general formulation for raindrop size distribution. J. Appl. Meteor., 33, 1494-1502.
•  Sempere-Torres,  D.,  J. M.  Porrà, and  J.-D.  Creutin, 1998: Experimental evidence of a general description for raindrop size distribution properties. J. Geophys. Res., 103, 1785-1797.
•  Sharma,  S.,  M.  Konwar,  D. K.  Sarma,  M. C. R.  Kalapureddy, and  A. R.  Jain, 2009: Characteristics of rain integral parameters during tropical convective, transition and stratiform rain at Gadanki and its application in rain retrieval. J. Appl. Meteor. Climatol., 48, 1245-1266.
•  Sorg,  A.,  T.  Bolch,  M.  Stoffel,  O.  Solomina, and  M.  Beniston, 2012: Climate change impacts on glaciers and runoff in Tien Shan (Central Asia). Nat. Climate Change, 2, 725-731.
•  Steiner,  M.,  J. A.  Smith, and  R.  Uijlenhoet, 2004: A microphysical interpretation of radar reflectivity–rain rate relationships. J. Atmos. Sci., 61, 1114-1131.
•  Tang,  Q.,  H.  Xiao,  C.  Guo, and  L.  Feng, 2014: Characteristics of the raindrop size distributions and their retrieved polarimetric radar parameters in northern and southern China. Atmos. Res., 135–136, 59-75.
•  Testud,  J.,  S.  Oury,  R. A.  Black,  P.  Amayenc, and  X.  Dou, 2001: The concept of “normalized” distributions to describe raindrop spectra: A tool for cloud physics and cloud remote sensing. J. Appl. Meteor., 40, 1118-1140.
•  Thompson,  E. J.,  S. A.  Rutledge,  B.  Dolan, and  M.  Thurai, 2015: Drop size distributions and radar observations of convective and stratiform rain over the equatorial Indian and west Pacific Oceans. J. Atmos. Sci., 72, 4091-4125.
•  Thurai,  M.,  W. A.  Petersen,  A.  Tokay,  C.  Schultz, and  P.  Gatlin, 2011: Drop size distribution comparisons between Parsivel and 2-D video disdrometers. Adv. Geosci., 30, 3-9.
•  Tokay,  A., and  D. A.  Short, 1996: Evidence from tropical raindrop spectra of the origin of rain from stratiform versus convective clouds. J. Appl. Meteor., 35, 355-371.
•  Tokay,  A.,  P. G.  Bashor,  E.  Habib, and  T.  Kasparis, 2008: Raindrop size distribution measurements in tropical cyclones. Mon. Wea. Rev., 136, 1669-1685.
•  Tokay,  A.,  W. A.  Petersen,  P.  Gatlin, and  M.  Wingo, 2013: Comparison of raindrop size distribution measurements by collocated disdrometers. J. Atmos. Oceanic Technol., 30, 1672-1690.
•  Tokay,  A.,  D. B.  Wolff, and  W. A.  Petersen, 2014: Evaluation of the new version of the laser-optical disdrometer, OTT Parsivel². J. Atmos. Oceanic Technol., 31, 1276-1288.
•  Ulbrich,  C. W., 1983: Natural variations in the analytical form of the raindrop size distribution. J. Climate Appl. Meteor., 22, 1764-1775.
•  Ulbrich,  C. W., and  D.  Atlas, 1998: Rainfall microphysics and radar properties: Analysis methods for drop size spectra. J. Appl. Meteor., 37, 912-923.
•  Ulbrich,  C. W., and  D.  Atlas, 2007: Microphysics of raindrop size spectra: Tropical continental and maritime storms. J. Appl. Meteor. Climatol., 46, 1777-1791.
•  Vivekanandan,  J.,  G.  Zhang, and  E.  Brandes, 2004: Polarimetric radar estimators based on a constrained gamma drop size distribution model. J. Appl. Meteor., 43, 217-230.
•  Wang,  G.,  R.  Zhou,  S.  Zhaxi, and  S.  Liu, 2021: Raindrop size distribution measurements on the southeast Tibetan Plateau during the STEP project. Atmos. Res., 249, 105311, doi:10.1016/j.atmosres.2020.105311.
•  Wen,  L.,  K.  Zhao,  G.  Zhang,  M.  Xue,  B.  Zhou,  S.  Liu, and  X.  Chen, 2016: Statistical characteristics of raindrop size distributions observed in East China during the Asian summer monsoon season using 2-D video disdrometer and Micro Rain Radar data. J. Geophys. Res.: Atmos., 121, 2265-2282.
•  Wen,  L.,  K.  Zhao,  G.  Zhang,  S.  Liu, and  G.  Chen, 2017: Impacts of instrument limitations on estimated raindrop size distribution, radar parameters, and model microphysics during mei-yu season in East China. J. Atmos. Oceanic Technol., 34, 1021-1037.
•  Willis,  P. T., 1984: Functional fits to some observed drop size distributions and parameterization of rain. J. Atmos. Sci., 41, 1648-1661.
•  Wu,  Y., and  L.  Liu, 2017: Statistical characteristics of raindrop size distribution in the Tibetan Plateau and southern China. Adv. Atmos. Sci., 34, 727-736.
•  Wu,  Z.,  Y.  Zhang,  L.  Zhang,  H.  Lei,  Y.  Xie,  L.  Wen, and  J.  Yang, 2019: Characteristics of summer season raindrop size distribution in three typical regions of western Pacific. J. Geophys. Res.: Atmos., 124, 4054-4073.
•  Yang,  L.,  X.  Li, and  G.  Zhang, 2011: Some advances and problems in the study of heavy rain in Xinjiang. Climatic. Environ. Res., 16, 188-198 (in Chinese with English abstract).
•  Yuter,  S. E.,  D. E.  Kingsmill,  L. B.  Nance, and  M.  Löffler-Mang, 2006: Observations of precipitation size and fall speed characteristics within coexisting rain and wet snow. J. Appl. Meteor. Climatol., 45, 1450-1464.
•  Zeng,  Q.,  Y.  Zhang,  H.  Lei,  Y.  Xie,  T.  Gao,  L.  Zhang,  C.  Wang, and  Y.  Huang, 2019: Microphysical characteristics of precipitation during pre-monsoon, monsoon, and post-monsoon periods over the South China Sea. Adv. Atmos. Sci., 36, 1103-1120.
•  Zeng,  Y.,  L.  Yang,  Z.  Zhang,  Z.  Tong,  J.  Li,  F.  Liu,  J.  Zhang, and  Y.  Jiang, 2020: Characteristics of clouds and raindrop size distribution in Xinjiang, using cloud radar datasets and a disdrometer. Atmosphere, 11, 1382, doi: 10.3390/atmos11121382.
•  Zeng,  Y.,  L.  Yang,  Z.  Tong,  Y.  Jiang,  Z.  Zhang,  J.  Zhang,  Y.  Zhou,  J.  Li,  F.  Liu, and  J.  Liu, 2021: Statistical characteristics of raindrop size distribution during rainy seasons in northwest China. Adv. Meteor., 2021, 6667786, doi:10.1155/2021/6667786.
•  Zhang,  A.,  J.  Hu,  S.  Chen,  D.  Hu,  Z.  Liang,  C.  Huang,  L.  Xiao,  C.  Min, and  H.  Li, 2019: Statistical characteristics of raindrop size distribution in the monsoon season observed in southern China. Remote Sens., 11, 432, doi:10.3390/rs11040432.
•  Zhang,  G.,  J.  Vivekanandan, and  E.  Brandes, 2001: A method for estimating rain rate and drop size distribution from polarimetric radar measurements. IEEE Trans. Geosci. Remote Sens., 39, 830-841.
•  Zhang,  G.,  J.  Vivekanandan,  E. A.  Brandes,  R.  Meneghini, and  T.  Kozu, 2003: The shape-slope relation in observed gamma raindrop size distributions: Statistical error or useful information? J. Atmos. Oceanic Technol., 20, 1106-1119.
•  Zhang,  G.,  J.  Sun, and  E. A.  Brandes, 2006: Improving parameterization of rain microphysics with disdrometer and radar observations. J. Atmos. Sci., 63, 1273-1290.
•  Zhang,  J., and  Z.  Deng, 1987: A Generality of Rainfall in Xinjiang. Meteorological Press, Beijing (in Chinese).
•  Zwiebel,  J.,  J.  Van Baelen,  S.  Anquetin,  Y.  Pointin, and  B.  Boudevillain, 2016: Impacts of orography and rain intensity on rainfall structure. The case of the HyMeX IOP7a event. Quart. J. Roy. Meteor. Soc., 142, 310-319.