Examining trends in temperature and precipitation mean/extremes over Gandaki Province, Nepal

Climate change is a global issue that poses a significant risk to humanity. A long-term change in climate has been observed at a global scale caused by the increase in greenhouse gases (GHGs), particularly carbon dioxide (CO₂) (Loo et al. 2015). This change has led to variations in precipitation patterns, temperature, wind patterns, and an increase in the frequency of extreme events (Loo et al. 2015). As per the Intergovernmental Panel on Climate Change (IPCC) report of 2022, the global average temperature rose by 1.1 °C between 2011 and 2020 compared with the average temperature in the late 19th century (Pörtner et al. 2022). The report highlights that this temperature rise is caused by the increase in GHGs (Pörtner et al. 2022). To understand the influence of climate on the frequency and intensity of extreme weather events, historical and future climate data have been widely studied (Felix et al. 2021). Variations in temperature and precipitation trends can easily alter the hydrological cycle, so these two variables play a crucial role in understanding the climate (Felix et al. 2021). The IPCC suggested that studies on historical and future climate trends are essential for developing strategies to cope with the impacts of climate change (Felix et al. 2021).

The University of East Anglia's Climatic Research Unit has found that there has been a significant rise in the global temperature following a decline from 1945 to 1973 (Loo et al. 2015). This increase is believed to be linked to anthropogenic causes such as carbon emissions and urbanization (Loo et al. 2015). This temperature change has altered the precipitation pattern in terms of its type, amount, intensity, and frequency (Azadani 2012). Eastern North America, southern South America, and northern Europe have seen an increase in precipitation, while the Mediterranean, most of Africa, and southern Asia have seen a decrease in precipitation (Azadani 2012). Additionally, the global climate model predicts that extreme precipitation events will become more intense during the 21st century worldwide (Li et al. 2019). Due to the influence of atmospheric circulation changes, extreme precipitation events are likely to intensify in the tropics and subtropics (Li et al. 2019).

Climate change has intensified many extreme events in South Asia as several South Asian countries have reported an increase in temperature and precipitation in recent years (Naveendrakumar et al. 2019). Naveendrakumar et al. (2019) has mentioned that India has experienced higher-than-usual temperature, with an average warming of 0.51 °C. Another study by Sheikh et al. (2015) found that there was a higher precipitation trend over the Himalayan region of South Asia.

The altitude of Nepal varies from 68 m in the south to 8,848 m in the north within a range of 193 km (Upadhayaya & Baral 2020). This variation in altitude within a short elevation has made climate change a very sensitive issue in Nepal (Upadhayaya & Baral 2020). As a result, Nepal is ranked the fourth most-vulnerable country to climate change (Joshi et al. 2019). Studies have shown that the mean annual temperature of Nepal has been rising by 0.06 °C/year, and the climate in the Himalayan region of Nepal is changing faster than the global average (Upadhayaya & Baral 2020). A study by Shrestha et al. (1999) on 49 stations in Nepal concluded that there was consistent and continuous warming after the mid-1970s. The study found that the warming was most pronounced in the higher altitude areas of Nepal and less significant in the Terai and Siwalik regions (Shrestha et al. 1999). Additionally, the warming was more pronounced in the winter season than in other seasons (Shrestha et al. 1999). However, the study did not find any significant trend in precipitation (Shrestha et al. 1999). A study on extreme climate in Nepal by Awasthi & Owen (2020) showed a significant positive and negative trend in the hot and cold extreme indices. The study showed a higher increase in the extreme temperature in the Himalayan and mountain regions. However, the study showed an insignificant increasing trend in the precipitation indices (Awasthi & Owen 2020).

Most research on climate change impact in Nepal has focused on the national level and these studies have often only focused on a single climate variable, i.e., temperature, precipitation, or extremes. While Nepal has seen an increase in extreme events, there have been few studies on climate extremes in the country. Also, few studies have been conducted with a primary focus on all three climate variables in Gandaki Province. Hence, this study was carried out to assess the climate change scenario in Gandaki Province by analyzing:

• the trend in temperature, precipitation, and their extremes and

• the correlation between the extreme climate indices.

The study will provide a better understanding of the trend of climate variables and their extremes in Gandaki Province. The public and private sectors can utilize this information to effectively formulate policies to adapt and mitigate the impacts of a changing climate.

The study area for this research is Gandaki Province (Figures 1 and 2). Gandaki Province is one of the seven federal provinces, which was established on 20 September 2015. The headquarters of the province are Pokhara. Geographically, the region is located between 27°20′N to 29°20′N and 82°52′E to 85°12′E. The highest point in the province is the Dhaulagiri Himal, which stands at 8,167 m, and the lowest point is the Triveni Susta of Nawalpur.

The total area of the province is 21,504 km², which is approximately 14.57% of the entire country's area (Neupane et al. 2021). The Himalayan region accounts for 5,819 km² (26.8%) of the total area, the hilly terrain covers 14,604 km² (67.2%), and the Terai terrain makes up 1,310 km² (6%) (Neupane et al. 2021).

The province stretches from the plains of Terai in the south to the cold Himalayan peaks in the north, resulting in a wide range of climatic conditions. This variation in climate, coupled with different agroecological regions, leads to a diverse range of vegetation, from subtropical/tropical to temperate/tundra. The Kaski district of Gandaki Province has the highest rainfall (2,711 mm) in Nepal, whereas the lowest rainfall is recorded in the Mustang district (258 mm) (Neupane et al. 2021).

Gandaki Province has a total population of 2,403,016, which is 9.06% of Nepal's total population (Neupane et al. 2021). The male population is 948,029, whereas the female population is 1,144,124 (Neupane et al. 2021). The region's urban population is 1,452,186 (60.5%), whereas the rural population is 943,652 (39.5%) (Neupane et al. 2021). The population of the Himalayan region is 19,990 (0.8%), of the hilly region is 2,072,162 (86.5%), and of the Terai region is 310,864 (12.7%). Around 14.19% of the total population of Gandaki Province is living in poverty (Neupane et al. 2021).

Daily data for minimum temperature, maximum temperature, and precipitation for 30 years (1990–2020) were acquired from the Department of Hydrology and Meteorology (DHM), Government of Nepal. These data were collected for five stations within Gandaki Province, i.e., Chame, Jomsom, Lumle, Khudibazar, and Dumkauli of Manang, Mustang, Kaski, Lamjung, and Nawalparasi districts. The stations were selected in a way that represents all the climatic zones of the province as far as possible. Of the five stations, two stations are located in the Himalayas above 2,500 metres above sea level (masl), one station is located in the hilly region at an elevation of 1,738 masl, one station is located in the Siwalik region at an elevation of 838 masl, and one station is located in the Terai region at an elevation of 183 masl. The elevation of the station ranges from 183 masl (Dumkauli) to 2,741 masl (Jomsom). CMIP 6 data were acquired from the World Bank climate change knowledge portal, https://climateknowledgeportal.worldbank.org/.

All the stations have missing data (Table 1). On average, the stations have 5.4% missing data (0.3%–15.8%) for temperature and 1.9% (0.3%–4.5%) missing data for precipitation. The Chame station has the shortest temperature and precipitation record (1990–2012), whereas the other stations have data from 1990 to 2020.

Ensuring the quality and completeness of the data was an essential step for this study. The temperature and precipitation data were inspected using RClimDex for outliers. RClimDex identifies outliers and erroneous data such as negative precipitation value and maximum temperature less than minimum temperature. We chose four standard deviations as a threshold for identifying outliers in RClimDex (Shrestha et al. 2017). The identified outliers were replaced by the average values from the previous day and the next day (DHM 2017).

Missing data for precipitation were filled with APHRODITE V1901 data (0.25° × 0.25°), which were obtained from https://www.chikyu.ac.jp/precip/. APHRODITE data are constructed from observed precipitation data along with additional pre-existing datasets (Yatagai et al. 2012). APHRODITE data were first bias-corrected using the linear scaling bias correction method (Shrestha 2015). Since APHRODITE data were not available after 2015, missing data after 2015 were filled by taking the average monthly values from previous years. Missing data for temperature were filled with CHELSA V2.1 data, which were obtained from https://envicloud.wsl.ch/. CHELSA is a global fine-scale climate dataset (Karger et al. 2017). CHELSA data were also first bias-corrected using the linear scaling bias correction method (Shrestha 2015).

Annual data and seasonal data for all the stations were calculated using the daily data. The seasons considered for this study are winter, i.e., December, January, and February (DJF); pre-monsoon, i.e., March, April, and May (MAM); Monsoon, i.e., June, July, and August (JJA); and post-monsoon, i.e., September, October, and November (SON).

The 17 extreme climate indices used in this study are summarized in Table 2.

RClimDex was used for the calculation of indices. It is a freely available R package with a graphical user interface and has been widely used (Shrestha et al. 2017). First, the upper and lower threshold values for maximum and minimum temperatures were defined. The upper threshold value for maximum and minimum temperatures was taken as the average temperature of MAM and DJF, respectively (Shrestha et al. 2017).

Based on the result of the annual trend in temperature, we can see a significant trend in the maximum and minimum temperatures at Chame, Khudibazar, and Dumkauli (Table 3).

The maximum temperature at Chame showed a significant increasing trend at the rate of 0.14 °C/year (Table 3, Figure 3). This is the highest positive trend seen among the five stations and indicates a strong warming trend at high altitudes. Similar results were published by Khand et al. (2020), DHM (2017), Shrestha et al. (1999), and Bajracharya et al. (2011), where a higher warming trend was observed in the high-altitude region. A similar trend can also be observed in the whole Hindu Kush Himalayan region (Shrestha & Aryal 2011).

Snow albedo feedback influences the temperature at higher altitudes (Pepin et al. 2022). The reduction of snow cover will alter the albedo of the region, causing an increase in the temperature (Shrestha et al. 1999). A study by Gurung et al. (2017) showed a decreasing trend in the snow cover in the Gandaki basin. This reduction of snow cover may be contributing to the increasing temperature trend in the high-altitude region (Shrestha et al. 1999). Similarly, the maximum temperature at Khudibazar showed a significant increasing trend at the rate of 0.03 °C/year (Table 3, Figure 4).

The minimum temperature at Chame showed a significant decreasing trend at the rate of 0.12 °C/year (Table 3, Figure 3), which is the highest negative trend seen among the five stations selected in this study. Similar results were published by DHM (2017) and Bajracharya et al. (2011). However, the minimum temperature at Dumkauli showed a significant increasing trend at the rate of 0.02 °C/year (Table 3, Figure 5).

A decreasing trend in the minimum temperature can be observed in the Himalayan and Hill regions, whereas an increasing trend can be seen in the Siwalik and Terai regions (Table 3). Similar findings were published by DHM (2017).

Comparing our findings with the multi-model ensemble CMIP6 climate model projection for the year 2015–2100, the model result showed a significant increasing trend in the annual maximum and minimum temperatures (Table 3). Under the SSP2-4.5 scenario, the annual maximum and minimum temperatures over the selected stations are projected to increase at the rate of 0.03 °C/year. In the case of maximum temperature, both our findings and CMIP6 projection results showed a similar trend; however, in the case of minimum temperature, the CMIP6 projection showed a different trend in the Himalayan and Hill regions compared with our findings.

The precipitation at Jomsom showed a significant increasing trend at a rate of 2.42 mm/year (Table 4, Figure 6). This result is consistent with the findings published by Bhadra et al. (2021). A similar increasing trend in precipitation was also observed in the Tibetan Plateau (Zhan et al. 2017). A study by Sigdel et al. (2022) over the Gandaki basin also showed a significant increasing trend in precipitation over the Himalayan region.

Fort (2014) has mentioned increasing temperature trends at higher altitudes as a cause for increasing precipitation. The increasing temperature trend could influence permafrost distribution causing less snowfall and more rainfall.

The climate in the Trans Himalayan region is believed to be influenced by subtropical westerlies and the Indian summer monsoon (Zhu et al. 2015). So, the increase in precipitation could be because of the weather system developed by the collision of monsoons and westerlies supported by jet streams (The Kathmandu Post 2021).

Jomsom and Chame both lie in the Trans-Himalayan region, but the precipitation trend is different for these two stations (Table 4). In an area of complex topography, especially with drastic elevation changes, precipitation is locally determined (You et al. 2015). Factors such as mountain shape and relief and land cover affect precipitation in such regions (You et al. 2015). Also, these deep valleys establish their own wind system causing different precipitation patterns (Chen et al. 2018). Further research needs to be carried out to find out which factors play a key role in the distribution of precipitation in areas of complex topography.

The remaining stations showed a decreasing trend but were not statistically significant. The highest decreasing trend is seen in the Lumle station of Kaski district at the rate of 15.68 mm/year, which is similar to the findings published by DHM (2017).

Comparing our findings with the multi-model ensemble CMIP6 climate model projection for the years 2015–2100, the model result showed a significant increasing trend in the annual precipitation at Jomsom, Chame, Lumle, and Khudibazar and an insignificant increasing trend at Dumkauli (Table 4). Under the SSP2-4.5 scenario, the highest increasing trend is seen at the Khudibazar station. Our findings and the CMIP6 projection showed an increasing trend at the Jomsom station; however, the result was different for other stations.

The analysis of the number of rainy days, i.e., precipitation >2.5 mm in a day, showed a significant trend at Jomsom and Khudibazar (Table 5). The number of rainy days is significantly increasing and decreasing by 0.5 days/year at Jomsom and Khudibazar, respectively. All the remaining stations showed decreasing trends but these were not statistically significant.

The increase in the maximum and minimum temperature, decrease in precipitation, and the number of rainy days at Khudibazar and Dumkauli indicate that heatwaves and drought events could become more common in the Siwalik and Terai regions.

The result from the MWP test to detect an abrupt change in the temperature and precipitation is presented in Table 6. There was no consistent year detected at the stations; however, all five stations showed a significant change-point year for maximum temperature. Due to the lack of metadata, further research needs to be carried out to understand the significance of the detected change-point years.

The seasonal maximum and minimum temperatures showed a significant increasing trend at most of the stations (Table 7). The change in the maximum and minimum temperatures at Chame is generally higher compared with the other stations. During the pre-monsoon and post-monsoon seasons, Chame showed a significantly high increasing trend in the maximum temperature in comparison with the other stations. Similar results were published by Marahatta et al. (2009). The highest significant increase in the maximum temperature is seen during the winter season at Chame (Table 7, Figure 7), which is similar to the findings published by Konchar et al. (2015). Shrestha & Aryal (2011) also found that warming in the winter season in high-altitude regions is more than in the other seasons.

Chame showed a decreasing trend in the minimum temperature with a significant decreasing trend during the monsoon season (Figure 8). Marahatta et al. (2009) also found a high decreasing trend in the minimum temperature at Chame during the monsoon season.

The increase in the seasonal maximum temperature was generally higher than that of the seasonal minimum temperature, which is similar to the findings published by Khand et al. (2020) and Mishra et al. (2014).

The seasonal precipitation showed a significant increasing trend at Jomsom during the monsoon season and Dumkauli during the pre-monsoon season at the rate of 1.5 and 4.53 mm/year, respectively (Table 8).

Except for Chame, all the stations showed an increasing trend in the pre-monsoon season (Table 8). This trend is similar to the findings published by Karki et al. (2017) and Sigdel et al. (2022). This rise in precipitation in the pre-monsoon season indicates an increase in intense thunderstorms over the region (Karki et al. 2017).

There is an increasing trend in precipitation seen only in the Himalayan region during the monsoon season (Table 8). Similar findings were published by Karki et al. (2017) where a significant increasing trend was observed in the Himalayan region of central Nepal during the monsoon season.

The post-monsoon and winter seasons mostly showed a decreasing trend over the region (Table 8). Similar findings were published by Panthi et al. (2015) where a decreasing trend in post-monsoon and winter seasons was observed over the Gandaki River basin from 1981 to 2012. Winter crops may suffer due to a lack of soil moisture as a result of decreasing trends of precipitation in the post-monsoon and winter seasons (Panthi et al. 2015).

Out of nine temperature indices used, five indices showed a significant trend (Table 9).

Jomsom showed an increasing trend (statistically insignificant) in frost days (FDs) at the rate of 0.63 days/year (Table 9, Figure 9). However, a study by Manandhar et al. (2012) over the Kali Gandaki River basin for the period of 1978–2007 found a significantly decreasing trend in the FD at Jomsom. The difference in the trend could be due to the data taken from the different timeframes. Icing days (IDs) were not recorded at any station (Figure 10), which is similar to the findings published by Manandhar et al. (2012).

The number of summer days (SUs) showed a significant increasing trend at Jomsom, Lumle, and Khudibazar stations at the rate of 0.24, 0.66, and 1.02 days/year, respectively. The minimum value of daily minimum temperature (TNn) and the maximum value of daily minimum temperature (TNx) showed a significantly decreasing trend at Dumkauli and Chame. The TNn and TNx values decreased at the rate of 0.06 and 0.43 °C/year at Dumkauli and Chame, respectively.

The minimum value of daily maximum temperature (TXn) showed a significant increasing trend at Chame, Lumle, and Khudibazar at the rate of 0.57, 0.08, and 0.09 °C/year, respectively (Table 9, Figure 11). This trend is similar to the findings published by Shrestha et al. (2017). This warming trend in Himalayan, Hill, and Siwalik regions could be due to GHG emission and changes in cloud cover (Shrestha et al. 2017).

The mean difference between maximum and minimum temperatures (diurnal temperature range, DTR) showed a significant increasing trend at Chame, Jomsom, and Lumle at the rate of 0.22, 0.11, and 0.04 °C/year, respectively (Table 9, Figure 12). This trend is similar to the findings published by Shrestha et al. (2017), where ten of the 12 hill/mountain stations of the Koshi River basin showed a significant increasing trend in the DTR. The positive trend in the DTR indicates that the maximum temperature is increasing faster than the minimum temperature.

Out of eight precipitation indices used, three indices showed a significant trend (Table 10).

The maximum one-day precipitation (RX1 day) showed a decreasing trend at all the stations except Jomsom (Table 10, Figure 13). However, a significant trend was seen only at the Khudibazar station where the RX1 day value decreased at the rate of 1.05 mm/year. This trend is similar to the findings obtained by Karki et al. (2017) where a decreasing trend was observed in the Hill and Siwalik regions of central Nepal from 1970 to 2012.

The consecutive dry days showed an increasing trend (statistically insignificant) at all stations except Dumkauli (Table 10, Figure 14). Similar findings were obtained by Shrestha et al. (2017) where 47 of the 50 stations showed an increasing trend in the consecutive dry days (CDD). The increase in CDD could lead to an increase in the frequency of drought.

The total precipitation in wet days (PRCPTOT) showed a significant increasing trend at Jomsom at the rate of 3.77 mm/year, and the remaining stations showed a decreasing trend (statistically insignificant) (Table 10).

The consecutive wet days (CWD) showed an increasing trend at Jomsom, Chame, and Lumle, a decreasing trend at Khudibazar, and no trend at Dumkauli (Table 10, Figure 15). Jomsom showed a significant increasing trend at the rate of 0.08 days/year. Similar results were obtained by Shrestha et al. (2017), where 33 stations showed an increasing trend and 17 stations showed a decreasing trend.

The finding indicates that extreme climate conditions with heavier rainfall and increasing drought period could become more common at higher altitude regions. It also indicates that the trend in temperature extremes is more significant than precipitation extremes, which is similar to the findings obtained by Shrestha et al. (2017).

The correlation between the annual trend magnitude of extreme climate indices is shown in Figure 16.

Based on the result, the DTR showed a high positive correlation (statistically significant) with TNn. TNn showed a high positive correlation (statistically significant) with CDD. TNx showed a high negative and positive correlation (both statistically significant) with TXn and TXx, respectively. TXn showed a high negative correlation (statistically significant) with TXx. FD showed a high positive correlation (statistically significant) with PRCPTOT. SU showed a high negative correlation (statistically significant) with R10.

The correlation between the trend magnitude of extreme climate indices with altitude is shown in Table 11.

RX5 day and CDD showed a significant positive correlation with altitude. This finding provides further evidence for the frequent occurrence of extreme climate conditions at the higher altitude region.

The study focused on the assessment of climate change in Gandaki Province, Nepal, by analyzing the trend in the average and extreme values of maximum temperature, minimum temperature, and precipitation as well as analyzing the correlation between the extreme climate indices from 1990 to 2020.

The result showed an increasing trend in the maximum temperature at all stations, where Chame and Khudibazar showed a significant increasing trend. For minimum temperature, the Himalayan and Hill regions showed a decreasing trend with a significant decreasing trend observed at Chame, and Siwalik and Terai regions showed an increasing trend with a significant increasing trend observed at Dumkauli. The seasonal maximum and minimum temperatures showed a significant increasing trend in most of the stations except for Chame, which showed a significant decreasing trend in minimum temperature in the monsoon season. The increase in the seasonal maximum temperature was found to be higher than the seasonal minimum temperature. Also, a significant change-point year was detected for the maximum temperature at all stations.

The result showed a significant increasing trend in precipitation at Jomsom and a decreasing trend in the remaining stations. The seasonal precipitation showed an increasing trend in the Himalayan region during the monsoon season. The number of rainy days showed an increasing trend at Jomsom and a decreasing trend in the remaining stations.

The majority of the temperature indices showed an increasing trend. The number of summer days showed an increasing trend, where significant changes were found at Jomsom, Lumle, and Khudibazar. The minimum value of maximum temperature also showed an increasing trend, where significant changes were found at Chame, Lumle, and Khudibazar. Similarly, the diurnal temperature range also showed an increasing trend, where significant changes were found at Jomsom, Chame, and Lumle.

The precipitation indices showed mixed results, which suggest complex processes in precipitation extremes. The maximum one-day precipitation showed a decreasing trend at all stations except Jomsom. However, maximum five-day precipitation showed an increasing trend in the Himalayan and Hill regions and a decreasing trend in Siwalik and Terai regions. The number of very heavy rainfall days showed a decreasing trend in Himalayan, Hill, and Siwalik regions and an increasing trend in the Terai region. The CDD and CWD showed an increasing trend in the Himalayan and Hill regions. RX5 day and CDD showed a significant positive correlation with altitude

The analysis presented here indicates that frequent and intense extreme climate conditions will increase in the future, causing significant loss and damage to lives and properties in the province. Hence, further local-level studies should be carried out to better assess and understand the changes in climate variables. Mitigation plans and strategies should be developed and implemented to tackle the changing climate.

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

The authors declare there is no conflict.