ENSO play a considerable influence on global weather patterns, by exerting complex ocean-atmosphere dynamics (Wang et al. 2023). El Niñoand La Niña are natural cyclical patterns of warming and cooling and ENSO has been recognized as the most pronounced interannual oscillation signal of the climate system (Webster et al. 1998). ENSO is characterized by an increase in sea surface temperatures that trigger extreme variations in precipitation, violent storms, flooding and other notorious events (Adamson 2021). Traditional ENSO events occur on the coast of Peru along the eastern equatorial Pacific, and the SST anomaly gradually extends westward (Huang et al. 1989; Zhang et al. 1999, Feng et al. 2003). Recently, a new type of ENSO event has been observed, marked by the development of its warm/cold center in the central Pacific (Kao et al. 2009). Various studies have highlighted the significance of distinguishing different episodes of SST in the Pacific when discussing ENSO impacts on floods (Zhang et al. 2014, Zhang et al. 2016). Several indices are used to describe ENSO, including the Multivariate ENSO Index (MEI), Southern Oscillation Index (SOI), and Sea Surface Temperature rise (NOAA 2002; Shrestha & Kostaschuk 2005). Previous studies (Simpson et al. 1993; Kahya & Dracup 1994; Wijeratne et al. 2023; Kim & An, 2018, Sahu et al. 2012, Nguyen & Li 2014) have focused on leveraging ENSO impact assessments to predict stream flow, providing valuable insights into water flow for periods ranging from 6 to 12 months. The studies have been extended to the diverse effects of ENSO on stream flows in the Ganges River in India(Whitaker et al. 2001) and the Mahaweli catchment in Sri Lanka (Zaroug et al. 2013). While an ENSO impact on precipitation in Nepal has also been identified (Shreevastav 2019), runoff was underestimated if rainfall was uniformly placed over large grid cells. At present, the specific basin-level relationship between ENSO and streamflow dynamics in transboundary rivers remains elusive. This research seeks to fill this gap by examining the effects of high and low (El Niño and La Niña) ENSO events on transboundary river floods, considering both regional and local dynamics.

The Indo-Gangetic Plain (IGP) witnesses extensive inundation, primarily due to inefficient land-use planning, inadequate floodplain management, and insufficient pre-development research (Simpson et al. 1993). Downstream flooding near the Nepal-India border is exacerbated by unauthorized infrastructure construction without proper land assessment, compounded by the absence of cross-drainage passages and embankments on the IGP (Kahya & Dracup 1994). A floodplain assessment identified the downstream region of the Karnali River Basin (KRB) as particularly vulnerable (Dingle et al. 2020; Aryal et al. 2020). El Niño years bring about water scarcity for irrigation, while La Niña years pose heightened flood risks during the monsoon season. Understanding these dynamics and preparing for forecasted ENSO years could significantly benefit society, particularly in areas designated as high-flood-risk zones (Dingle et al. 2020). The influence of ENSO on flood discharge variation (in the downstream KRB, station no. 280) ranges from a minimum of 3354.0 m³/s to a maximum of 23295.3 m³/s, with gauge heights observed ranging from a minimum of 7.93 m to a maximum of 15.2 m from the observed Government of Nepal Department of Hydrology and Meteorology (DHM) data during 1964 to 2020 (Dingle et al. 2020; Aryal et al. 2020). The calculated 150-year return period is shown to be 16.2 year. These findings shed light on how ENSO can influence flood occurrences, offering valuable insights for enhancing flood response strategies. The DHM has established crucial threshold water level gauge heights for warning and danger levels during the monsoon season which is essential for accurate flood prediction and effective mitigation. The severity of downstream flooding, for example highlighted by the discharge recorded on August 15, 2014, at Chisapani gauging station, underscores the importance of precise flood prediction models (Aryal et al. 2020).

This study highlights the importance of combining hydrological and hydrodynamic models (like HEC-HMS and HEC-RAS) with recent field data for accurate flood prediction in response to ENSO events. This is crucial for protecting communities and livelihoods in flood-prone areas. The study emphasizes the vulnerability of mountainous regions to floods during ENSO events (Shrestha et al. 2002). However, studies using HEC-HMS at the basin scale demonstrate a high correlation between estimated and observed rainfall-runoff hydrographs (Hamdan et al., 2021). Similarly, good agreement between observed and satellite-derived precipitation in estimating catchment runoff was found (Belayneh et al. 2020). The KRB is highly prone to flood disasters (Aryal et al. 2020). The main objective of this study was to identify ENSO years and their impact on river flooding in the Karnali River Basin. Specifically, it determines monsoonal precipitation patterns, frequency of flood forecasts, and the relationship between rainfall and discharge influenced by ENSO events in the KRB. During ENSO events, the flood response and IDF curve were developed, emphasizing the impacts of these events on precipitation and flooding in KRB. By examining the interplay between ENSO events and non-ENSO precursors, the research provides a more comprehensive understanding of the ENSO influencing floods in the KRB. The findings underscore the importance of considering regional climate phenomena in developing robust flood response strategies, crucial for safeguarding communities and their livelihoods.

The study area is divided into three regions i.e. the China-exclusive zone covering 3060.1 km², the region from the China border to the Chisapani hydrological station spanning 42890 km², and the total basin study area extending to 46,177.3 km² up to the Indian border. The DHM has established 47 precipitation stations and 1 hydrological station in the KRB. These stations are distributed across elevations ranging from 114 to 7,746m asl in the KRB transboundary river system (Adhikari et al. 2024). Extreme floods and droughts reach their peak during the monsoon months (June to September) in the Karnali River basin of Nepal (DHM 1998). Nineteen hydrometric stations each with a record spanning more than 30 years were selected for the study (Fig. 1).

The study explores the impacts of the El Niño-Southern Oscillation (ENSO) on river flooding in the Karnali River Basin (KRB) by examining precipitation patterns, flood response, and associated hydrodynamic processes. The study makes significant strides in understanding these complex interactions which are crucial for effective water resource management and flood risk assessment in the region. The Karnali River catchment area spans 42890 km², and encompasses the West Seti and Bheri tributaries (Adhakari et al. 2013). In this study, the six main ENSO events from 1983–2020 and their relationship with precipitation and discharge was evaluated (Appendix III). The daily intensity of precipitation from 1964 to 2020 and hourly intensity of the precipitation from 1983 to 2020 with ENSO were also evaluated using IDF curves (Appendix I&II). IDF data and ENSO impacts on river systems can play an important role for people planting cash crops i.e., farming, downstream of Karnali hydrological station towards the Indian border (Masood & Takeuchi, 2012,Costabile et al. 2021,Pandit et al. 2023,Adhikari & Panthee 2020). During La Niña years, the basin's annual precipitation fluctuated, with 1413 mm in 1983, 1596 mm in 2000, and 1283 mm in 2014. In contrast, during the El Niño year of 2015, the precipitation was lower, at 1190 mm. The precipitation characteristics differed significantly across different elevations, ranging from 225 m asl to 1472 masl (Fig. 4). During the 2014 La Niña event, the maximum observed precipitation was 499.8 mm at an elevation of 225 meters. In the KRB, monsoon precipitation plays an important role in shaping precipitation patterns, especially during strong ENSO events. The total (monsoon season) precipitation in the basin during ENSO years was 960 mm in 1983, 1290 mm in 2000, 972 mm in 2014, and 747 mm in 2015.

The KRB exhibits significant hydrological variation due to its diverse elevation range, from 129 m at Naubasta to 3430 m at Khaptad. Flow values were higher in 1983 and 2000 compared to 2014 and 2015. The Gumbel frequency method was used to analyze precipitation and extreme flood data (Adhikari et al., 2021; Suhaiza Selaman et al., 2007; Pepin et al., 2015). Monthly hydrographs of precipitation and discharge during four ENSO events (1983, 2000, 2014, and 2015) reveal that ENSO has a significant impact on precipitation and discharge, especially during the monsoon season (Fig. 3), which is particularly sensitive to ENSO influences (Geng et al., 2023). During the events in 1983 and 2000, the instantaneous peak discharge reached 21,790 m³/s, while in the strong El Niño year of 2015, the observed peak discharge at Hydrological Station No. 280 in Chisapani was 2545 m³/s. The annual mean discharge in Chisapani from 1964 to 2020 was 1361 m³/s, with peak modeled discharge in the Karnali River estimated at 29,910 m³/s, and a 100-year flood depth of 23 m (Aryal et al., 2019). Data from 1964 to 2020 show an annual maximum observed discharge of 1796 m³/s, an average of 1361 m³/s, and an instantaneous maximum of 21,700 m³/s, with a minimum discharge of 2365 m³/s and a standard deviation of 269 m³/s.

The study focuses on the importance of the Soil Conservation Service (SCS) method in evaluating rainfall duration and flood response by considering terrain variations, soil types, and land use (Jothityangkoon et al., 2013; Nishio & Mori, 2015). This approach was successfully applied using the HEC-HMS and HEC-RAS models for rainfall-runoff estimation and flood simulation (Muhammad, 2016) and has been effectively used in previous studies (Hussein et al., 2022; Zema et al., 2017; Masood & Takeuchi, 2012). The SCS approach enhances the estimation of flood dynamics, offering in-depth insights into the connections between ENSO events and significant floods, including rare catastrophic flood events that occur every 600–800 years (Raj et al., 2020; Shrestha & Kostaschuk, 2005b; Abdolrahimi, 2016). The relationship between sea surface temperature (SST) and significant ENSO effects on the 1D and 2D HEC-RAS models revealed that river channel shifts occurred at 2,000-meter intervals along both branches, with a higher frequency on the right bank. Hence, the analysis initially focused on the right bank using the 1D model, while the HEC-RAS 2D model was applied to examine both sides of the river spatially. The 1D model effectively analyzed flooding scenarios across the right bank, center, and left bank, predicting flood velocity and channel shifts during the 2015 ENSO event. This approach, consistent with previous studies (Masood & Takeuchi, 2012), was crucial in forecasting flood discharge and assessing future flood risks. Additionally, the study found moderate to strong positive correlations between precipitation and discharge, with a moderate correlation observed in January and February, and a strong correlation in June and November, indicating a direct relationship between seasonal precipitation and discharge patterns. The regression analysis of ENSO events, focusing on basin hydro-meteorological parameters, reveals mean precipitation (X4) and SST as the most influential factors (Supplementary Appendix V). Mean precipitation (X4) shows a strong relationship with discharge, with a coefficient of determination (R²) of 0.44, explaining approximately 44.21% of discharge variability. While SST has a lower R² value of 0.13, it continues to play a significant role, especially in seasonal discharge variations. During the dry season (December-February), SST exhibit weak correlations with discharge (R² = 0.003) and mean precipitation (R² = 0.05). However, during the monsoon season (June-September), both mean precipitation and SST show strong correlations with discharge, with R² values of 0.44 and 0.67, respectively.

This study analyzes the anomalies of crucial parameters using a three-year running mean method, to assess ENSO impacts on discharge. These parameters include the SST index (Wijeratne et al. 2023), annual pressure (mb), annual rainfall (mm), annual temperature (°C), and annual discharge (mm) from 1964 to 2020. These variables exhibit fluctuations characteristic of El Niño (E), La Niña (L), and Neutral (N) ENSO phases (Kiem & Franks, 2001; Whitaker et al. 2001). Similar to "elevation-dependent warming” in mountain regions of the world(Pepin et al., 2015) and hydrological databases(Santos et al. 2017) HESSD offers further insights into elevation dependent phenomena. This methodology accurately predicted flood discharge, velocity, and the extent of inundation in settlements along the river, helping to assess and mitigate flood risks in the KRB during ENSO events which demonstrated increased flood discharge during high return periods (Table 7). The application of both 1D and 2D models reinforces the critical need for spatial and temporal flood forecasting, particularly in ENSO-affected years. The 1D model was utilized to analyze flooding scenarios on the right bank, center, and left bank (Masood & Takeuchi 2012). The 2D model was employed temporally and spatially on both sides of the river channel to forecast flood response. This methodology accurately predicted flood discharge, velocity, and inundation in the settlements (Maharjan et al .2023).

The study faces several constraints that limit the scope and accuracy of its findings. One significant limitation is the relatively small number of monitoring stations used in the analysis. The limited spatial distribution of these stations restricts the identification of core regions within the basin and impacts the accuracy of the data. Moreover, the short length of the data records, combined with the inter-annual variability of El Niño and La Niña events, makes it challenging to conclusively determine the influence of ENSO on hydrodynamic processes in the KRB. the, Moreover, we acknowledge the importance of exploring atmospheric influences beyond ENSO, as demonstrated recent studies which used non-ENSO precursors to significantly improve Asian summer monsoon predictability (Wang et al. 2023;Raj et al. 2020). The calibration and validation show that the model is highly satisfactory. During the ENSO year, 2015, discharge was also calibrated (Fig. 7). Our findings demonstrate significant ENSO effects on KRB discharge and suggest further investigation into future projections using climate prediction data. This study provides valuable insights into the complex interactions between ENSO events and river flooding in the KRB. Despite several limitations, the findings highlight the importance of continued research and the development of predictive models to enhance water resource management and mitigate flood risks.

This research established a notable positive correlation (0.59) between basin mean precipitation and discharge, indicating precipitation as a key driver of the basin's hydrology. Modeled discharge showed limited associations with sea surface temperature (SST) and pre-annual mean SST. A moderate correlation (0.43) was observed between modeled discharge and basin average temperature, while a stronger positive correlation (0.740) was evident with basin mean precipitation. During the monsoon season, a statistically significant relationship emerged among precipitation, SST, and discharge. Moderate correlations persisted in the pre-monsoon (MAM) and post-monsoon (ON) seasons. The study rigorously investigated the impact of the 2015 El Niño event on extreme flooding in the Karnali River Basin, contrasting it with the La Niña years of 1983 and 2014, which surprisingly recorded higher annual precipitation. Utilizing sophisticated 1D and 2D HEC-RAS models, the 2015 flood event was effectively simulated. The simulated peak transboundary discharges from China through Nepal to the Indian border closely aligned with observed data at Chisapani, and significant alterations in river channel morphology were identified. Statistical analysis indicated basin-wide precipitation as the primary factor influencing discharge, with modeled discharge demonstrating a moderate sensitivity to temperature variations. Seasonal assessments further revealed distinct precipitation-discharge relationships during the monsoon, with less intense but persistent correlations in other seasons. By integrating intensity-duration-frequency analysis and hydrological-hydrodynamic modeling, this study provided a comprehensive assessment of flood behavior during the El Niño year. Comparative analysis across ENSO phases illuminated the complex interactions between ENSO-induced precipitation variability and flood dynamics. These findings offer critical insights for future water resource planning and management, particularly considering the anticipated increase in hydroclimatic variability associated with future ENSO and El Niño events in the transboundary Karnali River Basin of Nepal.