The map in Fig. 1 depicts the amount of accumulated rainfall over five days for various EPEs in India. The EPE that occurred over Assam and Meghalaya in the northeast (NE) received the highest amount of accumulated rainfall, followed by events in Madhya Pradesh and Rajasthan in central India (CI), Maharashtra (MH) in western India, and Gujarat (GJ). The accumulated rainfall spread over a larger area in GJ and MH, covering 825 sq. km and 600 sq. km, respectively. However, the EPE over CI and the NE were more concentrated and covered fewer areas, with 500 sq. km and 525 sq. km, respectively. It is noteworthy that while the EPE over the NE occurred during the initial phase of the 2022 summer monsoon, the other events occurred in July and August, after the full onset of monsoon in 2019 and 2021.
An enhanced view of the spatial variability of 5-day accumulated precipitation (column 1) from IMD for all EPEs, along with the climatology (column 2) and anomalies (column 3), are shown in Fig. 2. Based on IMD gridded data, the highest recorded five-day accumulated rainfall was 2645 mm over Phlangwanbroi village. This village is in the Mawsynram subdivision of East Khasi Hills district in Meghalaya (NE). Additionally, neighbouring areas experienced significant rainfall, with an accumulation value of around 2500 mm in 5 days. The rainfall amount decreased in the outer areas surrounding Phlangwanbroi, with an average of about 600 mm for the five-day accumulated period (Fig. 2a).
Similarly, Fig. 2d depicts the EPE that occurred in Maharashtra in 2021. The area affected was Wazarwadi, a village in Poladpur taluka within the Raigad district of Maharashtra. A five-day accumulated rainfall of 1395 mm was recorded in Wazarwadi. Notably, the region adjacent to Wazarwadi along the west coast of Maharashtra experienced widespread accumulated precipitation of approximately 500 mm.
Figure 2g shows the EPE over Madhya Pradesh and Rajasthan (CI) on 3rd August 2021. During this event, the highest recorded five-day rainfall occurred in Semli Phata, a small village in Shahbad Tehsil within the Baran district of Rajasthan, with 1054 mm rainfall in 5-days. The surrounding areas also experienced significant precipitation, averaging around 700 mm. The rainfall spread was less in this region and concentrated over a small area. Figure 2j shows an EPE that occurred in the year 2019 over GJ. Gujarat's highest five-day accumulation rainfall was recorded at 366 mm during this event. Similarly, the surrounding areas also received a significant rainfall of approximately 300 mm, with a substantial spread observed in western Gujarat.
Precipitation climatology for the period 1992–2022 of respective days of EPE occurrence is shown in column 2 (Fig. 2b, e, h, and k) for all EPEs. The 5-day climatology varies from 10–100 mm during the 30 years. The highest climatology was observed in the south of NE and CI with a range of 60–100 mm and the lowest being over GJ with 10 mm. Similarly, the 5-day anomalies of EPEs are shown in column 3 (Fig. 2c, f, i and l). As these EPEs led to heavy accumulated precipitation, which exceeded their climatology by 10-20-fold, these anomalies shape, spread, and magnitudes resemble the respective stocked precipitation patterns shown in column 1 for all EPEs. Overall, Fig. 2 highlights the varying intensities and geographical spreads of EPEs in different regions over the years, providing valuable insights into the climatic patterns and anomalies related to such occurrences.
Among the EPE studied, GJ, MH and CI are influenced mainly by the Arabian branch of the summer monsoon, and NE is at the core Bay of Bengal branch. NE receives considerable precipitation in the summer due to its geographical location and the influence of the summer monsoon. Geographically, NE is in the path of the summer monsoon winds, which blow from the Bay of Bengal and the Arabian Sea. These winds carry moisture-laden clouds that result in heavy rainfall when they encounter the mountainous terrain of NE. The topography of NE, characterized by hills and valleys, further enhances the rainfall. The hills act as a barrier, forcing the moist air to rise and cool, forming clouds and precipitation. The same is the case for EPEs over MH, where western Ghats act as gateways for the warm and moist winds from the Arabian Sea.
On the other hand, the valleys provide a favourable environment for the accumulation of moisture, resulting in increased rainfall over MH and NE. The differential heating of land and sea drives the summer monsoon in India. As the landmass of MH, NE, CI, and GJ heats up during the summer, it creates a low-pressure area, which attracts moist air from the surrounding oceans. This convergence of moist air leads to the formation of convective clouds and heavy rainfall over these regions. NE is near the Bay of Bengal, and CI, GJ and MH are close to the Arabian Sea, a warm and moisture-rich body of water. The warm sea surface temperatures and high evaporation rates from these seas contribute to the moisture content of the monsoon winds that reach these areas, resulting in increased rainfall in these regions.
Similarly, we compared the magnitude, spatial extent and spread of daily accumulated precipitation from IMD and ERA5 (supplementary figure S1) for the robustness of the IMD data used in the study. While both datasets show the occurrence of EPEs, the magnitude, location and spread of precipitation in ERA5 was much lower than IMD. Similarly, Fig. 3 shows spatially averaged daily accumulated precipitation for all EPEs. Except over CI, accumulated precipitation from IMD significantly leads the ERA5 during D0 by 80–150 mm. In the case of CI, accumulated precipitation from ERA5 is 80 mm higher than IMD. Among several reasons for the difference in precipitation, models, numerical methods used, and their impact on resolving physics and dynamics of the atmosphere, and data that went into preparing these reanalyses play a crucial role. Hence, it is essential to look after the dynamic and physical processes impacting the EPEs to understand better and forecast them in future. The following section focuses on the dynamical and physical processes to further illustrate the atmospheric variability during these EPEs.
Atmospheric dynamics and EPEs
In this section, we present the aspects of atmospheric dynamics and variables influencing the EPEs. Figure 4 shows all possible atmospheric variables that can affect the precipitation during EPEs occurrence. These variables include area-averaged hourly available potential energy (CAPE), convective inhibition (CIN), heat fluxes, surface temperature, and wind zonal and meridional components compared with daily precipitation magnitudes. Figure 4a shows the day-to-day variability of CAPE for all 4 EPEs. CAPE is primarily a thermodynamic parameter because it quantifies the potential energy available for convection in the atmosphere and represents the vertical buoyant energy available for an air parcel to rise in an unstable atmosphere.
On the other hand, CAPE is also influenced by dynamic factors. It is related to the vertical motion within the atmosphere and how it affects potential energy distribution. The presence of lifting mechanisms, such as fronts, mountains, or convergence zones, can enhance CAPE by aiding the ascent of air parcels. Thus, CAPE combines both thermodynamic and dynamical aspects to provide valuable information about the potential for convective storm development in the atmosphere. It assesses the energy available for convection (thermodynamic) and how the atmosphere's dynamics contribute to this potential (dynamical).
CAPE plays a crucial role in the development of convective precipitation, particularly in the form of thunderstorms. When there is a significant amount of CAPE present in the atmosphere, it indicates that the air is unstable and has the potential for strong updrafts. These updrafts can lead to the rapid ascent of moist air, condensation, and precipitation in the form of heavy rain or hail within thunderstorms. CAPE measures atmospheric instability that can contribute to developing convective precipitation, particularly in thunderstorms. High CAPE values are often associated with more intense and potentially EPEs.
In the case of all EPEs studied, daily CAPE over GJ has increased during D-2 to D0 (1800 J/kg), followed by NE, MH, and CI with values below 1000 J/kg. These values have decreased and show more significant fluctuation after D0. Similarly, CIN, which opposes the CAPE and resembles the atmosphere's stability, decreased over GJ from D-2 to D0, leading to higher instability on D0. While CIN during other EPEs was 10-fold lower (100 J/kg) than that in the case of GJ (1000 J/kg). Similarly, the variability of turbulent fluxes at the surface (SSHF and SLHF) during these EPEs is shown in Figs. 4c and 4d, respectively. SSHF and SLHF can also be considered dynamic and thermodynamic processes. SSF and SLF refer to the transfer of heat energy between the Earth's surface and the atmosphere, typically through conduction and convection. They play a role in the more extensive heat and moisture exchange process between the surface and the atmosphere, affecting weather and precipitation patterns. Both SSHF and SLHF during five days of all EPEs show peaks in the local noon (negative sign represents heat transfer from the surface to the atmosphere due to turbulence and winds). However, low (positive) SSHF during D0 shows heat transfer from the atmosphere to the land caused by the warming in the upper layer of the atmosphere due to condensation and heavy precipitation, as in the case of EPE over CI (25 W/m2). While SSHF over CI remained low/positive, values over other regions show strong diurnal variability. Similar SLHF peaks and dip patterns were noticed during all EPEs, and values over CI show low diurnal variability favouring warmer layers above the surface.
The surface temperature remains high over GJ compared to other locations but dropped 5oC on D0 from a peak of 32oC during D-2. The drop in temperature could be due to cloudiness and less direct solar shortwave radiation reaching the surface (Thandlam et al., 2023b; Thandlam & Rahaman, 2019). The surface temperature over all other EPEs shows less variability between 23-28oC but started to peak during noon from D + 1 (Fig. 4e). These peaks may be subject to incoming solar radiation and clear sky conditions. On the other hand, hourly precipitation over CI shows significant variations with two spells/peaks during D-1 and D0. NE shows persisting rainfall with three peaks during D-1, D0 and D + 1 (Fig. 4f). While hourly precipitation over GJ peaked during D0 and D + 1, MH had no such prominent peaks, and the precipitation rate was stable. Hence, GJ and CI events show a sudden occurrence of heavy precipitation.
Furthermore, the hourly changes in surface winds (zonal and meridional components) are shown in Figs. 4g and 4h, respectively. Prevailing westerlies have peaked during D0 and D + 1 over GJ, CI and MH. Their magnitude was dropped after D + 1. In contrast, NE shows prevailing easterlies until D0, which later changed to westerlies. NE and MH with strong orography, experienced prevailing southerly surface flow for five days. There was a substantial shift in the direction and magnitude of meridional winds over GJ. The flow over GJ turned from southerly to northerly on D0 and back after D + 1. The intensity of northerlies over CI has increased to 4 m/s on D0 and changed to southerlies on D + 1. Also, changes in zonal and meridional winds along the vertical columns for all EPEs are shown in Figs. 5 and 6, respectively. NE and MH show strong and established westerly flow from 900 hPa. However, the westerlies over CI and GJ are mostly confined to surface below 800 hPa and peaked at 400 hPa and 600 hPa, respectively, during D0. A robust northerly flow in the upper atmosphere (up to 200 hPa) during D0 was noticed for MH, CI and GJ, which later (during D + 1 and D + 2) decreased in magnitude and direction. At the same time, NE shows prevailing southerlies (in 900 − 600 hPa) for 5-days.
These changes in the surface winds are further explored concerning circulation, vertical velocity, and convergence/divergence at different layers of the atmosphere during all EPEs.
The vertical structure of the divergence/convergence during EPEs is shown in Fig. 7. All EPEs are associated with a robust negative divergence (convergence) on the surface and upper-level positive divergence. Except over GJ, a mid-level trough/ridge with substantial convergence/divergence was formed during D0. In the case of GJ, the prevailing convergence has intensified on D0, leading to strong convection. The more substantial the divergence of air in the upper troposphere, the more significant air is forced to rise. The shape of a 500 mb trough indicates dynamical strength, i.e., its potential to force strong rising motion in the atmosphere and, hence, heavy precipitation. Divergence in the upper atmosphere can lead to rising air at the surface (convergence).
As air rises, it cools and expands, condensing water vapour and precipitation and providing orographic and favourable local conditions (Rotunno & Houze, 2007). The vorticity shown in Fig. 8 for all these EPEs shows the robustness of the divergence/convergence discussed above. In regions with high vorticity, it can lead to the convergence of air masses. High vorticity can enhance the lifting of air, which is essential for rain or snow to form. In mountainous regions such as MH and NE, vorticity can be generated by the terrain's interaction with prevailing winds (Houze, 2012). This vorticity can induce orographic lift, forcing moist air to rise over the mountains. Thus, vorticity can impact precipitation by influencing air convergence enhancing lifting mechanisms. A prevailing and enhanced vorticity was found above 900 hPa over NE and CI, supporting a continuous moisture supply into the atmosphere. Prevailing vorticity near the surface over MH was enhanced on D0, leading to heavy precipitation. A strong and elongated patch (surface to 200 hPa) of vorticity noticed during EPE over GJ on D0 indicates the strong local convection and low pressure supporting strong convergence and heavy precipitation.
The vertical velocities shown in Fig. 9 also depict the convergence and vorticity described above, along with the instances of heavy precipitation. All EPEs show moderate to strong negative vertical velocities (as per ECMWF convention) during D0, leading to the stronger upward motion of wet and warm air and followed by a stronger subsidence (positive vortical velocity) after heavy precipitation. Notably, rapid subsidence with strong positive vertical velocities were noticed over GJ and CI, where the sudden occurrence of EPEs took place, and the subsidence was moderate for MH. Hence, a strong positive vertical velocity shows subsidence and inhibits the formation of clouds, leading to clear sky conditions and surface temperature rise due to direct insolation.
Atmospheric Thermodynamics and EPEs
In this section, we explored the impact of atmospheric thermodynamical variables on the occurrence of EPEs. The thermodynamic variables considered are RH, IWT and MSE. The atmosphere is more than 80% saturated during all EPEs. Higher relative humidity increases the chances of precipitation. However, it is essential to note that other factors, such as atmospheric instability, air temperature, and cloud condensation nuclei, also play a role in determining whether precipitation will occur. Furthermore, variability in relative humidity also affects the amount of rainfall received; rainfall is also maximized when relative humidity is high (above 80%). Interestingly, the subsidence of dry air after the heavy rainfall occurrence after D + 1 has reduced the RH to below 60% in the upper atmosphere (700 − 200 hPa) during all EPEs and formed a large region with less RH.
On the other hand, the vertical structure of the WVT shown for all EPEs in Fig. 11, derived from Eq. 1, shows the horizontal transport of water vapour (specific humidity) by the zonal and meridional winds. Water vapor transport was characterized via the specific humidity flux index (kg·kg− 1m·s− 1) calculated using (1), specific humidity q, and zonal (u) and meridional (v) wind component values for every available vertical level up to 100 hPa. The advective flux of specific humidity (kg·kg− 1m·s− 1) was calculated for the zonal direction (u. ∆qx) and for the meridional direction (v. ∆qy), where u and v are the horizontal wind speeds in the respective directions and ∆qx and ∆qy are the horizontal humidity gradients in the respective directions (Wypych & Bochenek, 2018).
Water vapour is the source of moisture for precipitation which can increase when a significant amount of moisture is transported into an area. Elevated levels of water vapour transport can result in heavy rainfall, especially if the moisture-laden air encounters a front or weather system that forces it to rise and cool, leading to condensation and precipitation (Houze, 2012; Kirshbaum & Smith, 2008; Rotunno & Houze, 2007). Also, moisture convergence from water vapour transport can contribute to developing extreme precipitation events. Except in the case of NE, where the source of water vapour is from the Bay of Bengal, all EPEs that occurred over other regions show significant advection of water vapour in the mid-tropospheric region (900 − 600 hPa) favouring the abundance of water vapour to form clouds and precipitate, providing favourable atmospheric conditions. The WVT was larger during D0 for the EPEs over GJ, CI and WI. However, the WVT over GJ is more concentrated at D0 and could be due to dominated local convection than a more prevailing concentration over CI and WI.
$$WVT=\sqrt{{\left(uq\right)}^{2}+{\left(vq\right)}^{2}}$$
1
Similarly, we also studied the vertical variability of MSE during all EPEs, as shown in Fig. 12. Moist static energy is defined as the sum of the potential, kinetic, and latent heat energy in an air parcel per unit mass. It can be expressed as in (2):
MSE = Cp * T + gz + Lvq (2)
Where Cp is the specific heat capacity of dry air; T is the temperature of the air parcel in Kelvin. g is the acceleration due to gravity. z is the height of the parcel above a reference point. Lv is the latent heat of vaporization. q is the parcel's specific humidity (mass of water vapour per unit mass of dry air). In extreme precipitation, MSE indicates the available energy for convection and can help meteorologists understand the potential for severe weather events (Marquet, 2015; Pauluis, 2007). Rising MSE can signify the intake of warm and moist air, which, when lifted, can release latent heat energy through condensation, contributing to heavy rainfall and extreme precipitation.
The MSE over NE during EPE shows a sustained 600 hPa barrier of lower MSE values trapped between higher values above and below in the atmosphere. Though mid-atmospheric MSE is lower before and after D0, for all other EPEs over GJ, CI and MH, a peaked MSE during D0 shows the strong availability of potential and kinetic energy in the atmosphere along with the enhanced latent heat fluxes (~ 350 kJ/kg) leading to heavy precipitation to occur. Hence, all the thermodynamical variables show a clear signature of enhancement during D0, resembling the rapid convection and increased instability in the atmosphere leading to the EPE.