1. Introduction
Drought is a natural phenomenon that often impacts water resources and socioeconomic activity on a regional scale anywhere in the world (Mishra and Singh 2010; Sheffield and Wood 2011; Markonis et al. 2021). In contrast to other natural disasters, drought incidents progress gradually in time, but their impacts generally span a long period of time. Hence, the drought analysis and its prediction will have significant implications for successful national policies and for alleviating drought impacts (WMO 2006). Drought monitoring and prediction is of utmost importance for the countries like India, where the rainfall is seasonal in nature. While agriculture is tuned to rainfall occurrence, any deficiency in rainfall will have a paramount impact on the Indian economy. The active and break conditions in the southwest monsoon (June–September) and northeast monsoon (October–December) are the key periods for monsoon drought examination on the Indian subcontinent. Numerous studies (Pai et al. 2011; Niranjan Kumar et al. 2013; Guhathakurta et al. 2017; Mishra et al. 2022) have focused on understanding climatic trends and spatial–temporal variability of monsoon droughts in India, particularly during the southwest monsoon. This season is critical as it contributes 70%–90% of the annual rainfall, impacting the Indian economy (Gadgil 2003; Niranjan Kumar et al. 2013; Mishra et al. 2022). Between 1871 and 2015, India experienced approximately 25 major drought years, with their frequency increasing over time (Niranjan Kumar et al. 2013; Guhathakurta et al. 2017; Mishra et al. 2022). The severe 2002 drought, one of the worst in this century, resulted in a 19% countrywide rainfall deficiency, affecting 300 million people across 18 states and causing a 13.4% drop in food grain production (Sarma 2004; Sarkar 2011). A similar situation occurred during the 2009 summer monsoon, with a 22% below-normal rainfall deficiency (Neena et al. 2011) and a 6.9% decline in food grain production (Sarkar 2011).
Drought recurrence in India is primarily attributed to delayed onset of rains, prolonged monsoon breaks, early withdrawal of the monsoon, and irregular distribution of rainfall. The India Meteorological Department (IMD) declares an “all-India drought” when seasonal rainfall is below −15% (Saith and Slingo 2006). However, recent research has emphasized that assessing drought for the entire season may not capture the full impact, as subseasonal variations significantly affect water resources and agriculture (Bhat 2006). Notably, severe meteorological droughts in India, like those in 1923, 2002, 1937, and 1907, often concentrated in specific months, such as July in the case of 2002, accounting for a significant portion of the seasonal deficit (∼56%) (Bhat 2006). The 2009 monsoon also started with substantial deficits in June (48%) and August (27%). These rapidly intensifying subseasonal droughts pose a significant threat to agriculture during the peak summer monsoon period. In the recent period of August 2023, central India and the southern peninsular region experienced a record-low monsoon rainfall, marking the driest month in 122 years since 1901. This historic deficit was characterized by two distinct monsoon breaks during 5–16 and 27–31 August, with the monsoon trough consistently positioned northward, resulting in unfavorable conditions for plains and crop cultivation.
Hence, many recent studies have recognized the importance of these droughts that develop on the Subseasonal to Seasonal Prediction project (S2S) time scale (weeks to months), posing a new challenge for prediction efforts on that time scale (Pendergrass et al. 2020). The potential implications of these subseasonal droughts over the Indian subcontinent have also been assessed in recent studies (Mishra et al. 2021; Mahto and Mishra 2020). It is also observed that more than 82% of the subseasonal droughts are found during the summer monsoon period over the Indian region (Mahto and Mishra 2020). The mechanisms underlying these droughts are multifaceted, encompassing internally driven processes linked to intraseasonal oscillations and the dynamics of monsoon–midlatitude interactions facilitated by the intrusion of Rossby waves (Krishnamurti et al. 2010; Krishnan et al. 2000, 2009; Goswami 2005).
Monsoon droughts are influenced by external factors like El Niño–Southern Oscillation (ENSO) and the Indian Ocean (Niranjan Kumar et al. 2013). The prevailing view attributes monsoon droughts to basinwide warm sea surface temperature (SST) anomalies in the equatorial Pacific associated with ENSO, but recent research challenges this perspective, noting that monsoon droughts can occur even when Pacific SSTs are near-neutral (Borah et al. 2020). It has been observed that monsoon droughts are also linked to subseasonal phenomena, driven by North Atlantic SSTs and mediated by Rossby waves. Predicting droughts at subseasonal scales is a complex task, encompassing time frames from as short as a week/month to as long as decades. In this study, we evaluate the performance of the National Centre for Medium Range Weather Forecasting (NCMRWF) Extended Range Prediction (NERP) system, based on the unified global coupled modeling system, for characterizing subseasonal droughts. This article introduces crucial datasets, explains the modeling process for NERP hindcasts, and discusses the drought index in section 2 and presents results and discussion in section 3. Section 4 summarizes the findings and conclusions.
2. Data and methods
a. NERP system
The NERP model configuration is based on the Met Office GloSea5 seasonal prediction system (MacLachlan et al. 2015) in Global Coupled 2.0 configuration (GC2.0; documented in Williams et al. 2015). The model is fully global coupled S2S ensemble system (Gupta et al. 2019a,b; Gera et al. 2021) consisting of Global Atmosphere 6.0 (GA6.0) and Global Land 6.0 (GL6.0) (Brown et al. 2012; Walters et al. 2017), Global Ocean 5.0 (GO5.0) (Megann et al. 2014), and Global Sea Ice 6.0 (GSI6.0) (Rae et al. 2015). The atmospheric model is nonhydrostatic and fully compressible that uses Arakawa C grid discretization on horizontal and Charney–Phillips vertical staggering with semi-implicit semi-Lagrangian discretization to solve equations of motion. The model has a terrain-following hybrid-height coordinate system with a horizontal resolution of 0.833° × 0.556° having 85 vertical levels covering up to 85 km. The land surface model (GL) is based on the Joint UK Land Environment Simulator (JULES; Best et al. 2011). The horizontal grid resolutions are the same as the atmospheric model but with four vertical levels and nine vegetation types. The NERP system uses the Nucleus for European Modelling of the Ocean (NEMO; Madec 2008; Blockley et al. 2013) model for its ocean component having an eddy-permitting horizontal resolution of 1/4° on a tripolar orthogonal curvilinear grid called ORCA02. The NEMO has 75 vertical levels reaching a depth of 6000 m with 1-m vertical resolution in the top 10 m of the upper ocean. The Los Alamos Sea Ice Model (CICE; Hunke and Lipscomb 2010) is employed in the NERP system that was developed in parallel with GA6.0/GL6.0 as part of GC2.0 configuration (Walters et al. 2017). The sea ice model and ocean exchange information at 3-h intervals with the other model components via the OASIS flux coupler. To resolve subgrid scale process and grid-level perturbations in the course of model integration, a stochastic kinetic energy backscatter v2 (SKEB2) method is employed (Bowler et al. 2009). Note that the SKEB2 scheme perturbs only atmospheric conditions, not the SSTs.
In this study, we employed the NERP hindcasts spanning 23 years from 1993 to 2015. For our hindcast simulations, we utilized initial conditions sourced from the European Centre for Medium-Range Weather Forecasts (ECMWF) interim reanalysis (ERA-Interim; Dee et al. 2011) data, while the Met Office GloSea5 reanalysis was consulted for ocean and sea ice data (MacLachlan et al. 2015). The ERA-Interim data, featuring 60 vertical levels and a 0.75° horizontal resolution, were reconfigured to match an L85 vertical grid (up to 85 km) and N216 regular grid resolution to facilitate the atmospheric model’s execution (Gupta et al. 2019a,b). The generation of hindcasts was rounded to six ensemble members, derived from lagged initial conditions spanning three preceding days. Each day’s conditions led to the creation of two ensemble members (Gera et al. 2021), resulting in a total of six members within the ensemble. Our analysis was based on hindcast data that initiated from a fixed start date, specifically the first day of each month. The model’s forecast covered a duration of the next 36 days, and we utilized the ensemble mean rainfall to calculate the standardized precipitation index (SPI) drought indicator at a monthly scale. For validation purposes, we compared our hindcast simulations with IMD gridded rainfall data (Pai et al. 2014). This dataset features a spatial resolution of 0.25° and encompasses daily rainfall records dating back to 1901. It has been meticulously adjusted to account for various factors influencing rainfall distribution across India, including orographic influences. The validation of the radiation and dynamical fields generated by the model is conducted by comparing it to two reputable sources: the National Oceanic and Atmospheric Administration (NOAA) interpolated outgoing longwave radiation (OLR) data (Liebmann and Smith 1996) and the ECMWF reanalysis (Hersbach et al. 2023). In the subsequent subsection, we provide a brief overview of the methodology employed to compute the SPI index.
b. SPI
Drought can be characterized by distinct indicators that are basically estimated using different hydrometeorological variables. Among several proposed drought monitoring indices based on meteorological, agricultural, hydrological, etc., the SPI (McKee et al. 1993) is used widely since it only involves precipitation for its computations. Moreover, the precipitation information is commonly accessible in any location via at-site observations or remote sensing estimations. Furthermore, the results based on SPI can be evaluated between different locations while conducting spatiotemporal analyses (McKee et al. 1993). Another element that makes SPI for its widespread application in drought monitoring is its ability to describe both short-term and long-term drought impacts through different time scales and regions with different climatic conditions. In principle, SPI is also useful to monitor floods in areas where scant hydrological data are available (Yuan and Zhou 2004) and can be used to quantify food grain productivity (Mishra and Cherkauer 2010; Patel et al. 2007).
Hence, SPI is widely used for drought assessment and real-time drought monitoring over the Indian subcontinent. For instance, the IMD uses the SPI for real-time drought monitoring and forecasting (https://imdpune.gov.in/imdone/caui.php), along with several other indices such as the aridity anomaly index (Thornthwaite 1948) and a standardized precipitation and evapotranspiration index (SPEI; Niranjan Kumar et al. 2013; Dhangar et al. 2019) from weekly to seasonal time scales. Shah and Mishra (2015) developed an experimental near-real-time SPI drought monitor using the Tropical Rainfall Measuring Mission (TRMM) rainfall data, which provided high-resolution drought information (district level) over India. SPI is one of the better indices for districtwide drought monitoring and is appropriate for investigating the active and break events over the Indian region (Pai et al. 2011; Mishra et al. 2022).
The SPI calculation involves fitting a gamma distribution to the historical precipitation record at each grid location. This distribution is then transformed into a normal distribution, resulting in a mean SPI value of zero for the desired location and period (McKee et al. 1993). Positive SPI values indicate wet conditions, while negative values indicate dry conditions. The magnitude of the positive or negative SPI values reflects the degree of deviation from the normal conditions. The SPI’s normalization allows for the representation of both wetter and drier climates in a consistent manner. This means that wet periods can also be monitored using the SPI. Additionally, the standardized SPI values have the same probability of occurrence, regardless of the time period, location, or climate. In the case of the NERP forecast data, the monthly sums are calculated by aggregating the daily values, considering that the simulations run effectively from the first of each month until 36 days. For a more comprehensive understanding of the SPI methodology and computational procedure, further details can be found in the works of McKee et al. (1993, 1995) and Edwards and McKee (1997).
3. Results and discussion
Figure 1 displays the monthly climatological rainfall distribution (June–August) using IMD gridded data and NERP hindcasts, offering an assessment of the model’s performance in terms of climatological mean at subseasonal scales. Most of the rainfall is observed over the Western Ghats, the core monsoon zone (CMZ; 18°–28°N and 65.0°–88.0°E), and the northeastern parts of the Indian subcontinent. Conversely, the lowest rainfall occurs in the extreme northwestern region and the extreme southeastern parts of the southern peninsula, commonly known as the rain-shadow region. The CMZ receives an average rainfall amount ranging between 15 and 20 mm day−1 during the core monsoon months of July and August. However, the areas with rainfall exceeding 20 mm day−1 are primarily along the Western Ghats and parts of northeastern India. The NERP model successfully captures the spatial pattern of the observed climatology (Fig. 1b). However, the model’s predicted rainfall magnitudes are generally higher than the observed values, particularly over the foothills of the Himalayas and the Indo-Gangetic Plain (IGP) region. In contrast, the model rainfall is lower than observed over the CMZ and along the west coast.
Climatological mean monthly rainfall over India from the (a) IMD observed data and the (b) NERP model hindcast runs for 1993–2015.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
To quantify the differences between the model hindcast and observed rainfall, we have calculated the relative mean error (ME), as shown in Fig. 2. The relative ME reveals a significant dry bias along the Western Ghats, with magnitudes of approximately 10 mm day−1. A smaller dry bias (<4 mm day−1) is observed over the CMZ. However, during the core monsoon months of July and August, a wet bias can be observed in certain parts of the CMZ. The wet bias (>4 mm day−1) is more pronounced in parts of northeastern India, including the foothills and the IGP region. The biases in the orographic rainfall belts in NERP hindcasts can be attributed to factors such as the diurnal cycle of convection processes, the horizontal resolution of the model, and the representation of topography (Gera et al. 2021). Additionally, a significant wet bias is noticed in parts of the western equatorial Indian Ocean (EIO), extending to the eastern EIO from week 1 to week 3 forecasts (Gera et al. 2021). This widespread equatorial convection could result in subsidence over the CMZ, leading to a dry bias in the model simulations. To substantiate our previous arguments, we examined the climatological rainfall time series averaged over the CMZ region from both model simulations and observational records for each month (Fig. 2e). A notable increase in dry bias with extended forecast lengths can be seen from Fig. 2e. The bias is most pronounced in June and July, while in August and September, the bias is less conspicuous.
(a)–(d) Monthly ME of NERP model hindcast rainfall predictions against IMD’s gridded data from 1993 to 2015. Regions with a significance exceeding the 95% confidence level are indicated by stippling. (e) Climatological (1993–2015) rainfall time series for each month area averaged over the CMZ.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
The CMZ holds significant importance in understanding interannual rainfall variability during the summer monsoon months due to its strong correlation with all-India summer monsoon rainfall (Rajeevan et al. 2010). Notably, the rainfall in August closely mirrors that of the entire country. Thus, we evaluate the SPI drought index at an interannual scale for different summer monsoon months in the CMZ region. This analysis helps mitigate observational biases caused by limited rainfall measurements, especially in complex orographic regions and remote parts of North India. Figure 3 presents the average SPI values over the CMZ for different months based on both IMD observations and NERP hindcasts. The correlation of the drought index between IMD and NERP hindcasts is significant (>0.7), particularly for August and September compared to June and July. The NERP system effectively captures the interannual variability of drought, with better prediction of high-frequency variability compared to low-frequency variability (Fig. S1 in the online supplemental material). It is important to note that these simulations are initiated every month on the first day and do not represent long-term seasonal or climate simulations that may better capture low-frequency variability. Additionally, biases associated with initial conditions can significantly impact the magnitude of SPI, as evident in Fig. 3.
Mean SPI values within the CMZ (18°–28°N and 65.0°–88.0°E) during (a) June, (b) July, (c) August, and (d) September, illustrating the interannual variation of droughts at subseasonal scales over a period of 1993–2015.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
The NERP system demonstrates a tendency to overestimate wet years with positive SPI values, particularly during June and July, when compared to observed data from IMD. Conversely, it tends to overestimate dry years with negative SPI values, especially in August and September. Interestingly, NERP simulations exhibit a significant correlation with IMD observations during the 2002 drought, which was primarily characterized by an extended dry period in July. However, it is also worth to note the overestimation of drought in the NERP system during the months of July and September 2015 (elaborated in section S5 in the supplemental material for comprehensive insights). Another noteworthy deficit in summer monsoon rainfall occurred during June–September (JJAS) 2009, as documented by previous studies (Neena et al. 2011; Francis and Gadgil 2010; Preethi et al. 2011). This period experienced two prolonged dry spells: 1 June 2009 (with 47% below-normal rainfall) and another from the end of July to the second week of August 2009 (with 27% below-normal rainfall). Figure 3 demonstrates that the NERP system fairly predicted the all-India drought episode in June 2009, while it overestimated the August 2009 episode.
Additionally, Fig. S4 (supplemental material) evaluates the accuracy of NERP hindcasts (1993–2015) by correlating SPI values with observed SPI from IMD gridded data for distinct summer monsoon months. To further assess the spatial distribution, we examine two drought cases in 2002 (July) and 2009 (June). In Figs. 4a and 4b, the July 2002 SPI drought index based on IMD gridded rainfall and NERP hindcasts is displayed. Both the model and observations indicate extreme drought conditions (SPI < −2.0) in July 2002, particularly over the CMZ (Figs. 4a,b). However, NERP hindcasts depict widespread extreme drought conditions across large regions of India relative to IMD observations. In the case of June 2009, except for the southern peninsular region, much of the all-India region is under severe drought with magnitudes less than −2 in IMD data (Fig. 4c). While the intensity is comparatively less in the NERP-predicted drought intensity, except over the foothills of the Himalayas. Thus, there is notable spatial heterogeneity in both cases even though the time series of the all-India drought is fairly represented in Fig. 3 for both cases. Nonetheless, the NERP system can predict all-India drought conditions at subseasonal time scales albeit with some differences in location and intensity. Additionally, please refer to supplemental material section S8 for further discussion on the model performance for all months where SPI has an absolute magnitude of 1 or more. It is important to note that drought episodes vary spatially for each season/month. Therefore, composite maps may not always provide a comprehensive assessment.
(a),(b) Spatial distribution of SPI drought during July 2002 from IMD gridded rainfall and NERP model. (c),(d) As in (a) and (b), but for June 2009.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
a. Subseasonal droughts triggered by tropical teleconnections
The differences in simulating and predicting subseasonal monsoon droughts have highlighted the need for a better understanding of the underlying internal dynamics of the summer monsoon droughts in NERP. However, it is well-established through theory and observations that the large-scale “external” forcing associated with ENSO is often accompanied by monsoon droughts (Prasanna and Annamalai 2012; Niranjan Kumar et al. 2013). Other factors, such as the Indian Ocean dipole (IOD) and local atmospheric variability, can also modulate the impacts of ENSO on the Indian monsoon and further influence the occurrence and severity of droughts. Nevertheless, it is important to recognize that the occurrence of droughts is a complex and multifaceted phenomenon influenced by a combination of internal and external factors. For instance, earlier reports showed that a large number of monsoon droughts in the last century are unrelated to ENSO (Neena et al. 2011; Borah et al. 2020). Hence, even though ENSO is a prominent external forcing mechanism, other internal driving factors of the monsoon system can also contribute to subseasonal droughts. The eastward-propagating Madden–Julian oscillation (MJO) and northward-propagating monsoon intraseasonal oscillation (MISO) are some of the well-known mechanisms through which the active/break cycles of the Indian summer monsoon (ISM) are modulated subsequently causing the occurrence and persistence of drought conditions. The passage of the MJO can bring a prolonged period of suppressed rainfall, exacerbating existing drought conditions in several geographic locations (e.g., Joseph et al. 2009; Neena et al. 2011). The noted differences in the location and intensity of drought conditions between the model and observations in Fig. 4 might also be associated with the position and phase of the drivers such as MSIO and MJO.
Figure 5 presents an assessment of the poleward propagation characteristics of the MSIO using OLR based on NOAA and NERP simulations during the drought years 2002 and 2009. The phase propagation of the MSIO is obtained by applying a 30–90-day bandpass filter to the OLR data, and a Hovmöller diagram is plotted with latitude on the y axis and time on the x axis, averaging over longitudes between 70° and 85°E. In the year 2002, it is important to note that the location of the MSIO is relatively consistent in both NOAA and NERP hindcasts. However, the phase speed of the MSIO in NOAA is quicker compared to the NERP system. This difference can be observed in July 2002 in Figs. 5a and 5b. Nonetheless, over the CMZ latitudes, a positive OLR anomaly is clearly visible. Therefore, the drought conditions in July 2002 are well-predicted in NERP hindcasts. During 2009, significant differences in the location and MSIO phase propagation in OLR were observed between the NOAA and NERP systems. Furthermore, in contrast to the MSIO propagation speed in 2002, NERP hindcasts show a faster propagation than NOAA. Specifically, when examining the subseasonal drought episodes in June and August 2009, an extended positive OLR anomaly is noticeable in NERP hindcasts over the central and northern Indian latitudes. This supports the observed differences in prevailing drought conditions shown in Figs. 5c and 5d. Hence, the disparities in the location and phase speed of the MSIO have significant implications for the observed differences in drought conditions in NERP hindcasts (refer section S5 of the supplemental material for more details).
(a),(b) Time–latitude Hovmöller diagram of 30–90-day bandpass-filtered daily OLR anomaly (W m−2) averaged over 70°–85°E for the period 1 May–30 Sep 2002 from NOAA OLR and the NERP model. (c),(d) As in (a) and (b), but for the year 2009. Solid contours represent a positive anomaly, and dashed contours represent a negative anomaly. Contours are drawn from −30 to 30 with a uniform interval 6.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
Another critical factor contributing to observed and model-predicted drought conditions is the influence of the eastward-propagating MJO. This complex phenomenon involves interactions among atmospheric circulation, convection, and moisture transport. Accurately capturing the MJO in models is challenging due to its subseasonal nature and complex dynamics (Ahn et al. 2017). While the MJO is highly prominent in the equatorial region during the Northern Hemisphere winter, it becomes relatively weaker during the boreal summer. Despite being less pronounced during this season, the MJO sporadically manifests over the summer monsoon regions. Monsoon droughts are also associated with the eastward-propagating MJO in the equatorial Indian Ocean and westward-propagating Rossby waves between 10° and 25°N (Wang and Xie 1997). These westward-propagating Rossby waves toward the Indian land play a crucial role in creating and sustaining conditions necessary for the monsoon break, subsequently leading to drought conditions. In this study, we assess subseasonal teleconnection links in the NERP hindcasts, along with reanalysis winds and observed OLR (a proxy for rainfall), to examine the geographical distribution and circulation anomalies associated with MJOs and their impacts on drought.
The analysis employs the methodology formulated by Wheeler and Hendon (2004), which is constructed using combined empirical orthogonal function analysis of various atmospheric variables [850-hPa zonal wind (u850), 200-hPa zonal wind (u200), OLR]. The u850 and u200 winds are obtained from ERA-Interim reanalysis as these data are used for the initialization of NERP hindcasts, alongside NOAA OLR data. This approach will enhance our ability to compare with model data and effectively track and characterize the behavior of the MJO, achieved through plotting it on a phase-space diagram as described by Wheeler and Hendon (2004). Figure 6 shows the MJO phase-space diagram for the severe drought months of July 2002 and June 2009 is shown from both reanalysis and model hindcasts. In comparing two distinct scenarios, one involving observed data and the other centered on hindcast simulations, it becomes evident that the phase propagation of MJO convection remains consistent across both cases, indicating a commendable level of agreement between the observed data and the modeled hindcasts. However, the analysis also uncovers discernible disparities in the amplitudes of the MJO signal as the model integration progresses (also refer to supplemental material section S6 for more details). This intriguing aspect suggests the dynamic nature of the MJO phenomenon and its sensitivity to the underlying model dynamics. These differences in amplitude prompt a deeper investigation into the underlying mechanisms that drive the MJO convection and its complex interactions within the model framework. Nevertheless, despite the observed amplitude variations, the overarching conclusion drawn from this analysis is that the MJO phase finds substantial representation within the NERP hindcast simulations. This alignment in phase representation showcases the modeling capabilities of the NERP system in capturing the fundamental temporal evolution of the MJO, thus adding to the credibility of the hindcast simulations during drought conditions over the Indian subcontinent. A composite assessment of MJO phases between the ERAI and NERP hindcasts in subseasonal scales is discussed in section S7 in the supplemental material.
(a) MJO life cycle in phase space during July 2002 from ERA-Interim (blue) and NERP model (red). (b) As in (a), but for June 2009. Please refer to Fig. S7b for a detailed depiction of the MJO propagation throughout the June–September period.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
The interaction between the atmosphere and the sea plays a major role in the monsoon variability across the Indian subcontinent at intraseasonal time scales by influencing the intensity and propagation of disturbances such as the MSIO and MJO. Consequently, it is essential to evaluate model biases in simulating these air–sea interactions. This evaluation entails examining the relationship between SST and precipitation on an intraseasonal scale, comparing observations with model outputs. For instance, in Figs. 7a–c, we present lead–lag correlations between precipitation anomalies and SST anomalies over the Arabian Sea (AR), Bay of Bengal (BoB), and the South China Sea (SCS) spanning 20 days before and after. A positive (negative) correlation when precipitation lags (leads) SST anomalies indicates an ocean-to-atmosphere (atmosphere-to-ocean) influence, suggesting either the SST driving the atmosphere or vice versa, respectively (Roxy et al. 2013; Konda and Vissa 2021). The strength of this correlation reflects the intensity of the driving force, while the lag (lead) time indicates the rate at which the atmosphere responds to SST anomalies and vice versa. Observationally, precipitation lags SST by approximately 10 days over the AR, whereas the lag is around 12 days over the BoB and SCS. This suggests a relatively faster ocean-to-atmosphere effect over the AR compared to the other regions. Within the NERP system, strong SST forcing is evident over the AR region, comparable to the BoB region, while the SCS shows a weaker SST forcing. Nonetheless, the NERP system demonstrates a quicker atmosphere response to ocean forcing, as indicated by the shorter lag relative to observations. Conversely, the atmosphere-to-ocean effect appears slower over the AR (around 12 days), with a faster response observed over the BoB and SCS (approximately 10 days). These distinct SST–precipitation relationships across different basins align with previous research findings (Roxy et al. 2013).
(a)–(c) Lead-lag correlation of precipitation anomalies with respect to SST anomalies, averaged over intraseasonal time scales and focused on the AR, BoB, and SCS regions as specified. (d)–(f) Hovmöller plots illustrating observed intraseasonal anomalies of SST (shading) and precipitation (contour levels displayed at 1 and −1 mm day−1) over the AR, BoB, and SCS regions, respectively, relative to the SST maximum at day = 0, for both observational data and the (g)–(i) NERP model. Regions with a significance exceeding the 95% confidence level are indicated by stippling.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
Figures 7g–i present a detailed assessment of the NERP system’s performance at intraseasonal scales relative to observations (Figs. 7d–f) using Hovmöller plots, illustrating the temporal evolution of SST and precipitation anomalies concerning an SST maximum (exceeding one standard deviation) across the AR, BoB, and SCS regions. Across all cases, a coherent northward propagation of anomalies is evident, albeit with variations in phase propagation, particularly noticeable over the AR region where the model’s phase propagation (Fig. 7g) appears relatively faster compared to observations (Fig. 7d). In each region, positive SST anomalies precede positive precipitation anomalies, suggesting a reasonably accurate representation of the SST–precipitation relationship. These regions play a critical role in influencing active and drought situations across the Indian subcontinent through the response of the poleward-propagating MSIO from the equatorial Indian Ocean. While a comprehensive analysis falls beyond the scope of this study, the results suggest a generally adequate representation of air–sea interactions.
Furthermore, a comprehensive analysis encompassing radiation and dynamical processes using the composite analysis (section S2 in the supplemental material) during drought occurrences yields a deeper understanding of teleconnections, thereby elucidating the underlying physical mechanisms at subseasonal scales. Illustrated in Fig. 8, our presentation initiates with a composite analysis of OLR during observed drought conditions extracted from the NOAA dataset, in conjunction with NERP hindcasts. Noteworthy is the fact that the drought conditions under examination belong to instances of prolonged dry conditions, commonly referred to as “long break” conditions. Extensive prior research has developed on the role of ocean–atmosphere coupling in instigating such extended breaks, consequently ending in drought episodes (Krishnan et al. 2006; Saith and Slingo 2006; Mohan et al. 2018). In particular, Joseph et al. (2009) have postulated that the genesis of extended monsoon break conditions is intertwined with air–sea interactions, while the eastward-propagating MJO in the equatorial Indian Ocean promotes their persistence, thereby contributing to drought development.
Composite anomalies of OLR during drought conditions over the CMZ for JJAS, spanning the years 1993–2015. (a)–(d) NOAA interpolated OLR data. (e)–(h) NERP hindcast information. Regions with a significance exceeding the 95% confidence level are indicated by stippling.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
Thus, we evaluated a composite analysis of a month-long drought (SPI drought index calculated on a monthly scale aggregating the daily rainfall; please refer to section 2b for more details) conditions spanning 1993–2015. In Fig. 8 (left panel), we examine observed OLR anomalies during subseasonal drought conditions across various summer monsoon months (June–September). Evidently, a distinct pattern emerges, marked by suppressed convection indicated by positive OLR values. This suppression is prominent over central and north-northwest India, the western coast, and the northern Arabian Sea regions. Analyzing the geographic distribution of these OLR anomalies provides insights into the interplay between regional anomalies during drought conditions and global phenomena at subseasonal scales. Notably, Fig. 8 (left panel) highlights a significant feature during drought conditions over the Indian subcontinent. This feature is associated with organized suppressed convection, forming a northwest–southeast (NE–SW) oriented band stretching from continental India to maritime territories. This well-defined pattern, documented by earlier studies (Wang and Xie 1997; Kemball-Cook and Weare 2001), suggests a link to suppressed convection dynamics. Additionally, Fig. 8 (left panel) reveals positive OLR anomalies over the Indian region and maritime territories, juxtaposed with negative anomalies over the eastern equatorial Indian Ocean (IO) and China. This intriguing pattern of positive–negative–positive–negative is a recognized “quadruplet” structure (Annamalai and Slingo 2001). The composite depiction of observed drought scenarios also unveils a symmetrical arrangement of suppressed and enhanced convection across the equator. This symmetrical behavior might imply Kelvin wave dynamics.
Meanwhile, the tilt of suppressed convection from maritime territories toward the Indian region could be attributed to Rossby wave dynamics, aligning with previous studies (Krishnan et al. 2000). These salient characteristics are represented convincingly within the NERP hindcast simulations (Fig. 8, right panel). Notably, the NE-SW tilted band, coupled with suppressed and enhanced convection over the equatorial IO, is remarkably well replicated. Another critical factor contributing to subseasonal drought development is the enhanced convection observed in the eastern equatorial Indian Ocean (EEIO), particularly noted in June and July (Figs. 8a,b). This increased convective activity over the EEIO leads to weakened monsoon circulation due to induced subsidence over the subcontinent (Krishnan et al. 2006). The NERP model effectively captures this coupled ocean response during drought conditions (Figs. 8e,f), albeit with some enhanced convection in June compared to observations. Nevertheless, the distinctive teleconnection features discussed above find faithful representation within the NERP system. Nevertheless, it is crucial to acknowledge that specific variations, like the geographical distribution and severity of droughts indicated by positive OLR anomalies, may vary across distinct months. The response of large-scale convective anomalies in the tropical western Pacific region, indicated by the negative OLR anomalies in Fig. 8, holds significant implications for drought development through widespread subsidence over the Indian subcontinent. This is further supported by the composite spatial maps of divergent circulation patterns and velocity potential from ERA5 and NERP system across different summer monsoon months (please refer to section S3 in the supplemental material for more details).
A more profound understanding of the disparities within the model hindcast will be elaborated in the context of anomalous circulation patterns (vectors) combined with specific humidity (q) distributions (shading) at the 850-hPa level, particularly amid observed drought conditions. These drought conditions are characterized by dry conditions with low humidity levels accompanied by negative anomalies. Such patterns are distinctly observable over North India and the northern Arabian Sea in the reanalysis data (depicted in Fig. 9, left panel) across various summer monsoon months. While the same overarching pattern manifests in the model hindcasts, a slight deviation is observed wherein the moisture levels over the northern reaches of the Arabian Sea are not as significantly reduced as depicted in the reanalysis composite. An additional noteworthy characteristic apparent at the 850-hPa level in both reanalysis and model hindcast during subseasonal droughts (as shown in Fig. 9) is the overall weakening of the summer monsoon flow. The opposite phase of the anomalous wind direction, as depicted in Fig. 9, contrasts with the prevailing southwesterly flow observed during the summer monsoon over India and the Arabian Sea. Notably, the occurrence of anomalous easterly winds over the Arabian Sea implies a substantial reduction in the moisture transport from the Arabian Sea toward the Indian subcontinent (Krishnan et al. 2000). The intrusion of these dry extratropical winds into northwestern India and Pakistan can result in a reduction of convective instability, thereby weakening the monsoon flow (Krishnan et al. 2009). This amalgamation of factors exacerbates the drought scenario that is well depicted in the model hindcasts.
Composite anomalies of 850-hPa specific humidity and wind vectors during drought conditions over the CMZ for JJAS, spanning the years 1993–2015. (a)–(d) ERA5 reanalysis data. (e)–(h) NERP hindcast information. Regions of specific humidity with a significance exceeding the 95% confidence level are indicated by stippling.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
Another crucial aspect evident during subseasonal droughts (Fig. 9) in the Indian region is the association with anomalous cyclonic circulation in the northwest Pacific. This characteristic becomes particularly pronounced in June and July in both reanalysis and model drought composites. The presence of anomalous cyclonic anomalies stretching across the northwestern Pacific acts as a notable marker of increased typhoon activity in that specific geographic region. This activity, in turn, exerts downstream effects on the Indian region. Mujumdar et al. (2006) have provided insight into the dynamic relationship between intensified typhoon activity in the northwest Pacific and its influence on atmospheric conditions in India. This complex interplay triggers an exceptional atmospheric subsidence phenomenon, where air sinks within the atmosphere. The consequential subsidence can disturb the conventional progression of the Indian summer monsoon, often ending in drought conditions. In conjunction with the aforementioned features, an important one that warrants attention is the notable presence of strong low-level westerly wind anomalies prevailing in the western equatorial Pacific (as depicted in Fig. 9). These potent low-level winds in the western Pacific have the capacity to incite equatorial jets and set off Kelvin waves within the ocean, thereby potentially exerting influence on air–sea interactions within the region. It is worth noting that the inception of a drought is indicated by the persistence of westerly wind anomalies in the western equatorial Pacific. Referred to as westerly wind events (WWEs), these anomalies are considered surface manifestations of the MJO and play a pivotal role in extending warm SST anomalies toward the central Pacific, thereby influencing the evolution of El Niño conditions (Saith and Slingo 2006; Wang 2005). As a result, these characteristics assume significance as they contribute to the development of drought conditions through teleconnections. Moreover, it is noteworthy that the model hindcasts (Fig. 9, right panel) effectively capture these characteristics in comparison to the reanalysis data, thereby providing a robust validation of their representation.
b. Subseasonal droughts triggered by midlatitude teleconnections
Besides tropical drivers, various factors such as the midlatitude circulation, North Atlantic variability, and midlatitude wave trains have been proposed as contributors to subseasonal droughts in India (Ding and Wang 2005; Krishnan et al. 2009; Umakanth et al. 2019; Borah et al. 2020). On time scales ranging from intraseasonal to interannual, the abnormal midlatitude circulation over western and central Asia, as well as East Asia, has the potential to induce unusual cooling in the middle and upper troposphere due to the influx of cold air. This cooling effect, in turn, diminishes the meridional thermal contrast across the Indian subcontinent and weakens the monsoon flow, creating conditions favorable for drought. To assess how these midlatitude teleconnections associated with cold air advection are represented in model hindcasts, we presented composite temperature anomalies in the midtroposphere (at 500-hPa level) during prevalent drought conditions over the Indian subcontinent.
It is noteworthy that the ERA5 reanalysis (Fig. 10, left panel) highlights significant advection of cold anomalous temperatures over northwest India, particularly during June and July. Nevertheless, cold anomalies are consistently observed during the summer monsoon months when drought conditions prevail. The model hindcasts reasonably align with the observed temperature patterns in the reanalysis albeit with a slight underestimation of their magnitude. An intriguing observation pertains to the weakening of the Tibetan high, particularly notable in the month of July, coinciding with the presence of cold anomalies. This weakening phenomenon exerts a significant influence on the development of droughts in India. For example, Ramaswamy (1962) proposed a compelling hypothesis that the intrusion of troughs in the midlatitude westerlies could act as a causal mechanism, triggering breaks by weakening the Tibetan high pressure system. Furthermore, on subseasonal time scales, some distinctions emerge, particularly regarding the greater weakening of the Tibetan anticyclone in the model hindcast relative to the reanalysis during the months of August and September. However, notwithstanding these minor distinctions, the model effectively reproduces the large-scale patterns during drought conditions. The results illustrated in Fig. 10 provide additional support for our earlier discussion (and Fig. 9), underscoring the role of cold and dry anomaly advection into northwestern India and Pakistan. This advection process contributes to the reduction of convective instability, subsequently leading to the weakening of the monsoon flow.
Composite of 500-hPa temperature anomalies and winds: (a)–(d) data from ERA5 reanalysis; (e)–(h) information from NERP hindcast. Regions of temperature with a significance exceeding the 95% confidence level are indicated by stippling.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
We further examine the upper-level dynamic interactions between extratropical circulation and monsoon convection in model hindcasts compared to reanalysis during subseasonal droughts, as illustrated in Fig. 11. Figure 11a presents anomaly composites of 200-hPa winds and geopotential for the model, while Fig. 11b displays the same for the ERA5 reanalysis. Notably, these circulation anomalies, as seen in Figs. 11a and 11b, are predominantly characterized by unusual westerly winds prevailing over the subcontinent during the months of June and July. These westerlies signify a weakening of the upper-tropospheric easterly flow. A striking feature evident in both Figs. 11a and 11b is the prominent cyclonic circulation anomaly, particularly pronounced in the midlatitude region east of the Caspian Sea. This cyclonic anomaly is embedded within the westerly wind pattern and extends horizontally from the Mediterranean region to approximately 85°E in June and July, further eastward to around 130°E in August. It is crucial to note that the anomalous westerlies over west-central Asia extend significantly southward, intruding into the Indo-Pak region and north-central India. This cyclonic anomaly exhibits an equivalent barotropic structure and is observable throughout much of the tropospheric depth, as elucidated by Krishnan et al. (2009). In fact, the implications of this cyclonic anomaly are discernible at the 500-hPa level, as depicted in Fig. 10 revealing cold air advection represented by negative temperature anomalies extending into the Indian region.
Composite upper-level (200 hPa) wind anomalies (vectors) and geopotential (shading) from (a)–(d) ERA5 reanalysis data and (e)–(h) NERP hindcast information. Regions of geopotential with a significance exceeding the 95% confidence level are indicated by stippling midst.
Citation: Journal of Hydrometeorology 25, 8; 10.1175/JHM-D-23-0171.1
Furthermore, this anomalous cyclonic circulation, characterized by a negative center of geopotential height anomalies in the upper troposphere (as shown in Fig. 11) and a negative (cold) center of temperature anomalies in the middle troposphere (as displayed in Fig. 10), is particularly notable in northwest India. The strong negative height anomalies in the upper troposphere are associated with the presence of a strong cold air mass (negative temperature anomaly) beneath it, in accordance with the hydrostatic balance. In Fig. 11b, the model hindcast demonstrates a remarkable ability to faithfully replicate and showcase the same distinct features that are evident in the reanalysis data presented in Fig. 11a. This skillful representation underscores the model’s proficiency in capturing the complex patterns of upper-level dynamic interactions during subseasonal drought periods.
The upper-level circulation anomalies in Fig. 11 are crucial in understanding regional weather patterns. In Fig. 11a of the ERA5 reanalysis, a prominent ridge is evident over East Asia, specifically around the 100°E longitude. This ridge is a recurring feature, particularly notable during the months of June and July. Remarkably, this ridge’s significance has been previously discussed by Raman and Rao (1981). According to their findings, this stagnant East Asia ridge could play a pivotal role in anchoring cyclonic anomalies over west central Asia and the Indo-Pak region during periods of monsoon drought conditions. In essence, this ridge appears to have a substantial impact on the broader weather dynamics of the region. Furthermore, this ridge-like feature is associated with a distinct eastward shift of the upper-tropospheric anticyclone, which, in turn, leads to increased rainfall activity over the Indo-China region (Krishnamurti et al. 1989).
Hence, understanding the complex nature of monsoon dynamics, particularly in initiating subseasonal droughts, is of utmost importance. It is not just about tropical factors; extratropical forces are equally significant, particularly in the early monsoon months. These forces involve various meteorological phenomena like atmospheric circulation patterns, jet streams, and interactions with midlatitude weather systems. They lead to temperature, pressure, and wind anomalies that stretch beyond the typical tropical boundaries. This complicated interplay between tropical and extratropical factors creates a complex web of feedback mechanisms and teleconnections, where changes in one region can affect the monsoon elsewhere. The NERP model effectively captures these dynamics, enhancing our understanding of subseasonal droughts in the Indian subcontinent.
4. Summary and conclusions
Drought is a gradual but impactful natural phenomenon that significantly affects water resources and economies. Monitoring and predicting droughts, especially in countries like India where agriculture relies on seasonal rainfall, is vital for policy formulation. Subseasonal droughts, characterized by abrupt onset and sharp escalation, represent a distinct category between short-term meteorological droughts and longer-term agricultural droughts. These subseasonal droughts, mostly occurring during the summer monsoon in India, have emerged as a recent concern. We assessed the NERP modeling framework based on the unified global coupled modeling system, for characterizing subseasonal droughts. Using the standardized precipitation index (SPI), a well-known metric for drought assessment, we monitored drought conditions across the Indian subcontinent. SPI effectively quantifies precipitation deficits and deviations from normal conditions and provides insights into drought severity and duration, offering a robust framework for our analysis. The subsequent key points encapsulate noteworthy findings and insights collected from the research conducted in this study:
-
The NERP system effectively reproduces observed climatic patterns but tends to overestimate rainfall near the Himalayan foothills and the Indo-Gangetic Plain while underestimating it in the core monsoon zone (CMZ) and along the western coastline. In our model assessment, we calculated mean errors (MEs), revealing a significant dry bias along the Western Ghats and a smaller one in the CMZ. During July and August, the CMZ exhibits a wet bias exceeding 4 mm day−1, with a more pronounced wet bias in northeastern regions, including foothills and the Indo-Gangetic Plain (IGP).
-
It is noteworthy that disparities in drought patterns over the Indian subcontinent are observed in the NERP model when compared to IMD observations, especially evident in the instances of 2002 and 2009. Nevertheless, the NERP system effectively identifies all-India drought conditions at subseasonal time scales albeit with variations in location and intensity. Therefore, a more in-depth evaluation is conducted on the internal dynamics of summer monsoon droughts. This evaluation considers both tropical and extratropical influences within the model hindcasts.
-
Tropical teleconnections, such as the northward-moving MSIO and eastward-propagating MJO, play a crucial role in drought onset and persistence. While observations and the NERP model consistently identify MSIO locations, differences in phase propagation during intense 2002 and 2009 droughts contribute to spatial variations in drought intensity across the Indian subcontinent. The NERP model effectively predicts the MJO phase during drought conditions, with a stronger magnitude than reanalysis data. Additionally, the NERP model emphasizes the robustness of the coupled system response in the near-equatorial Indian Ocean, a significant factor in subseasonal drought development, as enhanced convective activity in the eastern equatorial Indian Ocean weakens monsoon circulation, leading to subsidence over the subcontinent.
-
We assessed midlatitude teleconnections related to cold air advection in NERP model hindcasts by examining composite winds, temperature, and anomalies at different pressure levels during prevalent drought conditions across various summer monsoon months. The NERP model effectively represents anomalous easterly winds in the lower troposphere at 850 hPa over the Arabian Sea, coupled with anticyclonic wind anomalies over India, reducing moisture transport from the Arabian Sea to the Indian subcontinent, leading to drought conditions. Additionally, the NERP model realistically represents anomalous cyclonic circulation in the northwestern Pacific, associated with increased typhoon activity that triggers downstream droughts in the Indian region. The model also skillfully captures upper-level dynamic interactions during subseasonal drought periods, characterized by unusual westerly winds weakening the tropical easterly jet and a cyclonic anomaly within the upper-level westerly wind pattern, reducing thermal contrast across the Indian subcontinent and leading to weakened monsoon flow and drought-favorable conditions.
Hence, within the NERP hindcast simulations, we effectively assess the prevailing drought conditions and the corresponding patterns of associated teleconnections. The results indicate that the model effectively represents the interactions of various weather factors during droughts. Our study highlights the NERP system’s strength in simulating drought dynamics, helping us better understand subseasonal droughts. Additionally, this study highlights that droughts can result from both tropical and extratropical factors though their relative importance remains to be conclusively determined. This aspect could be a subject for future research exploration.
Acknowledgments.
This work was carried out under the MoES-Met Office UM Partnership project and the Weather and Climate Science for Service Partnership (WCSSP) India, a collaborative initiative between the Met Office, supported by the U.K. Government’s Newton Fund and the Indian Ministry of Earth Sciences (MoES).
Data availability statement.
The gridded IMD rainfall data are freely accessible at www.imdpune.gov.in, and the reanalysis data (https://cds.climate.copernicus.eu/) as well as OLR data (https://psl.noaa.gov/data/gridded/data.olrcdr.interp.html) are also available without charge. The NERP hindcast simulation runs will be made available upon request, subject to compliance with the NCMRWF data-sharing policy.
REFERENCES
Ahn, M.-S., D. Kim, K. R. Sperber, I.-S. Kang, E. Maloney, D. Waliser, and H. Hendon, 2017: MJO simulation in CMIP5 climate models: MJO skill metrics and process‐oriented diagnosis. Climate Dyn., 49, 4023–4045, https://doi.org/10.1007/s00382-017-3558-4.
Annamalai, H., and J. M. Slingo, 2001: Active/break cycles: Diagnosis of the intraseasonal variability of the Asian summer monsoon. Climate Dyn., 18, 85–102, https://doi.org/10.1007/s003820100161.
Best, M. J., and Coauthors, 2011: The JOINT UK Land Environment Simulator (JULES), model description–Part 1: Energy and water fluxes. Geosci. Model Dev., 4, 677–699, https://doi.org/10.5194/gmd-4-677-2011.
Bhat, G. S., 2006: The Indian drought of 2002—A sub-seasonal phenomenon? Quart. J. Roy. Meteor. Soc., 132, 2583–2602, https://doi.org/10.1256/qj.05.13.
Blockley, E. W., and Coauthors, 2013: Recent development of the Met Office operational ocean forecasting system: An overview and assessment of the new global foam forecasts. Geosci. Model Dev. Discuss., 6, 6219–6278, https://doi.org/10.5194/gmdd-6-6219-2013.
Borah, P. J., V. Venugopal, J. Sukhatme, P. Muddebihal, and B. N. Goswami, 2020: Indian monsoon derailed by a North Atlantic wavetrain. Science, 370, 1335–1338, https://doi.org/10.1126/science.aay6043.
Bowler, N. E., A. Arribas, S. E. Beare, K. R. Mylne, and G. J. Shutts, 2009: The local ETKF and SKEB: Upgrades to the MOGREPS short-range ensemble prediction system. Quart. J. Roy. Meteor. Soc., 135, 767–776, https://doi.org/10.1002/qj.394.
Brown, A., S. Milton, M. Cullen, B. Golding, J. Mitchell, and A. Shelly, 2012: Unified modeling and prediction of weather and climate: A 25-year journey. Bull. Amer. Meteor. Soc., 93, 1865–1877, https://doi.org/10.1175/BAMS-D-12-00018.1.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Dhangar, N., S. Vyas, P. Guhathakurta, S. Mukim, N. Tidke, R. Balasubramanian, and N. Chattopadhyay, 2019: Drought monitoring over India using multi-scalar standardized precipitation evapotranspiration index. Mausam, 70, 833–840, https://doi.org/10.54302/mausam.v70i4.277.
Ding, Q., and B. Wang, 2005: Circumglobal teleconnection in the Northern Hemisphere summer. J. Climate, 18, 3483–3505, https://doi.org/10.1175/JCLI3473.1.
Edwards, D. C., and T. B. McKee, 1997: Characteristics of 20th century drought in the United States at multiple time scales. Atmospheric Science Paper 634, Colorado State University, 172 pp.
Francis, P. A., and S. Gadgil, 2010: Towards understanding the unusual Indian monsoon in 2009. J. Earth Syst. Sci., 119, 397–415, https://doi.org/10.1007/s12040-010-0033-6.
Gadgil, S., 2003: The Indian monsoon and its variability. Annu. Rev. Earth Planet. Sci., 31, 429–467, https://doi.org/10.1146/annurev.earth.31.100901.141251.
Gera, A., and Coauthors, 2021: Skill of the extended range prediction (NERP) for Indian summer monsoon rainfall with NCMRWF global coupled modelling system. Quart. J. Roy. Meteor. Soc., 148, 480–498, https://doi.org/10.1002/qj.4216.
Goswami, B. N., 2005: South Asian monsoon. Intraseasonal Variability of the Atmosphere–Ocean Climate System, W. K. M. Lau and D. E. Waliser, Eds., Springer, 19–61.
Guhathakurta, P., P. Menon, P. M. Inkane, K. Usha, and S. T. Sable, 2017: Trends and variability of meteorological drought over the districts of India using standardized precipitation index. J. Earth Syst. Sci., 126, 120, https://doi.org/10.1007/s12040-017-0896-x.
Gupta, A., A. K. Mitra, and E. N. Rajagopal, 2019a: Implementation of Unified Model based global Coupled Modelling System at NCMRWF. NCMRWF Tech. Rep. NMRF/TR/01/2019, 59 pp., https://www.ncmrwf.gov.in/reports.php.
Gupta, A., A. K. Mitra, and E. N. Rajagopal, 2019b: Implementation of sub-seasonal to seasonal forecast system with NCMRWF global coupled model. NCMRWF Tech. Rep. NMRF/TR/04/2019, 69 pp., https://www.ncmrwf.gov.in/reports.php.
Hersbach, H., and Coauthors, 2023: ERA5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS), accessed 2 September 2023, https://doi.org/10.24381/cds.adbb2d47.
Hunke, E. C., and W. H. Lipscomb, 2010: CICE: The Los Alamos Sea Ice Model documentation and software user’s manual, version 4.1. Doc. LA-CC-06-012, 76 pp., http://csdms.colorado.edu/w/images/CICE_documentation_and_software_user’s_manual.pdf.
Joseph, S., A. K. Sahai, and B. N. Goswami, 2009: Eastward propagating MJO during boreal summer and Indian monsoon droughts. Climate Dyn., 32, 1139–1153, https://doi.org/10.1007/s00382-008-0412-8.
Kemball-Cook, S. R., and B. C. Weare, 2001: The onset of convection in the Madden–Julian oscillation. J. Climate, 14, 780–793, https://doi.org/10.1175/1520-0442(2001)014<0780:TOOCIT>2.0.CO;2.
Konda, G., and N. K. Vissa, 2021: Assessment of ocean-atmosphere interactions for the boreal summer intraseasonal oscillations in CMIP5 models over the Indian monsoon region. Asia-Pac. J. Atmos. Sci., 57, 717–739, https://doi.org/10.1007/s13143-021-00228-3.
Krishnamurti, T. N., H. S. Bedi, and M. Subramaniam, 1989: The summer monsoon of 1987. J. Climate, 2, 321–340, https://doi.org/10.1175/1520-0442(1989)002<0321:TSMO>2.0.CO;2.
Krishnamurti, T. N., A. Thomas, A. Simon, and V. Kumar, 2010: Desert air incursions, an overlooked aspect, for the dry spells of the Indian summer monsoon. J. Atmos. Sci., 67, 3423–3441, https://doi.org/10.1175/2010JAS3440.1.
Krishnan, R., C. Zhang, and M. Sugi, 2000: Dynamics of breaks in the Indian summer monsoon. J. Atmos. Sci., 57, 1354–1372, https://doi.org/10.1175/1520-0469(2000)057<1354:DOBITI>2.0.CO;2.
Krishnan, R., K. V. Ramesh, B. K. Samala, G. Meyers, J. M. Slingo, and M. J. Fennessy, 2006: Indian Ocean-monsoon coupled interactions and impending monsoon droughts. Geophys. Res. Lett., 33, L08711, https://doi.org/10.1029/2006GL025811.
Krishnan, R., V. Kumar, M. Sugi, and J. Yoshimura, 2009: Internal feedbacks from monsoon–midlatitude interactions during droughts in the Indian summer monsoon. J. Atmos. Sci., 66, 553–578, https://doi.org/10.1175/2008JAS2723.1.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 1275–1277.
MacLachlan, C., and Coauthors, 2015: Global Seasonal forecast system version 5 (GloSea5): A high-resolution seasonal forecast system. Quart. J. Roy. Meteor. Soc., 141, 1072–1084, https://doi.org/10.1002/qj.2396.
Madec, G., 2008: NEMO Ocean Engine. IPSL Note du Pole de Modelisation, 300 pp.
Mahto, S. S., and V. Mishra, 2020: Dominance of summer monsoon flash droughts in India. Environ. Res. Lett., 15, 104061, https://doi.org/10.1088/1748-9326/abaf1d.
Markonis, Y., R. Kumar, M. Hanel, O. Rakovec, P. Máca, and A. AghaKouchak, 2021: The rise of compound warm-season droughts in Europe. Sci. Adv., 7, eabb9668, https://doi.org/10.1126/sciadv.abb9668.
McKee, T. B., N. J. Doesken, and J. Kleist, 1993: The relationship of drought frequency and duration of time scales. Eighth Conf. on Applied Climatology, Anaheim, CA, Amer. Meteor. Soc., 179–186, https://www.droughtmanagement.info/literature/AMS_Relationship_Drought_Frequency_Duration_Time_Scales_1993.pdf.
McKee, T. B., N. J. Doesken, and J. Kleist, 1995: Drought monitoring with multiple time scales. Ninth Conf. on Applied Climatology, Dallas, TX, Amer. Meteor. Soc., 233–236, https://www.tib.eu/en/search/id/BLCP:CN008169111/Drought-Monitoring-with-Multiple-Time-Scales?cHash=fde84712dd804e39b292837db919d94d.
Megann, A., and Coauthors, 2014: GO5.0: The Joint NERC–Met Office NEMO Global Ocean Model for use in coupled and forced applications. Geosci. Model Dev., 7, 1069–1092, https://doi.org/10.5194/gmd-7-1069-2014.
Mishra, A. K., and V. P. Singh, 2010: A review of drought concepts. J. Hydrol., 391, 202–216, https://doi.org/10.1016/j.jhydrol.2010.07.012.
Mishra, V., and K. A. Cherkauer, 2010: Retrospective droughts in the crop growing season: Implications to corn and soybean yield in the Midwestern United States. Agric. For. Meteor., 150, 1030–1045, https://doi.org/10.1016/j.agrformet.2010.04.002.
Mishra, V., K. Thirumalai, S. Jain, and S. Aadhar, 2021: Unprecedented drought in South India and recent water scarcity. Environ. Res. Lett., 16, 054007, https://doi.org/10.1088/1748-9326/abf289.
Mishra, V., M. Mujumdar, and S. S. Mahto, 2022: Benchmark worst droughts during the summer monsoon in India. Philos. Trans. Roy. Soc., A380, 20210291, https://doi.org/10.1098/rsta.2021.0291.
Mohan, T. S., H. Annamalai, L. Marx, B. Huang, and J. Kinter, 2018: Representation of ocean-atmosphere processes associated with extended monsoon episodes over South Asia in CFSv2. Front. Earth Sci., 6, 9, https://doi.org/10.3389/feart.2018.00009.
Mujumdar, M., V. Kumar, and R. Krishnan, 2006: The Indian summer monsoon drought of 2002 and its linkage with tropical convective activity over Northwest Pacific. Climate Dyn., 28, 743–758, https://doi.org/10.1007/s00382-006-0208-7.
Neena, J. M., E. Suhas, and B. N. Goswami, 2011: Leading role of internal dynamics in the 2009 Indian summer monsoon drought. J. Geophys. Res., 116, D13103, https://doi.org/10.1029/2010JD015328.
Niranjan Kumar, K., M. Rajeevan, D. S. Pai, A. K. Srivastava, and B. Preethi, 2013: On the observed variability of monsoon droughts over India. Wea. Climate Extremes, 1, 42–50, https://doi.org/10.1016/j.wace.2013.07.006.
Pai, D. S., P. L. Sridhar, P. Guhathakurta, and H. R. Hatwar, 2011: District-wide drought climatology of the southwest monsoon season over India based on Standardized Precipitation Index (SPI). Nat. Hazards, 59, 1797–1813, https://doi.org/10.1007/s11069-011-9867-8.
Pai, D. S., M. Rajeevan, O. P. Sreejith, B. Mukhopadhyay, and N. S. Satbhai 2014: Development of a new high spatial resolution (0.25° × 0.25°) long period (1901-2010) daily gridded rainfall data set over India and its comparison with existing data sets over the region. Mausam, 65, 1–18, https://doi.org/10.54302/mausam.v65i1.851.
Patel, N. R., P. Chopra, and V. K. Dadhwal, 2007: Analyzing spatial patterns of meteorological drought using standardized precipitation index. Meteor. Appl., 14, 329–336, https://doi.org/10.1002/met.33.
Pendergrass, A. G., and Coauthors, 2020: Flash droughts present a new challenge for subseasonal-to-seasonal prediction. Nat. Climate Change, 10, 191–199, https://doi.org/10.1038/s41558-020-0709-0.
Prasanna, V., and H. Annamalai, 2012: Moist dynamics of extended monsoon breaks over South Asia. J. Climate, 25, 3810–3831, https://doi.org/10.1175/JCLI-D-11-00459.1.
Preethi, B., J. V. Revadekar, and A. A. Munot, 2011: Extremes in summer monsoon precipitation over India during 2001–2009 using CPC high-resolution data. Int. J. Remote Sens., 32, 717–735, https://doi.org/10.1080/01431161.2010.517795.
Rae, J. G. L., H. T. Hewitt, A. B. Keen, J. K. Ridley, A. E. West, C. M. Harris, E. C. Hunke, and D. N. Walters, 2015: Development of the Global Sea Ice 6.0 CICE configuration for the Met Office Global Coupled model. Geosci. Model Dev., 8, 2221–2230, https://doi.org/10.5194/gmd-8-2221-2015.
Rajeevan, M., S. Gadgil, and J. Bhate, 2010: Active and break spells of the Indian summer monsoon. J. Earth Syst. Sci., 119, 229–247, https://doi.org/10.1007/s12040-010-0019-4.
Raman, C. R. V., and Y. P. Rao, 1981: Blocking highs over Asia and monsoon droughts over India. Nature, 289, 271–273, https://doi.org/10.1038/289271a0.
Ramaswamy, C., 1962: Breaks in the Indian summer monsoon as a phenomenon of interaction between the easterly and the sub-tropical westerly jet streams. Tellus, 14A, 337–349, https://doi.org/10.3402/tellusa.v14i3.9560.
Roxy, M., Y. Tanimoto, B. Preethi, P. Terray, and R. Krishnan, 2013: Intraseasonal SST-precipitation relationship and its spatial variability over the tropical summer monsoon region. Climate Dyn., 41, 45–61, https://doi.org/10.1007/s00382-012-1547-1.
Saith, N., and J. Slingo, 2006: The role of the Midden–Julian oscillation in the El Niño and Indian drought of 2002. Int. J. Climatol., 26, 1361–1378, https://doi.org/10.1002/joc.1317.
Sarkar, J., 2011: Drought, its impacts and management: Scenario in India. Droughts in Asian Monsoon Region, R. Shaw and H. Nguyen, Eds., Emerald Group Publishing Limited, 67–86.
Sarma, J. S., 2004: Review and analysis of drought monitoring, declaration and management in India. International Water Management Institute Working Paper 84, 40 pp., https://www.preventionweb.net/files/1868_VL102135.pdf.
Shah, R. D., and V. Mishra, 2015: Development of an experimental near-real-time drought monitor for India. J. Hydrometeor., 16, 327–345, https://doi.org/10.1175/JHM-D-14-0041.1.
Sheffield, J., and E. F. Wood, 2011: Drought: Past Problems and Future Scenarios. 1st ed. Routledge, 192 pp., https://doi.org/10.4324/9781849775250.
Thornthwaite, C. W., 1948: An approach toward a rational classification of climate. Geogr. Rev., 38, 55–94, https://doi.org/10.2307/210739.
Umakanth, U., and Coauthors, 2019: Meridionally extending anomalous wave train over Asia during breaks in the Indian summer monsoon. Earth Syst. Environ., 3, 353–366, https://doi.org/10.1007/s41748-019-00119-8.
Walters, D., and Coauthors, 2017: The Met Office Unified Model global atmosphere 6.0/6.1 and JULES global land 6.0/6.1 configurations. Geosci. Model Dev., 10, 1487–1520, https://doi.org/10.5194/gmd-10-1487-2017.
Wang, B., 2005: Theory. Intraseasonal Variability in the Atmosphere–Ocean Climate System, W. K. M. Lau and D. E. Waliser, Eds., Springer, 307–360.
Wang, B., and X. Xie, 1997: A model for the boreal summer intraseasonal oscillation. J. Atmos. Sci., 54, 72–86, https://doi.org/10.1175/1520-0469(1997)054<0072:AMFTBS>2.0.CO;2.
Wheeler, M. C., and H. H. Hendon, 2004: An all‐season real‐time multivariate MJO index: Development of an index for monitoring and prediction. Mon. Wea. Rev., 132, 1917–1932, https://doi.org/10.1175/1520-0493(2004)132<1917:AARMMI>2.0.CO;2.
Williams, K. D., and Coauthors, 2015: The Met Office Global Coupled model 2.0 (GC2) configuration. Geosci. Model Dev., 8, 1509–1524, https://doi.org/10.5194/gmd-8-1509-2015.
WMO, 2006: Drought Monitoring and Early Warning: Concepts, Progress, and Future Challenges. WMO, 24 pp.
Yuan, W.-P., and G.-S. Zhou, 2004: Comparison between standardized precipitation index and z-index in China. Chin. J. Plant Ecol., 28, 523–529, https://doi.org/10.17521/cjpe.2004.0071.
