Exploring the effects of extreme events on cereal cropping systems in the Ganga River basin, Haryana

Climatic conditions have both direct and indirect implications for a wide range of socioeconomic activities. The planet is currently experiencing a concerning array of phenomena, including rising surface temperatures (Thanh et al. 2023), heatwaves and wildfires (Deb et al. 2020; Engström et al. 2022), cold spells, intense rainfall, severe cyclones (Song et al. 2022), rising sea levels, and other natural hazards (Laino & Iglesias 2023; Deb et al. 2024). These global changes are resulting in substantial human, agricultural, and infrastructural losses. As emphasized in the IPCC SREX 2015 report, extreme events refer to weather phenomena in which a climate variable exceeds or falls below a predefined threshold value within the historical range of observed values (Nicholls & Seneviratne 2015; Tong et al. 2019). The frequency of extreme climate events has risen due to ongoing global climate change (Hao et al. 2013; Han et al. 2023). This heightened frequency has resulted in a clear upward trend in meteorological disasters, agricultural setbacks, and socioeconomic damages (Botzen & Van Den Bergh 2009). According to the IPCC (2021), climate change has altered the hydrological cycle, leading to increased frequencies of flooding, droughts, and other consequential extreme events. The effects of climate change are evident across all continents and oceans, impacting both natural ecosystems and human systems (Adisa et al. 2018; IPCC 2021). The phenomenon has evolved into a substantial challenge for the global population (Dai et al. 2016), with noteworthy implications at both local and global levels. These consequences extend to critical areas including human health, hydropower generation, tourism, the textile industry, food security, and water resources (Tian et al. 2017; IPCC 2021). In various regions, researchers have explored variations in air temperature and precipitation extremes across space and time (Zhang et al. 2016). These studies have placed significant emphasis on understanding the seasonal and annual changes in the spatial and temporal distributions of various weather parameters, including fluctuations in their frequency and intensity (Dos Santos et al. 2011; Bell et al. 2018; Arnell et al. 2019; Tabari 2021). Utilizing threshold values for temperature and precipitation, the Expert Team on Climate Change Detection and Indices (ETCCDI) has computed a range of indices in various parts of the world (Zhang et al. 2011). The indices have been widely adopted by researchers, encompassing diverse climatic conditions worldwide (Zhang et al. 2011; Chervenkov & Slavov 2019; Mistry 2019). Notably, the impact of climate extremes has been particularly pronounced in Asian countries, with influencing factors such as population growth, susceptibility to tropical cyclones, rainfall patterns, low-lying islands, and coral reefs (Collins et al. 2019). Furthermore, the vulnerability to tropical cyclones has been a significant concern.

Simultaneously, climate change significantly affects agricultural activities, driving shifts in food production and processes at both regional and local levels (Arora 2019). In the context of agriculture, climate is often perceived as a conditioning factor, with crop growth intricately tied to climatic variations in both temporal and spatial dimensions. Researchers have extensively investigated the extremes of temperature and precipitation, concentrating on their remarkable variability throughout the world's many locations. Noteworthy examples include studies in Thailand by Limsakul & Singhruck (2016); Pimonsree et al. (2022), India by Ghosh et al. (2016); Pradhan et al. (2019); Dash et al. (2022); Dubey et al. (2022) China by Qin et al. (2015); Zhou et al. (2016); Tong et al. (2019); Gan et al. (2023), and Nepal by Talchabhadel et al. (2018); Bohlinger et al. (2019); Poudel et al. (2020); Chapagain et al. (2021). On a global scale, precipitation patterns have undergone shifts, leading to an increase in variation. Wet regions have experienced heightened rainfall, while arid areas have faced exacerbated dryness. Consequently, the influence of climate change on agriculture is poised to differ significantly among regions due to the distinct nature and magnitude of regional climates and variations in farmers' resilience (Dai et al. 2016). This divergence emphasizes the pressing need to comprehend the impact of extreme meteorological events on crop production, thereby advocating for a more robust monitoring framework for assessing agricultural repercussions (Dai et al. 2016; Hatfield et al. 2020). However, there exist few studies that delve into the potential impact of climate change on crop yields (Challinor et al. 2014).

In the realm of drought assessment, the utilization of indices such as the standardized precipitation index (SPI) and the standardized precipitation evapotranspiration index (SPEI) holds global significance owing to their incorporation of both precipitation and temperature data (Tirivarombo et al. 2018; Pathak & Dodamani 2020; Pei et al. 2020). In 1993, McKee et al. (1993) initially replaced the Palmer drought severity index with the SPI, applied across multiple time scales of 3 6, 12, 24, and 48 months (Deb et al. 2022). This transition was, however, well-suited for encompassing temporal, spatial, and precipitation probability distributions. The SPI effectively quantifies the probability of drought occurrence across extensive regions, considering various time scales. Notably, it establishes precipitation as a pivotal factor influencing both the intensity and duration of droughts (Pramudya & Onishi 2018; Ali et al. 2019; Moghim & Takallou 2023). The SPEI stands similar to the SPI, yet it incorporates surface evapotranspiration through the integration of minimum and maximum temperature data (Gao et al. 2017). The documented impacts of different temporal scales on the effectiveness of drought indices are well-addressed within the literature (Yihdego et al. 2019; Vicente-Serrano et al. 2021). Understanding the spatial and temporal significances of climate variability on agricultural productivity is essential at both global and local levels, despite the abundance of research on the impacts of global warming on precipitation and temperature patterns.

In the year 2002, a single drought event resulted in a decrease of the net sown area by 12 million hectares. This reduction in cultivated land further translated into a decline of 38 million tons in food grain production, contributing to a 3.2% decrease in the agricultural gross domestic product of India (Rathore et al. 2014). The numerous studies provide analyses of various types of drought occurrences and their distribution in India's history. Notable examples include research conducted by Chowdhury et al. (1989); Guhathakurta & Rajeevan (2008); Jain et al. (2015); and Das et al. (2016). Droughts in India are increasingly localized, with a notable spatial shift toward coastal South India, central Maharashtra, and the Indo-Gangetic plains (Mallya et al. 2016). Meteorological droughts are characterized by insufficient precipitation, while agricultural droughts can result from elevated temperatures such as heatwaves, causing a reduction in soil moisture levels (Wilhite 2011). Droughts, while affecting various socioeconomic aspects, mainly impact agricultural production by disrupting water availability (Shah & Mishra 2014). For instance, recurrent and prolonged droughts often result in a general drop in crop yields, posing a threat to food security and triggering various environmental impacts. However, this preliminary study only focused on extreme events that impact the crop. With this given background and existing knowledge gaps, this study provides (a) an analysis of regional district trends by assessing extreme indices derived from long-term, multi-station daily temperature and precipitation data in four districts of Haryana, namely Sonipat, Panipat, Yamuna Nagar, and Karnal; (b) temporal trends in long-term precipitation and temperature to determine the significant trends in extreme conditions; and (c) an investigation of the correlation between extreme climate indices and the yield fluctuations of crops in the selected districts.

This analysis is crucial for Haryana, where agriculture is a vital sector. With climate change posing a substantial risk to agricultural productivity and management, its impacts could reverberate throughout both local and national economies. The findings of this study have important implications for future agricultural planning and climate adaptation strategies in Haryana. By understanding the trends and correlations between extreme climatic events and crop yields, policymakers and farmers can develop targeted interventions to mitigate the adverse effects of climate change on agriculture. Moreover, the study highlights the need for continued monitoring and research to adapt agricultural practices in response to the evolving climate. Addressing these challenges is essential not only for ensuring food security in Haryana but also for safeguarding the region's economic stability in the face of climate variability.

The Indian Meteorological Department (IMD) gridded daily precipitation dataset, with a spatial resolution of 0.25° × 0.25° and daily maximum temperature data sets at 1° × 1° grid resolutions (Srivastava et al. 2009), was used for the period 1980–2020. The study that computed the climate indices, as recommended by the World Meteorological Organisation (WMO) Expert Team on Sector-specific Climate Indices (ET-SCI), was undertaken utilizing the R-software package known as ‘CLIMPACT2’ (Alexander & Herold 2016). The CLIMPACT2 package was also employed to determine both the SPEI and SPI for the designated study area. During the analysis of climate extreme indices, CLIMPACT2 requires specific meteorological variables, namely daily minimum and maximum temperatures, along with precipitation data. Within this framework, CLIMPACT2 systematically derived both the core and non-core ET-SCI Indices, shedding light on historical temperature and precipitation variations. Table 1 provides an overview of all core and non-core climatic indices utilized in this context. The baseline period of 1990–2000 was employed for index computation within the broader timeframe of 1980–2020. To ensure data quality, prior to index computation, rigorous quality control procedures were administered through CLIMPACT2, including homogeneity tests.

The ETCCDI proposed 27 core indices with a key emphasis on extremes to be derived from station daily data. The WMO's working group on climate change detection further refined this index list (Folland et al. 1999; Jones et al. 1999; Peterson & Manton 2008). The core and non-core indices were chosen to reflect the relevance to sectors, such as health, agriculture, and food security, as well as water resources and hydrology. These extreme indices were well-suited for capturing temperature and precipitation variations across the study area. They were categorized based on percentiles, absolute values, duration, and thresholds. Comprehensive descriptions of all utilized indices are provided in Table 1.

In this study, the Mann–Kendall (MK) test was used (Guan et al. 2015). The MK test is a nonparametric statistical test used to detect trends in time-series data. It is particularly useful when the data violates assumptions of normality or when the data are ranked rather than measured on a continuous scale. Both the SPI and SPEI are valuable tools for assessing drought conditions, with the SPI focusing solely on precipitation anomalies and the SPEI incorporating both precipitation and potential evapotranspiration to provide a more comprehensive measure of moisture balance. A more detailed process has been given in the Supplementary File S1.

For this analysis, the precipitation changes, extreme indices, and SPI analysis of the Haryana state were conducted using the IMD high-resolution gridded (0.25° × 0.25°) daily precipitation and temperature dataset. The analysis covered 40 years of IMD data, which were divided into two time periods: 1981–2000 and 2001–2020.

The data provided represents monthly precipitation changes (in mm) for different cities in Haryana, India. The values indicate the difference between the actual precipitation received in a given month and the long-term average precipitation for that month. A positive or negative value indicates above- or below-average precipitation. A plethora of studies have focused on precipitation trends over Haryana. A 120-year IMD gridded rainfall dataset analysis revealed that eastern Haryana districts experienced more consistent rainfall with less variability compared to western Haryana across all seasons (Chauhan et al. 2022). Similarly, Guhathakurta et al. (2020) analyzed monthly, seasonal, and annual data for a 30-year period (1989–2018) across all districts of Haryana and reported a significant decreasing trend in Ambala, Bhiwani, Charkhi Dadri, Kaithal, Panchkula, and Panipat districts. Malik & Singh (2019) investigated rainfall patterns in Haryana during the period of 1997–2014 at both daily and seasonal scales, finding a positive trend in monthly maximum and total rainfall. Furthermore, Nain & Hooda (2019) conducted a trend analysis at the monthly scale for 27 rain gauge stations across all districts of Haryana using IMD datasets over 42 years (1970–2011) and reported mixed trends for the stations. Anurag et al. (2018) conducted a monthly rainfall analysis, revealing extreme results with an increasing trend in Hisar and a decreasing trend in Sirsa district, Haryana.

Figure 3 provides insights into how minimum temperatures have changed monthly for all the selected districts. This information is valuable for understanding temperature trends and fluctuations in these regions. A positive value indicates above-average minimum temperatures, while a negative value indicates below-average minimum temperatures. The data show that minimum temperatures in March, April, May, July, August, September, October, and December have all experienced positive changes, indicating an increase in minimum temperatures across all the districts. The data also indicate that January minimum temperatures, particularly in the Sonipat district, have shown the most significant declines among the winter months, with similar trends observed in other districts (Figure 3). In summary, the analysis suggests that minimum temperatures in these districts have undergone changes between the two periods (1981–2000 and 2001–2020). While there are variations in specific months, the overall trend points toward higher minimum temperatures in these districts over the past two decades. This trend could be indicative of a warming pattern in these regions. However, it is important to conduct more comprehensive and long-term analyses to confirm and understand the larger-scale trends of temperature changes and their implications for the climate and ecosystems.

The climatic indices provide valuable information about the climate in different districts. The reduced diurnal temperature range indicates that the temperature variability between day and night is relatively smaller. In the table, the diurnal temperature range in selected districts varies from −0.26 to −0.33 °C, indicating that the temperature gap between daytime and nighttime temperatures has been decreasing annually during the agricultural period. This reduction can pose challenges for agriculture, as rapid temperature fluctuations might impact crop growth and development. The PCPTOT shows a decreasing trend in all the districts, with the maximum changes observed in the Panipat district, i.e., −176.65 mm, during the selected time periods. These negative values indicate that, on average, the districts received less precipitation than the long-term average. Notably, the Panipat district shows the most pronounced negative trend, indicating a significant reduction in precipitation during the study period. In contrast, Yamuna Nagar shows the least variation in precipitation, while Panipat exhibits the most significant fluctuations in the indices. The negative values for r10 mm, r20 mm, and r25 mm indicate that, on average, the districts received less precipitation compared to the long-term average. Particularly, the Panipat district shows a substantial negative trend, implying a reduction in precipitation during the selected period. In contrast, Yamuna Nagar displays a more stable trend in precipitation, with fewer fluctuations in the indices.

The negative values for rx1day, rx3day, and rx5day indicate a potential shift toward less extreme and intense precipitation events in the given time frame compared to the long-term average. The Panipat district has the highest negative change in rx1day, i.e., −7.40, while Yamuna Nagar shows a relatively lower negative change, i.e., −1.67. These negative values indicate a decrease in the highest daily precipitation, suggesting fewer instances of intense single-day rainfall events. Regarding the rx3day index, the maximum negative changes were observed in Panipat, i.e., −13.32, while Yamuna Nagar showed a positive change, i.e., 0.24, during the period. This positive trend indicates an increase in the maximum accumulated precipitation over a 3-day period compared to the long-term average. In contrast, the negative values imply a reduction in the intensity of multi-day heavy precipitation events, indicating fewer instances of prolonged intense rainfall. A similar pattern is observed in the rx5day index, which shows a negative trend during the period. These negative values indicate that extended periods of heavy rainfall covering 5 days have become less frequent or less intense. Overall, these climatic indices help in understanding the climate variability and specific weather patterns in Haryana districts. They are essential for climate monitoring, planning, and adaptation strategies in the region.

In the temperature indices, the sunshine duration (SU) showed a rising trend in the districts, with the change ranging from 0.35 to 0.58 in the years 1981–2000 and 2001–2020. In the yearly changes, SU showed a slight increase in all the districts. The Tmm represents the average temperature value observed over a 24-h period. It is calculated by summing up all the individual temperature measurements taken throughout the day and dividing by the number of measurements. Tmm showed an increasing trend in all the districts, with variations ranging from 0.30 to 0.36. This index helps understand the general atmospheric conditions in a specific area. Both tn10p and tn90p are useful for understanding changes in temperature extremes and identifying shifts in climate patterns. They can provide insights into how the frequency of cool nights and warm nights is changing over time, which is valuable for assessing climate variability and potential climate change impacts on local and regional scales. The tn10p showed a negative trend in all the districts, with values ranging from −7.84 to −8.85, indicating a decrease in the frequency of cold nights (Table 2). Alternatively, tn90p showed a rising trend in the selected districts, with positive values ranging from 4.68 to 5.66, suggesting an increase in the frequency of warm nights in these districts. The tnm, tnn, and tnx (minimum temperatures) indices represent the monthly minimum temperature, the monthly minimum of daily minimum temperatures, and the monthly maximum of daily minimum temperatures, respectively.

All three indices (tnm, tnn, tnx) increased during the selected period, indicating an overall warming trend in minimum temperatures in Sonipat, Panipat, Yamuna Nagar, and Karnal. The indices tx10p, tx90p, and txgt50p (warm days) represent the percentage of days when TN < 10th percentile, the percentage of days when TN > 90th percentile, and the percentage of days where TX > 50th percentile. Tx10p showed a negative trend in the districts, indicating a decrease in the frequency of cool days during the selected time period. The study also observed an increase in tx90p (hot days) in Yamuna Nagar and Karnal and a slight decrease in the Sonipat and Panipat districts. However, txgt50p increased in all the districts, with variations ranging from 5.80 to 8.35.

The indices txm, txn, and txx (maximum temperatures) represent the monthly maximum temperature, the monthly minimum of daily maximum temperatures, and the monthly maximum of daily maximum temperatures, respectively. The txm and txx indicate a gradual warming trend, while the txn shows an increase in the coldest day in the districts. In the study, the occurrence of tropical nights increased in all the districts, with variations ranging from 0.58 to 0.68. This index was used to assess the prevalence of warm nighttime conditions in the districts (Table 2). It provides valuable information about the climate and its potential impacts on ecosystems, agriculture, and human comfort. In summary, the selected districts showed some consistent trends, such as a decrease in total precipitation, a reduction in the maximum amount of rain that falls in 1-, 3-, and 5-day events, and a slight cooling of the diurnal temperature range. Additionally, there were noticeable changes in the frequency of cool and warm nights, with a significant reduction in cool nights and an increase in warm nights.

Trend analyses were performed for the period 1981–2020 in the different grid points that covered the selected districts. All climate extreme indices trends are tested for statistical significance using the MK test in the MAKESENS Excel template (Mann 1945; Kendall 1962). All extreme indices showed positive and negative trends in all the districts (Table 3). The trends of indices were denoted as NS, negative significant, PS, positive significant, N, negative, and P, positive. The results showed that most of the precipitation-related indices had a significant negative trend in the Panipat district. The Sonipat district showed a negative trend, but it was not significant, except for rx1day, which showed a positive trend. In the case of Yamuna Nagar, the precipitation indices showed a positive trend, except for rx1day. In the Karnal district, the indices showed negative trends, but there was no significance in the trend. The SU showed both negative and positive trends but was not significant in the districts. The tmm showed a significant positive trend in all the districts. In the indices tn10p and tn90p, there was a significant trend for all the districts, with tn10p showing a negative trend and tn90p showing a positive trend. The tnm, tnn, and tnx showed significant positive trends in the districts, indicating that the temperature was increasing in these areas. The TR showed a positive trend and was significant in all districts except for Panipat. The tx10p and tx90p had both negative and positive trends, and the txgt50p showed a PS trend except in Panipat. In the txm, txn, and txx indices, there was a positive trend in the districts (Table 3), indicating that both the coldest and warmest days were increasing in these districts.

The drought occurrences are distinctly identified by both the SPI and SPEI. Minimal disparities were observed between the SPI and SPEI datasets, regardless of the analysis time scale. The assessment encompassed the 3-, 6-, 12-, and 24-monthly SPI and SPEI for the period 1980–2020. This outcome indicates that in climates characterized by low interannual temperature variability, both drought indices primarily respond to precipitation variability.

In the Panipat district, the 12-month and 24-month periods showed dry conditions in the area. The drought intensity of the district exhibited an increasing trend from 2000 to 2020 (Figure 4(b)). The maximum changes in the dry condition were observed in 2017, i.e., −2.21 in the 24 month and −2.22 in 2002 in the SPI 12 month. In SPI, 3-month and 6-month also showed the same dry trends in the selected period. In Figure 4(c), a 3-month SPI described the increase of the wet period in the long-term period analysis; a 6-month SPI in the study area showed the same pattern. Long-term trends in rainfall were revealed by the SPI during these periods. The 24-month SPI showed the increasing wet condition of the Yamuna Nagar district, with the maximum wet change observed in 2011–2015 and 2017–2020, while dry conditions were observed in 2002–2010 (Figure 4(c)). The same pattern followed the 12-month SPI in the Yamuna Nagar district. In the Karnal district, the overall dry condition increased in all the SPI months. However, in recent years 2018–2020, wet conditions were noted in the 24-month SPI, while other SPI months showed fluctuations (Figure 4(d)). The 12-month SPI showed a dry condition increase between 2000 and 2018. In SPI, 3-month and 6-month dry trends increased, with the highest values observed being −2.95 in 2016 and −3.12 in 2012. It helped governments, water resource managers, and agricultural planners make informed decisions to manage water resources, assess potential impacts on crops, and respond effectively to changing climate conditions.

Turning to the Yamuna Nagar district, the 24-month SPEI demonstrated an escalating trend toward wetness. The most substantial transition toward wetness occurred between 2017 and 2020, reflecting the SPI trend, while dry conditions were prevalent from 2000 to 2011. A parallel pattern was observed in the 12-month SPEI for Yamuna Nagar, with the most significant shift recorded in 2010 and 2009 (−2.21 in the 24-month and −1.81 in the 12-month series). The 3-month and 6-month SPEI followed a consistent pattern in the Yamuna Nagar district (Figure 5(c)).

In the Karnal district, overall dry conditions increased consistently with the SPI across all months. However, during the recent period of 2018–2020, the 24-month SPEI showed wet conditions, while other SPEI intervals fluctuated between wet and dry conditions. The 12-month SPEI showed a dry condition increase between 2000 and 2010, and changes fluctuated from 2011 to 2020 (Figure 5(d)). In SPEI, 3-month and 6-month dry trends increased, with the highest values observed being −2.52 and −2.25 in 1987. By examining the SPEI values over time, the study could identify drought or wet periods in the region under consideration, assess potential impacts on crops, and respond effectively to changing climate conditions.

The figures illustrate the correlation between precipitation and temperature extreme indices and crop yields (rice and wheat) for the Haryana districts. The analyses were divided into the rice and wheat crops in the districts. The indices with monthly output were obtained as an average of the period from June to October for the rice and from November to April for the wheat crop, which was the period that best represented the growing season for the studied crops in the study districts.

Panipat showed fluctuations in rice yield during the study period, with the yield rising from 2.45 to 3.1. During this period, precipitation indices, including PCPTOT, r10 mm, r20 mm, r25 mm, and rx1day, increased significantly. This suggested a strong correlation between increased rainfall and higher rice yield. However, it also indicated the influence of other factors on rice production. The most significant increase in rice yield in Sonipat occurred from 1998 to 2020, when it rose from 2.1 to 2.68. During this period, PCPTOT, r10 mm, r20 mm, r25 mm, rx1day, and rx3day increased significantly, indicating an effective correlation between increased rainfall and higher rice yield (Figure 6). The highest changes in rice yield in these districts appeared to be strongly correlated with increased rainfall, as indicated by various precipitation indices. However, there were instances of yield fluctuations that could not be explained solely by precipitation, suggesting the influence of other factors such as temperature, soil quality, agricultural practices, or pest infestations.

The district of Yamuna Nagar underwent several significant changes over the years. In terms of wheat production, there was a notable increase from 1998 (3.24) to 2020 (5.36) before stabilizing. The key climate-related parameters that contributed to these fluctuations were precipitation-related indices (Figure 8). In 2020, there was a high total precipitation (PCPTOT) of 277.24, while in 2013, there was an even higher PCPTOT of 307.66. These above-average precipitation levels likely boosted wheat production. Conversely, in 2012, wheat experienced a rise in production (5.38) despite a low PCPTOT (41.29), possibly indicating that other factors, such as extreme rainfall events (rx1day, rx3day, and rx5day), negatively affected crop yield. Karnal also experienced significant changes in wheat production and climate parameters. Wheat production notably increased from 1998 (3.54) to 2020 (4.85), alongside rising PCPTOT values, reaching 202.82 in 2020. This suggests a positive correlation between precipitation and wheat yield. However, in 2016, there was a dip in wheat production (3.44) despite moderate precipitation (PCPTOT 33.56). This might be ascribed to other factors like changes in daily temperature ranges (DTR) or extreme rainfall events. Similar to Yamuna Nagar and Karnal, Panipat observed fluctuations in wheat production and climate parameters. Wheat production increased significantly from 1998 (3.32) to 2020 (4.75), coinciding with increased PCPTOT. However, in 2015, there was a noticeable drop in wheat production (3.58) despite a relatively high PCPTOT (129.2). This could be due to factors like temperature changes or the distribution of rainfall (r10 mm, r20 mm, and r25 mm). Sonipat, like the other districts, witnessed fluctuations in wheat production and climate indices. Wheat production increased from 1998 (3.31) to 2020 (4.89), along with a rise in PCPTOT (159.16). However, in 2015, there was a drop in wheat production (3.73) despite reasonable PCPTOT (107.17). This suggests that factors like temperature (DTR) and the intensity of rainfall events (rx1day, rx3day, and rx5day) could have influenced crop yield (Figure 8). The overall changes in these districts primarily revolved around variations in wheat production, often correlated with precipitation-related parameters. While increased precipitation tended to boost wheat yield, other factors such as temperature, extreme rainfall events, and distribution of rainfall could also play crucial roles in determining agricultural outcomes.

The most significant increase in wheat production in Yamuna Nagar occurred between 1998 and 2020, when it rose from 3.23 to 4.89. This period witnessed substantial improvements in various climate indices, including solar radiation (SU) and temperature-related indices (TN10P and TN90P), which could have contributed to the higher yield (Figure 9(a)). However, the increase in temperature extremes during this period might have negatively impacted production. Karnal experienced a substantial increase in wheat production, rising from 3.54 to 4.85. This increase coincided with improvements in solar radiation (SU) and temperature-related indices (TN10P and TN90P). However, in 2016, there was a significant drop in wheat production despite relatively stable climate conditions. In Panipat, the most noteworthy increase in wheat production occurred from 1998 to 2020, with a jump from 3.32 to 4.75. This period witnessed improvements in solar radiation and temperature-related indices. However, in 2015, there was a sharp decrease in wheat production, even though climate conditions remained relatively stable (Figure 9(b)). Sonipat experienced a notable increase in wheat production, with improvements in solar radiation and temperature-related indices. Like Panipat, Sonipat also saw a significant decrease in wheat production in 2015, despite favorable climate conditions. Local factors, including soil health and agricultural practices, could have influenced this decline. Other factors, like soil quality or pest infestations, might have played a role in this decline.

The most significant increase in pearl millet production in Yamuna Nagar occurred between 1998 and 2020, when it rose from 1.44 to 2.57. This period witnessed substantial improvements in various climate indices, including solar radiation (SU) and temperature-related indices (TN10P and TN90P), which could have added to the higher yield (Figure 11(a)). The rise in temperature extremes during this period may have had a positive impact on production. In Karnal, pearl millet production saw significant fluctuations between 1998 and 2013, increasing from 1 to 3.28, but then decreasing from 2014 to 2020. In Panipat, the most noteworthy increase in pearl millet production occurred from 2003 to 2011, with a jump from 1.81 to 4.20. However, from 2012 to 2020, there was a sharp decrease in pearl millet production, even though climate conditions remained relatively stable (Figure 11(b)). Sonipat experienced a notable increase in pearl millet production from 1998 to 2020, climbing from 1 to 2. However, Sonipat's pearl millet yield showed variations from selected periods, despite favorable climate conditions. Only the Sonipat district maintained pearl millet yield consistently in all the years except 2011. The yield remained relatively constant from 2014 to 2020. Local factors, including soil health and agricultural practices, could have influenced this decline.

Exploring the effects of extreme events on cereal cropping systems in the districts adjacent to Yamuna River Basin in Haryana presents several significant challenges. One major hurdle is the availability and quality of data, as high-resolution and consistent climate and yield data are often limited, leading to potential inaccuracies in the analysis. Additionally, the spatial variability within the districts further complicates the assessment, as differences in climate conditions, soil types, and agricultural practices influence how extreme events affect cereal crops. Uncertainty in climate and crop models adds another layer of complexity, with potential discrepancies between model predictions affecting the reliability of results. Finally, socioeconomic factors such as market access and resource availability play a crucial role in determining crop resilience, making it challenging to isolate the effects of extreme events from other contributing influences. Addressing these challenges is crucial for developing effective strategies to mitigate the impacts of extreme weather on cereal cropping systems in the region.

This study examined how changes in climate over spatial and temporal scale affect agricultural production in the districts of Haryana. First, regional trends of extreme climate indices were obtained based on long-term observational daily climate data from Yamuna Nagar, Karnal, Sonipat, and Panipat. Based on the results obtained, a notable increase in average maximum and minimum temperatures is evident in the selected districts during the study period. These results also suggest that the SPI and SPEI for Haryana's districts, including Yamuna Nagar, Sonipat, Panipat, and Karnal, spanning from 1981 to 2020 at various monthly scales, underscore a growing concern about escalating drought conditions. The SPI trends indicate prolonged periods of dryness, particularly in Sonipat and Panipat, with a notable intensification of drought in recent years. Particularly noteworthy are the prolonged drought events observed in Sonipat and Panipat, persisting for extended periods from 2000 to 2020. This is underscored by a consistent upward trajectory in the SPEI intensity. In contrast, Yamuna Nagar experiences shifts between wet and dry conditions, with recent years showing an increasing trend toward wetness, while Karnal exhibits fluctuating conditions. These findings emphasize the multifaceted nature of drought dynamics and the significance of employing indices like the SPEI, which integrate both precipitation and temperature data, for effective drought monitoring and climate adaptation strategies in the region.

The analysis of the correlation between precipitation and temperature indices and rice yield in the districts of Yamuna Nagar, Karnal, Panipat, and Sonipat underscores the complex interplay of climatic factors on rice production. While there are notable correlations between certain precipitation indices, such as PCPTOT and heavy rainfall days, and rice yield, the relationships are not uniformly strong across all districts, suggesting the influence of other factors like irrigation practices. The highest increases in rice yield in these districts were strongly correlated with increased rainfall, as indicated by various precipitation indices. However, fluctuations in rice yield, particularly in Panipat and Karnal, could not be solely attributed to precipitation, implying the influence of other factors such as temperature, soil quality, agricultural practices, or pest infestations. Yamuna Nagar stands out as a district where most temperature indices exhibit an upward trend, aligning with increasing rice yields, implying a positive correlation between temperature and rice production. Conversely, Panipat and Sonipat districts experience relatively constant solar radiation (Su) indices, resulting in lower yields compared to other districts. However, even in these districts, rice production responds differently to various temperature indices, with some displaying decreasing trends in rice yields.

In this study, an in-depth examination of temperature indices and their impact on pearl millet yield production in the districts of Yamuna Nagar, Karnal, Sonipat, and Panipat from 1998 to 2020 revealed notable variations and correlations. Temperature-related indices, including solar radiation (SU), TMM, TN10P, TN90P, TNM, TNN, and TNX, provided insights into the complex relationship between temperature dynamics and pearl millet yields across these regions. The districts exhibited distinct patterns, with Yamuna Nagar consistently displaying the lowest pearl millet yields, while the other districts showcased higher yields, particularly in areas with greater solar radiation and temperature-related indices. The negative correlations observed between pearl millet production and PCPTOT, as well as rainfall-related variables, highlighted the preference for lower precipitation levels and reduced extreme rainfall events for increased pearl millet yields in these regions. Notably, Sonipat, Karnal, and Panipat exhibited consistent negative trends in correlations between pearl millet production and precipitation-related variables, emphasizing the impact of these climatic factors on crop outcomes. Conversely, the Yamuna Nagar, with its higher precipitation levels, displayed lower pearl millet yields. Moreover, the examination of extreme rainfall events suggested an upward trend in pearl millet yield in Panipat, while Yamuna Nagar exhibited negative trends, suggesting that heavy rainfall over consecutive days could lead to reduced crop production. These findings underscore the intricate relationship between climate indices and pearl millet yields, offering valuable insights into crop management and strategies for adaptation in agriculture.

The assessment of indices in relation to wheat yield in the districts of Yamuna Nagar, Karnal, Panipat, and Sonipat provides valuable insights into the complex association among climate variables and agricultural outcomes. Across these districts, wheat production generally exhibits a positive correlation with precipitation, particularly moderate and consistent rainfall. Higher total precipitation (PCPTOT), an increased number of moderate to heavy rainfall days (r10 mm), and even days with extremely heavy rainfall (r20 mm and r25 mm) tend to coincide with higher wheat yields. However, it is essential to maintain a balance, as excessive extreme precipitation can lead to flooding and crop damage. Moreover, fluctuations in temperature indices reveal nuanced effects on wheat production. While increased solar radiation (SU) is generally associated with higher yields, the impact of other temperature-related indices varies by district. These findings underscore the need for region-specific agricultural strategies that consider the interplay of both precipitation and temperature factors to optimize wheat production.

The variations in wheat, pearl millet, and rice production across these districts during the study period highlight the complexity of factors influencing crop yields. While certain years saw significant increases in production, often linked to favorable climate conditions, there were also instances of yield declines despite relatively stable climates. These findings suggest that environmental factors such as the frequency of days with maximum temperatures exceeding 25 °C, the highest recorded daily maximum temperatures, and the variation between daily maximum and minimum temperatures can negatively impact crop production in Haryana's districts. Moreover, temperature and precipitation extremes significantly influence crop yields. This underscores the importance of local factors, such as soil quality, pest infestations, and agricultural practices, in determining crop outcomes. To ensure food security and promote sustainable agricultural practices in these regions, policymakers and farmers must adopt a holistic approach that considers both climatic and agronomic factors in wheat and rice production. Additionally, ongoing monitoring and adaptation to changing climate conditions will be crucial in addressing the challenges posed by climate variability in agriculture.

The authors are thankful to IMD (India) and the Department of Economic and Statistical Analysis, Haryana for providing the data for the current analysis. This work also acknowledges the Nexus Gains initiative of CGIAR for funding.

S. K. D., V. K., and P. C. V. conceptualized the article and developed the methodology; S. K. D. rendered support in formal analysis; S. K. D. and P. D. figures; S. K. D. and V. P. wrote the manuscript and edited the article; S. K. D., P. C. V., V. K., and P. D. manuscript reviewed and edited the article.

This study was conducted as part of the Nexus Gains initiative of the CGIAR supported by the CGIAR Trust Fund contributed by various funders: https://www.cgiar.org/funders/.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.