4.1 Synoptic and Climatic Conditions/climatology
Large scale monsoon circulation and moisture distribution controls the spatial and vertical extension of clouds during the monsoon seasons (RJ11). Sathiyamoorthy et al. (2004) explained that a strong Tropical Easterly Jet produces convective storms over most head of BOB, and Indian Ocean able to produce the anvil clouds associated with deep clouds. Figure 5 shows the spatial distribution of monsoon circulation and specific humidity at various pressure level (5a) 850 hPa, (5b) 700 hPa, (5c) 500 hPa and (5d) 200 hPa respectively.
Low level westerly jet stream during monsoon seasons coming from the AS across the south peninsula is the most prominent feature of the monsoon circulation at 850 hPa (RJ11). During the monsoon seasons, low level moisture (850hPa) is transported from both AS and the BOB to the central India and Himalayan foothills (Fig. 5a). The maximum specific humidity exceeds more than 16 g/kg over the central India along the monsoon trough. Another maximum is also observed northeast India and south equatorial Indian Ocean. At 850 hPa beyond 30°N, the strength of westerly low level jet decreases and turns into the easterlies associated with the monsoon trough and a region of strong moisture gradient aligned in Himalayan Foothills. The low-level moisture also remains higher (increase) but it is relatively lower Hindu Kush Mountain. At 700 mb, the higher moisture concentration is observed over land regimes of south East Asia and westerly winds are shifted/established to southern side of the AS and BOB (Fig. 5b). A weaker northerly is observed over southern Pakistan and the northern AS at 500 mb (Fig. 5c). At upper level (e.g., 200 mb) the westerly jet stream moves northward, placing its southernmost edge at 30°N (Figs. 5d). We also observed a Tropical Easterly Jet along 12°N across the south Peninsula at 200 hPa (Sathiyamoorthy et al. 2004). A right entrance with left exists quadrants with large scale divergence induced the convection.
The average precipitable water (PR_wtr; Fig. 6a) and integrated water vapour (IWV; Fig. 6b) is maximum over BOB, head BOB and AS. Over these areas evaporation increases the boundary-layer moisture with the surface latent heat fluxes contribution. The low-level southwesterly winds (jet) take this moisture from the ocean towards the Indian subcontinent (Fig. 5). The advected winds from AS and BOB, upon landfall encounter dramatically different with the land surface and hence the surface heat fluxes are not same over these two regions (Niyogi et al., 2002; Douglas et al., 2006). South of Eastern Himalaya has local maximum in surface latent heat (Fig. 6c) which is the results of evapotranspiration and these areas is characterized by irrigated cropland, wetlands, and forests. However, a local minimum in surface latent heat fluxes is observed near the Thar Desert, where surface sensible heat flux dominates (Fig. 6d). The local minimum of surface latent heat flux over Desert area does not extend farther downstream (i.e. to the northeast), although these down-streams consist of few very intense convection (Zipser et al. 2006; Houze paper). The desert area is characterized by the negative surface sensible heat flux and Thar Desert consists of extremely hot low-level air that moves to relatively cooler surface of the western indentation, which is on average characterized by moisture soil. There has a low moisture content over Afghanistan and western Pakistan, which extends up to North Africa and the Middle East and is a large-scale feature (Annamalai et al. 1999). As a result, there is a strong moisture gradient observed near India–Pakistan border. This moisture gradient separates the moist monsoonal air over the Indian Peninsula from the dry air that pervades North Africa and the Middle East (Medina et al. 2010).
3.2. Spatial distribution of different types of cloud systems defined in the present study
Figure 7 shows the spatial distribution of different types of cloud systems (7a) CCs, (7b) DCCs, (7c) DCSs and (7d) ICSs defined in the present study in each 1° x 1° grid boxes. An important feature is observed, and based on the intensity (ICSs), depth (DCCs) and types of the cloud systems (DCSs and CCs) locations are varying and earlier also observed in the TRMM and GPM bases analysis (RH11; RH10; Liu and Zipser 2015; K19a; KS19). The CCs are nearly uniformly distributed over the Indian subcontinent, however a close inspection shows few favouable regions of of occurrences of CCs (Fig. 7a). At the Himalayan belt (IGP,, Central and Eastern Himalaya Foothills (CHF & EHF), Meghalayan plateau (MEP)), Tibetau PleatauMayanmar Barma coast (MWC) along with IWC including WG have higher frequency of CCs. Central India consists of higher frequency of CCs. AS near the coastal areas and HOB follows the mountainon areas, whereas east coast of BOB and southern BOB have much less frequency of CCs.
The locations of DCCs change remarkably compared to CCs and now we can have higher frequency of DCCs only at head BOB along with MEP, CHF and EHF (Fig. 7b). GT also consists of comparable number of DCCs, whereas BOB have higher frequency of DCCs compared to AS. WG has also higher frequency of DCCs, when compared to surrounding ocean and land, but very less when compared to BOB, CHF and EHF. DCSs show completely different pictures and now the DCSs are more close toward the mountain slopes both at east and west coast of India. DCSs are higher at Himalayan slopes including WHF, CHF and EHF along with MEP. Myanmar and Barma coast (MWP) along with GT also consist of higher frequency of DCSs. MEP consist of Khasi hills, one of the rainiest areas on the Earth’s and consists of higher frequency of DCS too. Intense based cloud systems shows the clearest difference between the mountain, land and oceanic areas. Deep and intense convective systems clearly shows that WG has more intense convective systems (Ze) compared to deep convective systems (K14), e.g., WG has higher chance of intense CCs compared to deeper CCs (Fig. 7c;d). Myanmar coast has more DCSs compared to WG (K14). ICSs are more frequent at the slope of the Himalaya, MEP and Eastern Himalaya foothills and related with the moisture convergence (RH10; RH11). AS and BOB also shows the different characteristics based on different types of the cloud systems. BOB has more DCCs, DCSs and ICSs compared to AS and already observed in the TRMM and GPM based analysis (Kumar and Bhat 2015).
All the distributions are showing two important features. First, variability in east and west cost of India and second, difference between land and ocean dominated areas. For example, the highest occurrences of DCCs are observed over the HOB and adjoining mountain areas followed by the IGP and EHF. East cost of mountain areas namely MYM and coast, Khasi hills and GT consist of higher number of DCCs. WG and AS also have smaller DCCs. Overall west side of India have a smaller number of DCCs compared to east coast of India. The spatial distribution of DCSs and ICSs show the different characteristics compared to DCCs. The DCSs are mostly concentrated at the north- east area, central India and WGs, whereas ICSs are located the Himalayan belt. If we compare the locations of different types of the cloud systems with the locations of the maximum amount of rainfall, we see the higher rainfall over HOB and GT matches with the DCCs and DCSs. The locations of WG and WHF rainfall is associated with the more ICSs. Within the WG, north WG is associated with the ICSs whereas south WG rainfall is mostly matching with the locations of DCSs. Secondary maxima near the Eastern Ghats are associated with the DCCs and DCSs.
4.3. Characteristics and internal structure of deep convective cores (DCC)
The average vertical profiles for DCCs are shown in Fig. 8b over the selected areas. The average vertical profiles show the peak nearly 8–9 km altitude and decreases as we move both upward and downwards from 9 km altitude (Dodson et al. 2011; K23). To investigate the regional differences clearly, we enlarged the vertical profiles below (< 6 km) and above (> 11 km). The in-large images show that mountainous areas such as WHF, IGP along with central India are showing an interesting feature and all these areas gain a higher Ze values at higher altitude (> 12 km). This may be associated with the updraft speed and flow along the mountain areas (House 2012). Liu et al. (2007) reflected that in CPR measurements a higher Ze in upper altitudes is result of dense, large, reflective particles generated in DCCs because of higher vertical velocities. WG, GT and AS have the least Ze above the 12 km altitude and related with the weaker intense vertical profiles. However, below the freezing altitude (~ 5 km), opposite characteristics are observed. For example, hilly areas (WHF and IGP) have the least Ze whereas western side of India (GT and BOB) have the highest Ze values. These variations are related with the cloud microphysics and cloud droplets formations (Liu et al. 2007).
Figure 9 shows the CFAD (contoured frequency by altitude; Yuter and Houze 1995) for DCCs over the selected areas. CFAD shows the relative occurrences of Ze in each 1 dBZ (between − 30 to 20 dBZ) and 240 meters vertical intervals. The regional differences among the selected areas are very small and all the CFAD’s show nearly the similar structure/patterns. An arc shaped is observed in all the CFAD’s (Dodson et al. 2011) and maximum Ze is observed in mid troposphere (6–9 km) with modal value lies near 10 dBZ. This is consistent with the results of Bacmeister and Stephens (2011) and K23 where deep convective clouds had curved outlines. Below the melting layer (indicated by the vertical discontinuity around 5 km), Ze decreases as the signal is attenuated by heavy precipitation and bigger raindrops (Luo et al. 2008). This may be related to ice microphysical processes. The higher sized hydrometeors fall out quickly in the downdraft compared to small sized hydrometeors. And, since Ze is proportional to the sixth power of diameter assuming Rayleigh scattering and ignoring phase of hydrometeors, small-sized hydrometeors leave the less Ze. The higher Ze (~ 10 dBZ) is related with the supercooled liquid cloud particles along with the overlying precipitating particles associated with deep clouds (Liu et al. 2007). These CFADs are consistent with the previous studies associated with deep convection using CloudSat profiles (Bodas-Salcedo et al. 2008; Satoh, Inoue, and Miura 2010; Nam and Quaas 2012; K23). All the CFAD has a curved bow is arc shape with high Ze or concentration with the mixed phase region. A bow structure in the average vertical profiles and CFAD is indicating the that both ice (warm) microphysical growth process from the top of higher (lower) estates in the vertical structure.
4.4. Cloud top height distribution of DCCs
To examine the internal structure of DCCs, the following parameters are analyzed: a) cloud top height (CTH; maximum height of − 28 dBZ), b) Echo top height (ETH, the maximum height reached by the 0- and 10-dBZ CPR echoes, referred as ETH_0 dBZ and ETH_10 dBZ, respectively), c) the distance between the CTH and ETH, which indicates the fuzziness (Luo et al. 2011) of cloud tops. Figure 10 shows the cumulative frequency distribution (CDF) and frequency distribution of CTH and ETH of DCCs over selected areas. 0 dBZ separates the precipitating and non- precipitating clouds (Houze 1993; Steiner and Houze 1995) and can reveal the height from where the precipitation starts falling in DCCs. CTH has the highest regional differences, whereas ETH_10 has the least reginal differences. West and east coast differences are also highest in CTH and least in ETH_10 dBZ.
The CTH distributions are also showing that IGP has the highest fraction of deeper DCCs (Fig. 10a;d), and 50% DCCs are crossing the 15.5 km altitude, whereas over WG, the corresponding altitude is 13.5 km. Overall more than 20% DCCs are crossing the 13 km altitude over all the areas. ETH_0 is showing nearly similar feature as the CHT, but the regional differences is less compared to CTH, and now WHF starts gaining the higher-level ETH_0 e.g. in the precipitation regimes (RH10; RH11). Nearly 32% DCCs are crossing the 14 km altitude over IGP and WHF, whereas over WG and AS only 10% DCCs are crossing the 14 km altitude (Fig. 10b;e). ETH_10 shows the different characteristics, and now WHF has highest ETH_10 altitude and more than 20% DCCs are higher than 12 km altitude whereas only 8% DCCs are crossing the 12 km altitude over BOB and AS. In the interesting way WG also gain higher altitude ETH_10 whereas BOB matches with the AS (Fig. 10c;f). WHF has less CTH, but highest ETH_10 dBZ and reflects that once the precipitation starts, they are intense and deeper one (RH10; RH11 and Zipser et al. 2006). These characteristics are related with the hydrometeors size difference in clouds and precipitation (Liu et al. 2008; RJ11; Sindhu and Bhat 2013; K23). It clearly reflected that at WHF deep convection is mostly favored by the precipitation sized particles (earlier observed in TRMM/GPM based observations; KS19; RH10; RH11), whereas over oceans deep convection is favored by the cloud-size hydrometeors (KS19).
Figure 11 (a-c) shows the box plot for CTH (a), ETH_0 dBZ (b) and ETH_10 dBZ (c) altitudes. In each boxplot the lower and upper line shows the 25 and 75% percentiles of the DCCs, whereas red line and green circle are showing median and average height of DCCs. Within all the selected areas the median and average CTH is higher than 14 km, which is the located near the tropical tropopause layer (TTL). Over WG, the least fraction of CTH is reaching the 14 km altitude. DCCs which are reaching near the TTL are very important as at and above the TTL, they can either dehydrate or hydrate the TTL, and depends on the size of ice crystals lifted in the upper atmosphere due to strong convection (Jensen et al. 2007). IGP and HOB have the highest median and average CTH. The boxplot of ETH_0 shows that WHF starts gaining the higher altitude however still IGP has the highest average ETH_0 dBZ. This becomes more evident in the boxplot of ETH_10 dBZ and WHF has the highest average and median ETH_10 dBZ altitude compared to other areas, and clearly reflects that deep convection is mostly favored by the precipitation sized particles over WHF (RH10; RH11; Qie et al. 2014) compared to EHF, central India and BOB, where cloud sized particles are important.
Figure 11 (e-f) shows the differences of CTH-ETH_0 (e) and CTH-ETH_10 (f) and known as fuzziness. These figures are indicating the how the cloud tops are packed between the precipitating (> 0 dBZ) and non- precipitating areas (< 0 dBZ; Luo et al. 2011; 2018; K23). The differences between CTH and ETH shows the difference between the altitude where the cloud and precipitation sized hydrometeors can reach. The less fuzziness (small distance between CTH and ETH) indicates the more packed convective tops, and dominated by the smaller sized particles (Luo et al. 2018), whereas higher fuzziness indicates that convective tops in the DCCs can carry the higher sized hydrometeors in the upper troposphere (Luo et al. 2011). Land on ocean differences is also observed here. WHF, EHF and WG have least width compared to eastern continent of India; e.g., they have more packed clouds. It is observed that WHF and WG DCCs consist of large sized of precipitation sized hydrometeors at their tops (KS19) e.g., the chance of precipitation size hydrometeors in DCCs can reach deeper into the atmosphere (RH10; RH11). Eastern coast and side of India (HOB and MYN) have the higher differences between CTH and ETH and indicated that they have less sized of hydrometeors at higher altitude in DCCs (KS19) or they have higher area of less Ze values and associated with anvil or stratiform clouds over these areas (Kumar 2016; Houze and Schumacher 2003; Zudeima 2003). Oceanic areas (AS, BOB, HOB) have the higher differences between the CTH and ETH.
Konwar et al. (2012) explained the relationship between drop size, concentration and cloud top height in the monsoonal clouds using the aircraft measurement. He showed that DSDs parameters depend on the meteorological conditions. They also showed that CCN advected to Indian continent from the west play a vital role in the cloud droplets formation, and growing of them into the precipitation sized hydrometeors. KS19 and Kumar and Silva 2020 also showed the relationship between the hydrometeors size, concentration and intensity at various altitude using TRMM/GPM observations. They also observed the higher sized of hydrometeors at western side (WG and WHF) of India compared to eastern side in the deeper and intense precipitating cells. RJ11also showed the climatological mean effective radius (MER) over South Asian regions using CloudSat observations. They showed that westerly advected CCN and giant CCN (GCCN) over WG simulates the higher cloud drops over WG compared to IGP. However, both studies are showing the opposite characteristics, as CloudSat mostly measures the cloud sized hydrometeors, where TRMM/GPM measures the precipitation sized hydrometeors. The differences between two studies clearly shows the early droplet formation at oceanic areas, but because of the difference in the growing mechanism the precipitation sized particle and rainfall dominate over different areas. The higher concentration of CCN is associated with the weaker updraft speed (Zipser and Lutz 1994; Lucas et al. 1994), whereas higher updraft speed over land areas carries the hydrometeors at higher altitude whereas weaker updraft speed leaves the higher hydrometeors over oceanic areas, which also leads to lower ETH. Long-term TRMM/GPM observations showed that WG has more shallower echoes compared to Myanmar, However GT has higher mid-level precipitating cells (K14; Kumar and Bhat 2017; Kumar 2018). K14 explained that higher GCCN and liquid water leads the dominance of warm rain process in the WG linked with the shallower cloud tops, whereas MYM along with BOB have more cloud ice, graupel and snow, and indicates the higher possibility of cold rain which is coming from mixed phase processes and having the deeper clouds. At the same time strengthen and the direction of wind pattern at different pressure level and moisture content also modify the cloud characteristics specially near the mountain (Zhang et al. 2018; Kumar et al. 2019b; Kumar et al. 2020b).
4.5. DCS and ICS characteristics
Figure 12 shows the characteristics of DCSs and ICSs defined in the present study. Figure 12a shows the distribution of horizontal span of DCSs. This horizontal span is different from the what we observed through the TRMM and GPM (Median et al. 2010; Kumar and Bhat 2016) as TRMM/GPM could produce a vertical cross section of the three-dimensional precipitating cloud systems, which CloudSat can’t. CloudSat only provides the horizontal span along the scanning track, which is quite important as they provide the horizontal extension of non-precipitating areas/extension of the clouds (< 0dBZ) (Awaka et al. 1997). To get the detailed information of DCSs and ICSs, we divided them into three categories; small (< 120 km), medium (between 120 and 320 km), and large (> 320 km) based on number and width of DCSs and ICSs (Luo et al. 2011; K23).
Figure 12a shows the frequency occurrences of different sized DCSs populations over the selected areas. AS and BOB have the highest fraction (~ 55%) of largest DCSs followed by the GT and EHF (~ 50%). Also, WHF and IGP have least fraction of large DCSs. Opposite characteristics are observed for small sized DCSs, and WHF and IGP have the highest fraction of small sized DCSs (40% and 35% respectively), whereas AS and BOB have the least fraction of small sized DCSs (~ 20%). Figure 12b shows that boxplot of DCSs horizontal span over the selected areas. In each boxplot the upper and lower edge shows the 25 and 75 percentiles of the DCSs horizontal span; the red line shows the median, and the green dots indicate the average horizontal span of DCSs. Again, AS and BOB have the highest median (~ 400 km) and average horizontal span (~ 790 km) compared to other selected areas and followed by GT (median: 375 km, average: 750 km). Along the Himalayan belt, WHF and IGP have the least median and average DCSs horizontal span (~ 200 km), whereas EHF has the highest median (~ 350 km) and average (~ 500 km) DCSs span. At the east coast of BOB and MYN has the least median and average DCSs span (~ 220 km and 400 km respectively). Figure 12c shows the CTH for the DCSs over the selected areas and it is observed that regional differences in CTH is less compared to DCCs. In the small regional differences, WHF and WG has least average CTH (~ 12), however median CTH over WG is higher (~ 13 km) compared to WHF (~ 11.75 km). BOB and IGP have higher median CTH (~ 14.5 km) followed by MYN (~ 14 km). Over WG and AS in the ICSs and DCCs, the rapid precipitation formation due to collision and coalescence process in warm clouds does not allow the hydrometeors to go deeper into the atmosphere (Konwar et al.2014; Maheskumar et al. 2014).
Figure 12d shows the frequency occurrences of three different size categories to the total population of ICSs over selected areas, and more than 50% ICSs have horizontal span less than 120 km, except over AS. AS has the highest fraction of small ICSs (~ 70%) followed by WG and GJ-AS. So, it is very rare that cloud systems could be large as well as intense over the Western side of India (see the DCSs population). However opposite characteristics are observed at the eastern part the India. At that part the ICSs could be large as well as intense also, and BOB and GT have the highest fraction of large ICSs. Figure 12e is showing the horizontal span of ICSs and overall horizontal span is less compared to DCSs. GT, BOB and HOB have the largest and highest average horizontal span width (~ 400–500 km) and even their frequencies are higher. These cloud systems are associated with the higher frequency of broad stratiform precipitation over these areas (RH10; RH11; K19a). Western side of India including AS, WG, GJ-AS and WHF have the least median and horizontal span ICSs ( ~ < 250 km). Figure 12f is showing the CTH of ICSs and we can clearly see Western side (AS, WG, MTC) have the least average (< 9 km) and median (< 6 km) CTH (Kumar and Bhat 2019) for ICSs compared to eastern side, where average (~ 10–11 km) and median (~ 8 km) CTH is higher. However, WHF also shows the higher average and median CTH, and could be associated with the intense precipitating cells over WHF (Zipser et al. 2006; RH10; RH11; Kumar and Bhat 2016; KS19).
It is important to observe that the regional differences are higher in height based convective systems (DCSs) compared to intensity (Zemax) based convective systems (ICSs) for horizontal span. DCSs show that east-west differences in horizontal span but not in the CTH and intense based ICSs show the different characteristics and do not show the much east west difference in horizontal span but a clear difference are observed in CTH. Overall ICSs have less horizontal span and CTH compared to DCS, and shows that in intense precipitation the rainfall is falling down at early stage which do not allow them grow them more horizontally and vertically (Konwar et al. 2012; K14). Similar features are observed using the TRMM/GPM based precipitation features (PFs), where AS and BOB were consisting of largest PFs near to the coastal boundaries (K19a). They also showed that largest area of PFs are contribution to higher surface rainfall compared to intense rainfall events over AS, GJ-AS BOB, HOB but they did not grow vertically. However deepest PFs are observed mostly land dominated areas namely IGP and CHF. Hirose and Nakamura (2005) also showed that horizontal dimension of precipitating systems over AS is smaller than the systems over BOB using TRMM PR observations. RH10 and RH11 also studied the wide convective core (WDC) and broad stratiform regions (BSR) over the India during summer monsoon seasons and showed that WDCs are higher along the CHF, Khasi Hills and over the BOB. They also observed the BSR, associated with convective systems are highest at eastern side of the BOB. They argued that WDC and BSR near and over BOB is related with the occurrence of MCSs with broad stratiform regions. They also observed BSR over the AS and northern Eastern Ghats but their frequencies are very less.
4.6. Cloud characteristics
To investigate the cloud characteristics, we analyzed several parameters (a) CCs CTH, (b) Ave Ze (Zeave), (c) Maximum Ze (Zemax), (d) horizontal span of CCs and (e) width of the CCs (Htcc- Hbcc). Figure 13a shows the CCs CTH distribution over the selected areas. All the areas show the bimodal distribution in CCs CTHs as earlier observed in CloudSat based observations (Das et al. 2017; Chen et al. 2018; K23). The primary peak in CCs CTH lies between 3–4 km (low level clouds) altitude except for WHF, central India and ETH where primary peak occurs at 7 km altitude. The secondary peak occurs at 14–15 km altitude and associated with the high-level clouds (Subhramanyam and Kumar 2012). GJ-AS has the higher fraction (15%) of low level of CCs (~ 3 km) followed by WG and AS (~ 10%, ~ 3.5 km) e.g., mostly at the west side of India. WHF and ETH have higher fraction (~ 12%) of CTH (7–8 km). IGP and main monsoon zone nearly shows the flat curve between 4–7 km altitude compare to other areas and reflected the mid-level clouds (Kumar and Bhat 2017). The orographic areas have higher CTH and could be associated with the orographic lifting through elevation rise where the wind pattern at the upward side can take the cloud hydrometeors at higher level (Houze 2012; Rh10; Zhang et al. 2018; Kumar et al. 2019b; Kumar et al. 2020b). BOB has highest fraction (~ 12%) of CCs CTH near secondary peak (~ 14 km). WHF has the least fraction (~ 6%) of CCs nearly 14 km altitude and showed that cloud sized hydrometeors are not reaching deeper into the atmosphere, which is either associated with either anvil or stratiform precipitation (RH10; RH11). CloudSat observations are also showing the higher frequency of CTH below 2 km or less (CF18), which TRMM/GPM PR is not able to detect because of strong surface echoes. Overall, all the selected areas have less than 5% CCs, which are crossing the 15 km altitude, compared to DCCs, where 65% DCCs are crossing the 15 km altitude (Fig. 10). The fraction of CCs, which are crossing the 10 km altitude (deep convection > 10 km; Houze et al. 2007; Fu et al. 2018; Kumar 2017b) are highest over WHF (7%), whereas least over GJ-AS among all the selected areas. For all other areas CCs, which are crossing 10 km altitude lies between 4 to 5%. The fraction of CCs with CTH < 5 km (e.g., shallower tops Kumar and Bhat 2017; Kumar 2018) lies between 5 to 8%, with highest fraction at WG and EHF (8%) and lowest fraction is at WHF and GJ-AS (5.5%).
Figure 13b shows the distribution of Zave for CCs over the selected areas. The Zeave has a single peak at around − 18 dBZ over the selected areas, and them all the areas are showing a hump between − 10 and − 6 dBZ, although the regional differences are not significant over the selected areas. The Zave has very less possibility (2%) of rain (Ze > 0 dBZ; Liu et al. 2008). Figure 13c shows the boxplot for the maximum Ze over the selected areas and the Zemax clearly shows that less the 25% are not CCs do not have rain (Ze > 0 dBZ; Liu et al. 2008). It clearly shows that regions near to orography and mountain Hills have higher Zemax and for more than 75% CCs Zemax is higher than 0 dBZ. Average Zemax is highest over the GT and EHF. WG is also showing the higher fraction of Zemax and associated with the shallower but intense cloud systems. GJ-AS has the least Zemax and associated with semi air condition in this region. Figure 13d is showing the box distribution of horizontal span of CCs with median, 25 and 75 percentile and average values over the selected areas. The median horizontal span is less than 50 km over all the selected areas and in a small regional difference AS and GJ-AS have the least median span (< 40 km), whereas along the Himalayan belt e.g., WHF and EHF have little higher horizontal span width (~ 45 km). Distribution of 75 percentile of CCs horizontal span shows higher reginal differences and GT has the highest horizontal span (~ 120 km) followed by EHF and Central India (~ 110 km). Overall western side/coast of India has less horizontal span compared to eastern coast and Himalayan Foothills and observed in TRMM and GPM based measurement earlier also and discussed earlier also (RH10; RH11; Kumar et al. 2019a). The average horizontal span of CCs (green lines) is showing the higher regional differences and now the differences between west and east size/coast is clearly visible. At the western side, WG has the highest horizonal span (~ 130 km), whereas GJ-AS has least (~ 100 km) horizontal span. Overall GT has highest average horizontal span (~ 180 km) followed by BOB (~ 170 km) and HOB (~ 150 km) are these regions also showed the higher area of stratiform precipitation in TRMM observations (RH10; RH11). All the selected areas over Himalayan Foothills (WHF, IGP and EHF) have nearly same horizontal span.
Figure 13e is showing the box distribution of total CCs width (Htcc - Hbcc) with median, average width along with the 25 and 75 percentile CCs width (lower and upper edge of the box) over the selected regions. The median width is nearly 1 km over all the areas except over the GJ-AS, which is nearly 0.75 km. 25 and 75 percentiles of CCs width are showing the higher regional differences, and, mountain and land areas such as over Himalayan foothills and topographic regions (WHF, IGP, EHF and MYN) have the higher CCs width, whereas AS and GJ-AS have the least CCs width. It may be related with the higher horizontal or non-rainy precipitation area. The average CCs width is showing a clear regional differences and eastern coast/side of India along with Himalayan Foothills have higher CCs width (~ 1.5 km) and compared to western side (less than 1.5 km over AS and GJ-AS), and least over the GJ-AS. May be arid and dessert areas near GJ-AS is restricting the CCs horizontal and vertical development of CCs (Houze et al. 2007; Median et al. 2010). Again, it is observed the regional difference is higher in horizontal span compared to vertical reach into the atmosphere. West side of India has small sized CCs with less CTH, but they can be very intense specially over WHF and WG.
4.7. Zonal and meridional variation of cloud properties
To investigate a zonal and meridional variation of cloud related parameters, we plotted the various CCs parameters at each 10-degree latitude (longitude) belt against longitude (latitude) to investigate the zonal and meridional variations. Figure 14 (a-c) shows the zonal variation of various CCs parameters (a) CTH, (b) CCs horizontal span and (c) CCs Zemax for different latitude belts. The cloud depth, horizonal span and intensity increase as they shift from west to east at Indian subcontinent except for the subtropical latitude belt 25°N-30°N. At the subtropical latitude belt (25°N to 30°N), which consists of Himalayan Foothills (WHF, CHF and EHF) shows the opposite nature compared to tropical longitude belt (10–15°N, 15–20°N and 20–25°N), where the geographic locations such as orography, moisture convergence and capping inversion mechanism play a vital role in formation of clouds followed by precipitation (Swaymer et al. 1947, Houze et al. 2007, Medina et al. 2010). At the subtropical latitude the CTH decreases from west to east, however their intensity (Zemax) and horizonal span increases and associated meteorological condition and mechanism which increase the rainfall (Konwar et al. 2012; K14). It shows that at the west side of India (< 75°E) at high latitudes, CCs do not develop much horizontally but (Fig. 14), but they could be deeper one (Houze et al. 2007, Zipser et al. 2006; Kumar and Bhat 2016).
The highest CTH are observed between 75°E-90°E, where their horizontal span also increases except for the latitude belt (10°N-20°N). A sudden peak in horizontal span and Zemax is also observed in near 75°E (10°-15°N) and associated with the AS and WG, where AS and WG have large fraction of large DCSs and small ICSs and WG has more intense rainfall events during summer monsoon seasons (Fig. 14). It shows that the once CSs cross the coastal boundary, they tend to develop more in horizontal, gain the intensity and CTH up to 95°E between 10°-20°N. Oceanic CCs have much less CTH, size and intensity compared to land and topographic areas. Beyond 95°E, except at the lowest latitude belt (10°-15°N) all the CSs related parameters decrease near the EHF and east coast of BOB and west coast of MYN (RH10; RH11; Kumar and Bhat 2019). Near the HOB (85°-95°E & 15°-20°N) there is an increase in horizontal span width (blue curve) which is lagging the CTH (peak near 82-84E) and leading the Zemax (peak in 85°- 90°E).
Figure 14(d-f) shows the meridional variation of CCs related parameters (d) CTH, (e) CCs horizontal span and (f) Zemax at different longitude belt. We can see that oceanic areas over all the longitudinal belt have less CTH (below 10°N) and horizontal span with less intensity specially at the west side (reg and green). But at the extreme east side of longitude belt shows different characteristics. Once the CCs enters in orographic and land areas, CCs CTH and Zemax are increasing but their horizontal span is decreasing. It shows that CCs are more horizontally developed, but with less CTH and intensity over ocean compared to land and orographic areas whereas CCs have less horizontal span (e.g., horizontally developed) but they are intense. The less CTH with less Zemax but higher horizontal span over the oceanic areas is associated with higher fraction of stratiform and anvil clouds which are not contribution to much of the rainfall (RH10; RH11; Kumar 2016; K19a).
Zemax is increasing at the Himalayan foothills and orographic (WG & MYN and e.g, at higher longitude and latitude areas, beyond 25°N), but their horizontal span is less compared to oceanic areas. A higher Zemax is proxy for higher sized of hydrometeors (KS19, Kumar and Silva 2020) and rain rate. The CCs related three parameters shows the mechanism involved in CCs development at different areas. Most of the CCs at high latitude are less wide but more intense and higher CTH, whereas, at low latitude, CCs are mostly wide and less intense (Fig. 14a). The CCs occurs at mid- and subtropical latitudes also associated with the mid latitude’s cyclones, so they are also less vertically developed compared to CCs at low latitudes. CCs at high latitudes at the center and Eastern Himalayan Foothills, orography can simulate them at and they have higher CTH, less horizontal span and high Zemax. At Himalayan Foothills is the effect of moisture convergence and capping inversion mechanism, and moisture transported for the AS and BOB (south of the continent). The LHF and SHF (as mentioned earlier) also increases the buoyancy and responsible for higher CTH and horizontal span. At the mountainous areas daytime intense solar heating developing local instability conditions, favorable for deep convective clouds (Houze 2012; RMZ10). Zonal differences in CTH are higher at west of India, however difference in horizontal span and Zemax is higher at east side/coast of India. CTH differences are higher at the west side of India compared to eastern side.