Spatio-temporal trends and variability in extreme temperature and precipitation indices in the Kashmir Valley, North Western Himalayas

The average air temperature of the planet Earth has seen substantial warming and reports indicate that over the past century, surface temperature has risen by 0.7 °C with different areas warming to different degrees and land displaying greater warming than oceans (Trenberth 2011). The warming is unprecedented in the last three decades (Hartmann et al. 2013; IPCC 2013), affecting the whole Earth system (Lavinia 2011), and is largely attributed to the human lead-enhanced concentration of Greenhouse Gasses (GHGs) to levels never seen before in the recent past (Bindoff et al. 2013; Qin 2014; Wang et al. 2017). The future trends of these GHGs indicate further warming of the global surface air temperature (IPCC 2013, 2018; Incoom et al. 2023). The impacts of changes in global climate are observed worldwide in various forms and various extents. Climate change has contributed to apparent shifts in the frequency and severity of some kinds of extremes (IPCC 2014; Melillo et al. 2014; Donat et al. 2019; Sharma & Goyal 2020) with the warming of systems contributing to enormous shifts in the extent and severity of hot and cold events, hurricanes, droughts, thunderstorms, and rainstorms, with major environmental and socio-economic effects (Li et al. 2019; Pfleiderer et al. 2019; Zhang et al. 2019). These extreme events are altering the magnitude of the predicted climate; they, therefore, help us to perceive climate change. (Tour'e et al. 2017). Climate-related disasters now account for a major part of the increasing global disaster losses and thus, the understanding of climate extremes concerning global warming is of utmost importance for policymaking (Donat et al. 2013; Liu et al. 2019).

Mountain environments are characterized by their diversity and complexity and are very sensitive to climate change (Beniston 2003). Global climate change is affecting these topographically complicated regions like the Himalayas in wide-ranging ways favoring the triggering of extreme climatic events (Fort 2015). They come along the critically and speedily affected ecosystems and can be feigned by any alteration in temperatures and precipitation patterns (Zemp et al. 2009). Around 12% of the human population lives within or near the edges of the mountains and almost half of the world's human population directly or indirectly benefits from mountain resources (Fort 2015; Rasul & Molden 2019). Mountain ecology suffers more from climate change perils, such as extreme climate events, whose magnitude has increased by many folds during the last three decades. Risks from climate action failure and extreme weather are the most severe risks on a global scale over the next 10 years (GRR 2022). The study of climate extremes under the global warming scenario is thus of vital importance at a global scale (Alexander 2016; Zhou et al. 2016; Hadri et al. 2021) and on a regional scale (Franzke 2015; Sein et al. 2018; Naveendrakumar et al. 2019; Manzoor & Ahanger 2022; Pervin & Khan 2022). In regions like the Himalayas, the frequency of extreme climate events has increased over time the indicator-based approach can help assess the prevailing trends in climate and extreme events with greater precision.

Most of the studies on the Kashmir region in northwestern Himalaya (Islam et al. 2008; Shafiq et al. 2016, 2018a, 2019a, 2019b; Zaz et al. 2019) focus on the mean climate variability. Studies that have studied temperature change in the northwestern Himalayan region using long-term data are that of Bhutiyani et al. (2007) & Dash et al. (2007) who described a significant rise in surface air temperature in this region by about 1.6 and 0.9 °C, respectively, during the last century. Dimri & Dash (2012) examined the winter temperatures in the northwestern Himalayas and reported an increasing trend notably in mean maximum temperature (1.1–2.5 °C) during 1975–2006. Islam & Rao 2013 observed that the annual average temperature has shown an increasing trend significance test (Student's t-test at 95% confidence level) in Kashmir Valley at the rate of 1 °C during 1961–2005. The seasonal precipitation has shown variable trends in the western Himalayas. According to Li et al. (2018), precipitation has shown increasing trends in the summer and decreasing trends in winter. However, as shown by Shafiq et al. (2019a, 2019b) a consistent decreasing trend in annual precipitation at the rate of −5.1 mm/year (statistically significant at 90% confidence interval) is observed during 1980–2016.

The extreme climate over south and central Asia was studied by Klein Tank et al. (2009) who found that around 70% of the studied observational stations reflect increasing (decreasing) trends for warm days/nights (cold days/nights). According to Roy & Balling (2004), the increasing trend in precipitation events is visible from the northwestern Himalayas to the Deccan plateau during the last century. The percentage number of cold nights is decreasing while the percentage number of warm nights is increasing during the winter season in the western Himalayas (Dimri & Dash 2012) and the frequency of light and heavy precipitation events have increased significantly in the Hindu Kush Himalayan (HKH) region during 1961–2012 (Zhan et al. 2017).

Studying climate extremes is of very high importance particularly when the climate risks are increasing in the region that nestles a large human population of about 8 million people who depend on the delicate balance of the climate in the region. Besides the area nestles a large number of glaciers which are rapid response indicators of climate change (Dimri & Dash 2012; Shafiq et al. 2018b, 2020a, 2020b). The region has, over time, become more vulnerable to weather-induced disasters (Bhat et al. 2019; Shafiq et al. 2020a, 2020b; Ahsan et al. 2021). Moreover, climate change will likely increase exposure to natural or economic hazards, particularly because in many mountain regions, poverty is higher than and food insufficiency (Kohler et al. 2014).

The present study is carried out to assess and discuss the extreme temperature and precipitation study of Kashmir Valley. The study area, a small and narrow longitudinal valley with a higher altitudinal range, part of the Upper Indus Basin (UIB) in the Western Himalayas is home to around 8 million people. This study is based on the analysis of temperature and precipitation extremes in Kashmir Valley using observational data from six meteorological stations for 1980–2017. The study applies 25 indices to investigate the climate extremes in the Kashmir Valley. These indices are offered by the Expert Team on Climate Change Detection and Indices (ETCCDI, http://etccdi.pacificclimate.org) and have been extensively used by researchers to inquire into the spatio-temporal variability in climate extremes (Franzke 2015). This study is significant in terms of its novelty in using an indicator-based approach to study the climate extremes in this region of the Himalayas. The area like other parts of the Himalayas and elsewhere is witnessing climate change impacts such as the recession of glaciers (Romshoo et al. 2020) change in the form of precipitation (Romshoo et al. 2015) and depletion of glaciers and streamflow associated with changes in life and livelihood of mountainous communities (Bolch et al. 2012; Azam et al. 2018; Rasul & Molden 2019). Horticulture activities predominantly apple cultivation form the backbone of the economy of Kashmir Valley, which is heavily dependent on the existing climate conditions and the water resources in the region. The study is anticipated to form a basis for understanding and formulating climate change adaptation and water resource management strategies in the horticulture- cum agriculture-dominated region The findings of the study are expected to further increase the hazard assessment related to extreme weather perils that may prove vital in developing the necessary disaster risk reduction measures for climate-induced hazards in the region.

The observations for the study comprised daily maximum and minimum temperatures and daily precipitation of the Indian Meteorological Department (IMD), Pune, India for the six meteorological stations from 1980 to 2017 (Table 1). These stations are spread across the varied topographies of the valley. The quality of the data was assessed using the series method and autocorrelation and homogeneity tests among others. Furthermore, the data quality was also analyzed using quality control and homogeneity assessment modules of RClimDex and RHtestV3 programs, respectively (Zhang & Yang 2004; Zhang et al. 2005) which assess the daily maximum temperature less than daily minimum temperature, daily temperature values greater than 70 °C, leap days, all values corresponding to an impossible date and any non-numeric values among others. Moreover, for assessing the inter-relationship between the extreme precipitation events and various teleconnection patterns, the Indian Summer Monsoon Index (ISMI) and Western Disturbance Index (WDI) were used. Data for ISMI were downloaded from the Monsoon monitoring page (http://apdrc.soest.hawaii.edu/projects/monsoon/seasonal-monidx.html) and for WDI, which corresponds to the difference of geopotential height at 200 and 850 hPa levels, as suggested by Midhuna et al. (2020).

Expert Team on Climate Change Detection and Indices (ETCCDI) has come out with a core set of 27 extreme climate indices to enable a better understanding of observed climate change and weather extremes (Zhang et al. 2011). These indices describe the amplitude, frequency, and persistence of temperature and precipitation climate extremes. A user-friendly R-based program (RClimDex) was used for assessing extreme climate events http://etccdi.pacificclimate.org/ or https://www.wcrp-climate.org/etccdi (Zhang et al. 2011). In this study, out of 27 indices, we have used 25, including 15 temperature indices (Table 2) and 10 precipitation indices (Table 3). The extreme climate change analysis in this study was based on percentile thresholds for both precipitation and temperature calculated using the base period (1980–2010) analysis. The lower limit of the threshold was set for the 10th percentile of the time series and the upper limit was set for the 90th percentile of the time series for both temperature and precipitation. Instead of defined thresholds, the explanation for using mostly percentile thresholds is that the amount of days reaching percentile thresholds is more uniformly spread in space and is significant. These indices allow simple tracking of patterns in event severity or frequency.

As mentioned earlier, daily temperature and precipitation data for six meteorological stations were obtained from IMD, Pune from 1980 to 2017 to estimate the extreme precipitation and temperature changes in the study area. Besides, the manual data quality analysis, various statistical tests were also used like autocorrelation analysis and others. Data quality testing was also done using the inbuilt package of RClimdex 3.2.0 package (Zhang & Yang 2004). Time series with more than 20% of the missing values were omitted from the study as part of manual data calibration. To maintain the temporal consistency in data, the outliers in both temperature and precipitation time series data were identified. Any of the quality management steps or unreasonable data were found and corrected, such as days with a rainfall of negative or greater than 500 mm, days with a minimum temperature equal to or greater than the maximum temperature, and days with a minimum temperature greater than the maximum temperature. Furthermore, to avoid computational errors, the missing data in the time series were replaced with (−99.9) which is a recognized format of the RClimdex software package (Peterson et al. 2002; Aguilar et al. 2005).

The basic aim of the calculation of extreme climate is to monitor variability in moderate extremes and to get a major overview of the climate change and variability occurring across several spatial and temporal domains (Min et al. 2011). The study involves the investigation of 15 temperature indices (Table 2) and 10 precipitation indices (Table 3). The indices were classified into certain homogenous classes for both temperature and precipitation. The categorization was based on the nature and characteristics of particular indices under consideration. For temperature analysis, the indices were broadly categorized into Absolute indices, warming indices, and cooling indices. Similarly, for precipitation, the broad classes were Intensity indices, Frequency indices, and duration Indices. All the trends for the selected extreme climate indices were calculated annually. All the extremes were calculated using the percentile indices based on the reference period, 1980–2010, for the lower limit using the 10th percentile and the upper limit using the 90th percentile of the specific times series. RClimDex 3.2.0 employs the bootstrapping process to avoid potential biases within the reference duration associated with the present inhomogeneity during the trend calculation of percentile-based indices (Zhang & Yang 2004). Moreover, the recommended standard criteria were adopted after following the methods, as reported in the user manual of RClimdex at http://etccdi.pacificclimate.org/RClimDex/RClimDexUserManual.doc. All the correlations were calculated in R studio using the corrplot package.

For the present study, non-parametric statistical tests were used for analysis. The magnitude of the trend was estimated using Sen's slope estimator (Sen 1968) and the statistical significance of trends was computed using the Mann–Kendall test (Mann 1945; Kendall 1975) at different confidence levels. These trends were estimated for individual meteorological stations and the entire Kashmir region as well by averaging the values of individual stations.

The null hypothesis of the MK test states that there is no discernible trend in the time series of data. The existence of a significant trend, either positive or negative, in the data series conditional on the score, is tested when the null hypothesis is rejected.

If the sample size is greater than 10, the test statistic S is normally distributed.

Finally, Q_med is tested using a two-sided test at the 100 (1 – α) % confidence interval, and the nonparametric test can be used to determine the true slope.

The results of the extreme temperature analysis have been divided into several categories based on their character and nature which include, absolute indices, warming indices, and cooling indices. The trends for different extreme climate indices for the entire Kashmir and six meteorological stations for temperature have been given in Table 4, presenting the percentage of meteorological stations showing positive and negative trends along with their p-values.

Following the methods used in extreme temperature analysis, precipitation indices were also categorized as intensity indices, frequency indices, and duration indices. The trends for different extreme precipitation indices for the entire Kashmir and six meteorological stations have been presented in Table 5. Table 5 reveals the percentage amount of meteorological stations showing positive and negative trends along with their p-values.

In the case of extremely wet day precipitation (R99p), as shown in Figure 8(b), the average contribution was 7.29% to the total precipitation from 1980 to 2017 with an average contribution ranging from 0 to 26.74%. The analysis of R99p/PRCPTOT showed a highly fluctuating trend with the rate of increase. Overall, the contribution of very wet day precipitation (R95p) and extremely wet day precipitation (R99p) increased from 1970 to 2017. Furthermore, the contribution of extreme heavy precipitations increased from 2000 onwards.

In this study, Pettit's test was used to observe the abrupt changes (breakpoints) in the extreme climate time series of temperature and precipitation, as given in detail in Table 6. The analysis of the table reveals that most of the extreme indices showed abrupt changes from 1980 to 2017 with most of the variables showing abrupt changes during the 1990s. Except for TNn, other absolute temperature indices showed abrupt changes towards the end of the 1990s and early 2000s with TNn showing abrupt change during 1987 which indicates the transformation of climate from relatively cool to relatively warm periods. Quite early abrupt change in TNn shows that warming started quite early in extreme minimum temperature. The increase in DTR was witnessed from 1997. Furthermore, the majority of the cooling and warming indices showed abrupt changes during the late 1990s. ID showed an abrupt change in 2011 which indicates that Ice days showed a decrease towards the later part of the study. The analysis of Pettit's test on extreme precipitation revealed that most of the indices show abrupt changes during the 2000s with few exceptions. While RX1d showed abrupt changes towards the early part of the study in 1985, CDD showed it towards the later part of the study in 2014. RX1d and R99p showed abrupt changes in 2013.

The perusal of results indicates that most of the indices have a positive correlation with total precipitation except CDD. Barring a correlation between PRCPTOT and CWD, all other indices showed a correlation >0.6 (p = 0.05). SDII was also well correlated with all other extreme precipitation indices (except CWD and CDD). The correlation of CWD and CDD was mostly negative but all insignificant.

Many studies have been conducted that have signalled the impact of ISMI and WDI on precipitation in Kashmir. As shown in Table 7, the Western Disturbance Index (WDI) has a significant positive correlation with PRCPTOT and R10 mm, which indicates that WDI has a considerable influence on precipitation variations in Kashmir, while the Indian Summer Monsoon Index (ISMI) has shown a negative correlation with these two indices. Western disturbances or Mid-latitude westerlies are upper air circulations that are prevalent and most abundant in the winter season over the entire western Himalayas bringing most of the winter precipitation. Furthermore, most extreme precipitation indices have shown a negative correlation with the ISMI, thus signifying the lower effect of the Indian monsoon, which is not able to significantly cross the mighty middle Himalayas (Pir Panjal) to cause effective precipitation over the study area. WDI has shown a significant negative correlation with CDD, thus indicating that lower values of WDI enhance the persistence of precipitation deficit in the study area.

Significant spatio-temporal trends in extreme climate indices have been observed in the study area. The results showed a notable difference from one meteorological station to another. Studies conducted over the Himalayas have recognized and confirmed a dominating trend in extreme climate indices by linking the trend to the existence of lofty mountainous ranges, complex topography, and linkages to the changing patterns of saturated water vapour and temperature (Kazemzadeh et al. 2021). The present study establishes the imprints of global warming on climate extremes in the region. The increase in the frequency and occurrence of the warm temperature-based extreme indices is consistent with global warming but the magnitude of change varies spatially due to the large-scale atmospheric circulation patterns associated with it (Meehl & Tebaldi 2004). The Arctic Ocean's climatic circulation, with its alternating positive and negative phases, has the strongest influence on the extremes in the Northern Hemisphere (Thompson & Wallace 2000). Many studies have linked the substantially increasing monsoonal patterns of extreme events in the Himalayan Region and subsequent increase in the study area to teleconnections (Chevuturi et al. 2018; Ahsan et al. 2021; Bhardwaj et al. 2021). Our study shows a considerable decrease in the cold extremes and a rise in the hot extremes in the study area. The findings are consistent with similar studies conducted on extreme events locally and globally. Projections for the future show increasing trends in the warm temperature-based indices, while decreasing trends were found in the cold temperature-based indices for the 21st century (Ahsan et al. 2021). These indices could have a direct bearing on the socio-economic set-up of the study area owing to its fragile environment. Climate change can cause disasters in the region with floods and drought-like situations due to extreme wet and dry spells. Long dry spells are an emerging concern in the study area that lead to acute water scarcity and an increase in drought episodes (Himayoun & Roshni 2019).

The present work examines the extreme temperature and precipitation indices, as suggested by ETCCDI over Kashmir from 1980 to 2017. Widespread changes were seen in the extreme temperature indices mainly owing to the warming which has been recorded for individual meteorological stations and Kashmir as a whole during the last more than three decades. The absolute extreme temperature indices of TXx, TXn, TNx, and TNn have all seen increasing trends. The amount of changes witnessed in daily maximum temperature was greater than the daily minimum temperature which is further demonstrated by increasing DTR. The results for the warming indices of TX90p, TN90p, GSL, SU25, and WSDI have also shown an increasing trend from 1980 to 2017, whereas the cooling indices (TX10p, TN10p, FD, ID, and CSDI) have shown a negative trend. The results indicate a decrease in frost and ice days, while significant increases in hot days and nights which will have serious ramifications over the study area. The analysis of the step changes in the extreme temperature time series indicates that most of the changes have occurred or witnessed an upward trajectory during the 1990s. Furthermore, changes in the extreme precipitation indices have not shown a significant pattern compared to extreme temperature changes. The trend review reveals a heterogeneous pattern of positive and negative variations on the individual station scale in most of the extreme precipitation indices. Most of the extreme precipitation indicators suggest that extreme precipitation adds to overall precipitation indicating wetter conditions. CDD has shown a positive trend for Kashmir as a whole and five out of six stations which indicates that dry days have increased with most of the precipitation intensity, precipitation frequency, and precipitation duration indices registering a positive trend, thus leading us to the conclusion that increases in PRCPTOT and SDII are mostly due to extreme precipitation events. The analyses of the abrupt changes in extreme precipitation time series, however, indicate no coherent pattern and are distributed well along with the time series.

Recent decades have seen an increase in extreme weather events in the Himalayas like Leh cloudburst of 2010, the Kedarnath floods of 2013, the Kashmir floods of 2014, and the Chamoli glacier burst of 2021 (Roy & Balling 2004; Bhardwaj et al. 2021) among several others which are a direct consequence of climate change. Changes in extreme temperature and precipitation events will enhance in the future and multiply the negative impacts which occur in these areas. The study area has witnessed several flood-like situations and extreme flood events recently owing to the changes in extreme precipitation which have significant implications for water-related hazard risk management. The increase in extreme climate events poses a significant threat to agriculture, water resources, snow and glaciers, tourism, settlement, and ecosystem services of the study area. The study region is undergoing tremendous changes in urban areas, coupled with massive changes in the land system that increase its susceptibility to climate hazards, it is important to develop strategies for climate risk management in Kashmir. The current study focuses on establishing drifts in extreme climate indices, which is a less studied field in the study area and hence forms a base for future studies interested in estimating the sector-specific impacts of extreme climate change in the Kashmir Valley. The results of this study will help do away with the negative consequences of climate change and help policymakers in adapting and mitigating its impacts in the study area.

The authors would like to thank the Indian Meteorological Department (IMD) for providing the observational data for this research. The authors also thank the Editor Ahmed Kanawy, PhD and anonymous reviewers for their comments and suggestions on the earlier version of the manuscript that greatly improved the content and structure of this manuscript. The authors would also like to thank their respective organizations for constant support and infrastructure facilities.

M.U.S., Z.U.I., and P.A. conceptualized the study. M.U.S., Z.U.I., and R.M. did data curation. M.U.S., Z.U.I., and R.M., A.F. did formal analysis. M.U.S., Z.U.I., R.M., and A.P.D. investigated the study. M.U.S. and R.M. performed the methodology. R.M. and M.U.S. collected the resources; M.U.S. did software analysis; P.A. and A.P.D. supervised the study; R.M., M.U.S., and A.P.D. validated the study; M.U.S. and A.F. visualized the study; M.U.S., Z.U.I., and A.F. wrote the original draft.

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

The authors declare there is no conflict.