Keywords

1 Introduction

The Brahmaputra is the fourth-largest river in the world in terms of annual discharge (Jungclaus et al., 2007). Being a sub-basin of the Ganges-Brahmaputra-Meghna basin, the Brahmaputra River Basin (BRB) plays a key role in the socio-economic development of over 131 million people living in the basin countries: China, Bhutan, India, and Bangladesh (Barua et al., 2019, Alam et al., 2021). The river frequently floods large areas in Bangladesh and northeast India during the annual monsoon rains with devastating consequences (Rao et al., 2020). Over the last few decades, the basin experienced major floods in 1954, 1974, 1987, 1988, 1993, 2004, 2007, 2008, 2010, 2012, 2016, 2017, and 2019 (Khalequzzaman, 1994, 2020; Hossain et al., 2019). Damage to property and loss of lives due to flooding have increased in recent years. For example, the flood in 1998, which displaced ~30 million people and caused ~1000 deaths (DMB, 1998), inundated 69% of Bangladesh (Mirza, 2003). The number of major floods that covered more than 35% of the land area in Bangladesh has increased from 1 in 32 years during the 1954–1986 period to 5 in the following 20 years during 1987–2007 (Dasgupta et al., 2011). In the Brahmaputra River, the number of years when the total annual flow exceeded 120 billion cubic meters has increased from 10 during 25 years between 1955 and 1980 to 18 during the following 25 years between 1981 and 2005 (Ray et al., 2015). In order to minimize the damage caused by such devastating floods, it is imperative to understand the changing patterns of floods in space and time for the BRB. This study uses the most up-to-date data available in the public domain on projected global warming and precipitation changes due to climate change that have the potential to impact future floods in the BRB basin.

The intensity and duration of flood events can potentially change as the climate warms. Climate change-induced temperature warming will likely increase precipitation in the basin and may shift the snow melting time in the headwaters. (Alam et al., 2021). Extreme precipitation is also expected to increase in the future (Immerzeel, 2008). Since the basin is susceptible to flooding, any substantial changes in precipitation and streamflow, especially in the wet season, can lead to increased flood risk. An increase in flooding is the most important change risk in northern and eastern India, adjoining Nepal, and Bangladesh (Sharma & Sharma, 2008).

Numerous studies reported that climate change-induced warming will influence the magnitude and intensity of precipitation, river discharge, and flood extent in the BRB (Ahmed et al., 2021; Dutta et al., 2020). Scientific evidence shows that most glaciers are retreating in South Asia’s Hindu Kush Himalayan region, including the ones that feed the Brahmaputra River discharge (National Research Council, 2012). Ding et al. (2006) found that the retreat of glaciers in this region is related to the increasing temperature. Yao et al. (2008) suggested that the glacial retreat increased more than 5.5% of river runoff from the plateau. Such observations align with broader scientific evidence of glacial retreat in the eastern and central Himalayas, including those feeding the Brahmaputra (National Research Council, 2012). On the other hand, Ge et al. (2004) observed that precipitation in the upper BRB shows an overall increasing trend, with the most remarkable increasing trend in spring, no significant change in fall and winter, and a decreasing trend in summer. Haque et al. (2021) concluded that compared to the historic baseline years, the mean annual temperature, amount of precipitation, discharge, and sediment yield in the BRB will likely increase in the 2030s (30 years during 2011–2040), 2050s (30 years during 2041–2070), and 2080s (30 years during 2071–2100). The implication of such changes has been quantified in several studies (e.g., Haque et al., 2021; Ghosh & Dutta, 2011, 2012; Alam et al., 2016, 2021). Climate change is likely to increase the peak flow of the Brahmaputra River by about 28% (Ghosh & Dutta, 2011). Change in monthly flow for August is likely to rise by 4, 5, and 8% for the years 2020s, 2050s, and 2080s, respectively (Paul, 2014). Alam et al. (2016) investigated the sensitivity of long-term discharge (mean and maximum of 30-year period) of the BRB to climate change and found that changes in discharge are correlated with changes in precipitation and temperature. Climate models predict an intensified monsoon and higher flood risk with global warming (Alam et al., 2021; Ray et al., 2015). In summary, earlier studies agree that climate change will change river discharge volume and seasonality in the BRB, with a higher impact during the latter part of the century (2080s).

Since global warming and potential changes in precipitation patterns will play a key role in the hydrologic responses of the BRB, it is crucial to analyze the extreme precipitation in the basin based on the most updated Global Circulation Model (GCM) projections. In particular, the new Intergovernmental Panel on Climate Change (IPCC) scenarios using the Coupled Model Intercomparison Project phase 6 (CMIP6) data incorporated the effect of changes in socio-economic projection and political environment in addition to global warming (Meinshausen et al., 2020). The probability estimation of extreme precipitation depends on the tail behavior of the distribution that models extreme events. The tail behavior is described by the shape parameter of the extreme value distribution (Coles, 2001). A heavy tail (the shape parameter is greater than zero) implies a higher probability of more frequent extreme events (Serinaldi & Kilsby, 2014).

The future peak streamflow based on projected precipitation and global warming can be useful to assess flood risk. These will help in developing an improved prediction of monsoon flooding which is an essential part of BRB management and could reduce loss of life and minimize the economic damage in the basin. With that in mind, there are three distinct objectives of this study: (a) analyze the basin-scale changes in mean annual precipitation for the future periods, (b) analyze the tail behavior of projected precipitation in the basin under various IPCC scenarios, and c) project the future peak streamflow of the basin based on CMIP6 precipitation and projected climate change that will result in global warming by 1, 1.5, 2, and 2.5 degrees.

We chose to calculate the discharges and corresponding river stages in Brahmaputra River at Bahadurabad, Bangladesh using 1, 1.5, 2, and 2.5 °C of global warming for the following reasons. First, as per the IPCC Sixth Assessment Report, the likely range of total human-caused global surface temperature increase from 1850–1900 to 2010–2019 is 0.8 °C to 1.3 °C, with the best estimate of 1.07 °C (IPCC, 2021). Second, the Paris Agreement, a legally binding international treaty on climate change, was adopted by 196 countries at COP 21 held in Paris on December 12, 2015, to limit global warming to well below 2 °C, preferably to 1.5 °C, compared to preindustrial levels (UNFCC, 2021). Third, the IPCC (2021) predicted that 1.5 °C global warming is likely to occur between 2030 and 2052. Fourth, although it is desirable to limit global warming well below 2 °C, the current actions and policies promised and practiced by most countries in the world through Nationally Determined Contributions (NDC) are not enough to reach the desired goal. The Climate Action Tracker (CAT), an independent scientific research group that analyzes the progress in tackling climate change, concluded that the average global temperature will likely increase by 2.7 °C by the end of the century (CAT, 2021). To be consistent with our calculation of discharges and river stages for each 0.5 °C increase, we have used 2.5 °C along with 1, 1.5, and 2 °C.

Several studies determined the future flow in the BRB at Bahadurabad using different GCMs and RCMs (Gain et al., 2011; Ghosh & Dutta, 2011; Pravettoni, 2015). Those studies used precipitation data for shorter periods. For example, Pravettoni (2015) used the precipitation data for 1983–2010, and Gain et al. (2011) used a discharge-weighted ensemble based on the inputs from the 12 General Circulation Models (GCMs) that are older than the CMIP6. The degree of knowledge about the impact of climate change on future precipitation events under various SSPs has improved since the last CMIP5 data (IPCC, 2021).

We calculated future extreme flows and corresponding river stages that will cause flooding in the BRB using the most up-to-date precipitation data generated by the CMIP6. Our study is the first of its kind to use the CMIP6 data to determine the changes in the future stream flows and river stages that are calculated using a field-based rating curve at Bahadurabad in the BRB.

2 Study Area

The BRB, the mainstream of which originates from the Tibetan Plateau, is the largest trans-Himalayan river basin, occupies 580,000 km2 area, and encompasses parts of the territory, ecosystems, people, economies, and politics of China (50.5% of the basin area), Bhutan (7.8%), India (33.6%), and Bangladesh (8.1%) (see Fig. 16.1). The BRB is characterized by three distinct physiographic zones, namely, the Tibetan Plateau (elevation is greater than 3500 m) in the western and northern parts of the basin, the Himalayan belt (elevation ranges between 100 m and 3500 m), and the piedmont-floodplain zone (elevation is less than 100 m). These three zones occupy about 44.4, 28.6, and 27% of the basin area, respectively (Immerzeel, 2008).

Fig. 16.1
A study area map depicts the location of Bahadurabad station and grid points indicated all over the region.

A map of the Brahmaputra River Basin. The star symbols show CMIP6 precipitation grid locations, and the red square indicates the Bahadurabad hydrometric station

Full size image

The basin’s climate is monsoon (May–September) driven, and 60–70% of the annual precipitation is received during the monsoon period. However, the upstream flow is generated by groundwater and glacial/snowmelt (Barua et al., 2019). The Brahmaputra basin receives an annual average 1,100 mm of precipitation (Immerzeel, 2008). The precipitation is high in the southern (downstream) part and is low in the northern part (upstream) of the basin. Duration and the total amount of precipitation have changed between 1951–1980 and 1981–2007 in the basin. The major changes occurred in the southern part of the basin where both duration and amount of precipitation events have increased during recent years (Pravettoni, 2015). The annual average streamflow of BRB is ~21,250 m3/s at the Bahadurabad station (located near the downstream end, see Fig. 16.1). The streamflow during the dry season (December–May) and wet season (June–November) indicated strong seasonality having the flow 7,565 m3/s for the earlier and 30,840 m3/s for the later season (Alam et al., 2021).

3 Data and Methods

We acquired daily precipitation outputs from nine GCMs used in CMIP6 historical and future experiments. These nine GCMs were chosen based on data availability for the study area. In addition, these nine GCMs reproduced the historic precipitation data for the BRB that closely matched the recorded historic precipitation values of 1,100 mm/year (Pravettoni, 2015). Table 16.1 shows a description of the GCMs selected for this study. The historical experiments include data from 1950 to 2014, and the future experiments include data from 2015 to 2100. We used three Shared Socio-economic Pathways (SSPs), namely, SSP126, SSP245, and SSP585 (Bottinger & Kasang, 2021; Riah et al., 2021). As included in the sixth IPCC report (IPCC AR6), compared to the previously used Representative Concentration Pathways (RCPs), the new SSPs have been improved to a great extent. The SSPs represent various socio-economic developments and atmospheric greenhouse gas concentration pathways (Meinshausen et al., 2020). These are (1) SSP126, a sustainable future pathway, low mitigation and adaptation challenge; (2) SSP245, an intermediate challenge for mitigation and adaptation, and (3) SSP585, fossil-fuel development, greenhouse gas emission through the twenty-first century, and mitigation challenges dominate. We have used these three scenarios to highlight the differences in three distinct pathways that range from the best to worst possible outcomes for the planet in the future.

Table 16.1 List of GCMs selected for the study
Full size table

We used a rating curve to estimate corresponding river stages for calculating the future streamflow of the Brahmaputra River at Bahadurabad station (see Fig. 16.2). The rating curve is developed by Bangladesh Water Development Board and is currently operational to forecast flood stages in Bangladesh (Sarder Udoy Raihan, Sub-divisional Engineer, Flood Forecasting and Warning Centre, Bangladesh Water Development Board; personal communication).

Fig. 16.2
A line graph plots water level versus discharge. The values are plotted for rated discharge and observed discharge. Both values depict an increasing trend.

Rating curve of the Brahmaputra River at Bahadurabad station (Sarder Udoy Raihan, Sub-divisional Engineer, Flood Forecasting and Warning Centre, Bangladesh Water Development Board; personal communication)

Full size image

The methodology used in this study is as follows:

  • Step 1: The daily precipitation data for the historical and climate change period of the 2080s from nine GCMs and three SSPs were downloaded from the CMIP6 database. The resolution of the GCMs ranges from 100 to 250 km. Therefore, GCMs were regridded at 100 km grid spacing using spline interpolation technique before the multi-model analysis was carried out on the 103 grid locations as shown in Fig. 16.1.

  • Step 2: The peak over threshold (POT) approach was used to sample the extreme precipitation events. The 90th percentile (P90) was considered to define the threshold value above which peaks are sampled. We employed the likelihood ratio test to choose the optimum threshold and used the extremal index to ensure the independence of exceedances (Masud et al., 2021). The generalized Pareto distribution (GPD) was fitted for these extreme events to analyze the tail behavior.

  • Step 3: The behavior of precipitation extremes is described by modeling the tails of the GPD distribution. The tail index of the fitted GPD distribution was calculated and compared between the climate normal period (30 years from 1981 to 2010) and the climate change period of 2080s. A heavy-tailed distribution generates more frequent and severe extremes compared to a light-tailed distribution.

  • Step 4: The basin-scale mean precipitation changes (ΔP%) for the 2080s were calculated with reference to the climate normal period. Then, four static temperature changes (ΔT) for the 2080s with respect to a climate normal period were set (1, 1.5, 2, and 2.5 °C) to consider the target global warming.

  • Step 5: The annual peak discharge (QMax (m3/s)) of the Brahmaputra River at Bahadurabad for the 2080s was estimated using the established multivariable regression equation developed by Alam et al. (2021):

    $$ {Q}_{\textrm{Max}}=\textrm{47,110}+2655\ast \Delta T+731\ast \Delta P $$

where ΔT is the changes in mean temperature (°C) and ΔP is the changes in mean precipitation (%) from the climate normal period at the basin scale. We have used the above equation to calculate the maximum flow, as opposed to the minimum and mean flow, because we are interested in extreme precipitation events that will result in flooding in the future.

  • Step 6: Finally, the corresponding river stages for calculated discharge values from Step 5 were determined using a rating curve developed by Bangladesh Water Development Board for the Brahmaputra River at Bahadurabad. The river stage values that exceed the flooding stage of the river were analyzed to assess the future flood risk for the Brahmaputra River at the Bahadurabad location in Bangladesh.

4 Results and Discussion

4.1 Changes in Basin-Scale Precipitation

Figure 16.3 shows the simulated basin-scale mean annual precipitation for the historical (1950–2014) and future (2015–2100) periods based on nine GCMs and three SSPs. The blue-shaded band zone shows the range of historical precipitation, and the orange-shaded zone shows the range of future precipitation based on these scenarios. The thin lines indicate simulated precipitation for various GCMs and SSPs, whereas the bold lines indicate the ensemble mean. A careful observation of the plots reveals that there exists a great number of uncertainties in simulated precipitation in historical and future scenarios. The SSP126 and SSP245 demonstrate a mild increase in basin-scale mean annual precipitation, whereas the SSP585 results in a higher increase in mean annual precipitation for future periods.

Fig. 16.3
Four graphs a to d plot precipitation versus years for various models. The values are plotted for historical and future precipitation. All the graphs depict a fluctuating pattern.

Historical (1950–2014) and projected (2015–2100) basin-scale mean annual precipitation in the Brahmaputra River Basin based on nine GCMs and three SSPs

Full size image

Figure 16.4 shows boxplots of basin-scale precipitation changes for the 2080s compared to the climate normal period based on the nine GCMs and the three SSPs. As expected, the interquartile ranges of projected precipitation, which is a measure of uncertainties associated with the projections, grow as we move to the distant future. The cross marks demonstrate the ensemble mean of the projected changes in precipitation. The percentage changes in basin-scale mean annual precipitation for the 2080s compared to the climate normal period for the SSP126 range from 0.85% to 14.84% with an ensemble mean of 7.48% (Fig. 16.4). The GCM ACCESS-CM2 projected the highest increase (14.84%), and the CAMS-CSM1-0 projected a slight decrease (−0.85%) in mean annual precipitation (Fig. 16.3a). The percentage changes in basin-scale mean annual precipitation for the 2080s compared to the climate normal period for the SSP245 range from −3.31% to 18.01% with an ensemble mean of 8.01% (Fig. 16.4). The EC-Earth3-Veg-LR projected the highest increase (18.01%), and the CAMS-CSM1-0 projected a significant decrease (−3.31%) in mean annual precipitation (Fig. 16.3b). The percentage changes in basin-scale mean annual precipitation for the 2080s compared to the climate normal period for the SSP 585 range from −2.13% to 31.12% with an ensemble mean of 19.37% (Fig. 16.4). The MRI-ESM2-0 projected the highest increase (31.12%), and the CAMS-CSM1-0 projected a considerable decrease (−2.13%) in mean annual precipitation (Fig. 16.3c). These results are generally in agreement with previous studies (Almazroui et al., 2020; Caesar et al., 2015). As shown in Fig. 16.3d, all SSPs (SSP126, SSP245, SSP585) project a moderate increase in basin-scale precipitation in the early to mid-twenty-first century (2015–2050). The projected changes in precipitation in the near decades are not very sensitive to the emission pathways as the radiative forcing is very similar during this time. From 2050 onwards, SSP585 projects a significant increase in mean annual precipitation, and SSP126 and SSP245 show a moderate increase. Ensemble mean of the mean annual precipitation in 2080s is almost similar for SSP126 and SSP245, indicating that higher precipitation change is only expected in the worst scenario (SSP585) that will result from high warming and low mitigation measures in the future.

Fig. 16.4
A box plot depicts changes in precipitation in percentage. The values are plotted for S S P 126, S S P 245, and S S P 585. The S S P 585 has the highest value.

Boxplots of basin-scale mean annual precipitation changes for the 2080s compared to the climate normal period of 1981–2010 based on nine GCMs and three SSPs. The cross marks demonstrate the ensemble mean of the projected changes in precipitation

Full size image

4.2 Tail Behavior of Extreme Precipitation Events

Figure 16.5 shows the tail index (shape parameter of GPD distribution) for historical and three future climate change scenarios (SSPs) based on the ensemble mean of nine GCMs. The tail indices determined for precipitation extremes show that by the 2080s, the BRB will experience more intense precipitation events under various SSPs as compared to the historic period. The extreme precipitation will increase under all SSPs. The reason behind the increase in extreme precipitation is attributed to enhance tropical convective precipitation in this region (Ge et al., 2021). Such an increase will likely be observed in the upstream part of the basin in the Tibetan Plateau and the downstream part of the basin in Bangladesh.

Fig. 16.5
Four spectral maps of the study area plot latitude versus longitude, depicting tail index values from negative 0.25 to 0.25 for historical, S S P126, S S P 245, and S S P 585.

Tail index (shape parameter of GPD distribution) for historical and three future climate change scenarios (SSPs)

Full size image

Figure 16.6 shows the kernel density plot (a nonparametric probability density function plot) of the tail index. A density plot is a representation of the unknown distribution of a numeric variable and a smoothed version of the commonly used histogram. The tail index clearly shows that extreme precipitation will increase in the future (positive values of tail index and right side of the density curves), and floods are likely to occur due to such extreme heavy precipitation (Chen & Zhai, 2013) although changes in antecedent soil moisture, storm extent, and snowmelt will pose a grand challenge for such a cause-effect relationship (Sharma et al., 2018).

Fig. 16.6
A line graph plots density versus the tail index. The lines are plotted for historical, S S P 126, S S P 245, and S S P 585. The lines depict a bell curve pattern.

Kernel density plot (a non-parametric probability density function plot) of the tail index

Full size image

4.3 Future Peak Streamflow and River Stage

Table 16.2 shows the estimated peak discharge of the Brahmaputra River at Bahadurabad based on the projected changes in precipitation from nine GCMs and three SSPs and four target temperature changes (1, 1.5, 2, and 2.5 °C) in the 2080s. Figure 16.7 shows a Boxplot representation of these projected streamflows. The corresponding river stages are shown in Fig. 16.8. The cross marks in Figs. 16.7 and 16.8 demonstrate the ensemble mean of the projected streamflow and stages, respectively.

Table 16.2 Projected future (2080s) peak discharges of the Brahmaputra River at Bahadurabad based on precipitation changes projected by nine GCMs and three SSPs and four target temperature changes
Full size table
Fig. 16.7
Three box plots depict the projected streamflow of the Brahmaputra River Basin at Bahadurabad station. The values are plotted for the models S S P 126, S S P 245, and S S P 585. All the models depict an increasing trend.

Projected future (2080s) peak discharge of the Brahmaputra River at Bahadurabad based on precipitation changes projected by nine GCMs and three SSPs and four target temperature changes. The cross marks demonstrate the ensemble mean

Full size image
Fig. 16.8
Three box plots depict the projected water level of the Brahmaputra River Basin at Bahadurabad station. The values are plotted for the models S S P 126, S S P 245, and S S P 585. All the models depict an increasing trend.

Projected future (2080s) peak stage of the Brahmaputra River at Bahadurabad based on precipitation changes projected by nine GCMs and three SSPs and four target temperature changes. The cross marks demonstrate the ensemble mean. The dotted line indicates the current danger level (flood stage) for the Brahmaputra River at Bahadurabad, Bangladesh

Full size image

The baseline (climate normal period) maximum streamflow of the Brahmaputra River at Bahadurabad is 47,110 m3/s (Alam et al., 2021). Note: this peak streamflow does not represent the instantaneous peak flow over the climate normal period rather represents maximum discharge averaged over the climate normal period. The corresponding river stage at Bahadurabad, as derived from the rating cure, is 18.98 m Public Works Datum (PWD).

It was found that for all SSPs, the projected maximum discharge, and corresponding river stage increase linearly with the target warming scenarios. This aligns with the results from Alam et al. (2016, 2021). For SSP126, an increase of 1-, 1.5-, 2- and 2.5-degree in basin-scale temperature in the 2080s will potentially increase the maximum streamflow at the Bahadurabad location by 17%, 20%, 23%, and 26%, respectively (ensemble mean). For SSP245, an increase of 1, 1.5, 2, and 2.5 degree in basin-scale temperature in the 2080s will potentially increase the maximum streamflow at the Bahadurabad location by 18%, 21%, 24%, and 27%, respectively. For SSP585, an increase of 1, 1.5, 2, and 2.5 degree in basin-scale temperature in the 2080s will potentially increase the maximum streamflow at the Bahadurabad location by 36%, 39%, 41%, and 44%, respectively. We found that the projected maximum discharge in the 2080s for SSPs is relatively higher than those estimated using CMIP5 scenarios by Alam et al. (2021). Alam et al. (2021) showed maximum discharge will increase by 16% (median) for ensemble scenarios. However, we find maximum discharge increase is even higher for a relatively optimistic scenario (SSP126) and a 1-degree increase in temperature.

Because of the nonlinear relationship between streamflow and stage, the percentage changes in river stage compared to the baseline stage of 18.98 m PWD range between 2.06% and 4.87% for various target temperature and SSPs. For SSP126, a 1-, 1.5-, 2-, and 2.5-degree increase in basin-scale temperature in the 2080s will potentially increase the maximum river stage at the Bahadurabad location by 2.06%, 2.39%, 2.70%, and 3.01%, respectively (ensemble mean). For SSP245, a 1-, 1.5-, 2-, and 2.5-degree increase in basin-scale temperature in the 2080s will potentially increase the maximum river stage at the Bahadurabad location by 2.14%, 2.46%, 2.78%, and 3.09%, respectively. For SSP585, a 1-, 1.5-, 2-, and 2.5-degree increase in basin-scale temperature in the 2080s will potentially increase the maximum river stage at the Bahadurabad location by 4%, 4.30%, 4.59%, and 4.87%, respectively. According to the Flood Forecasting and Warning Centre, Bangladesh Water Development Board (http://www.ffwc.gov.bd/), the Flood Danger Level (FDL) at Bahadurabad is 19.50 m PWD, and the recorded Highest Water Level (HWL) at Bahadurabad is 21.16 m PWD. Our analysis shows that for the SSP126, the projected river stage (ensemble mean) exceeds the FDL for a 2.5-degree increase in basin-scale temperature in the 2080s. For SSP245, the projected river stage exceeds the FDL for a 2-degree and 2.5-degree increase in basin-scale temperature in the 2080s. For the SSP585, the projected river stage exceeds the FDL for all warming scenarios. However, it should be noted that this projected river stage only demonstrates the mean value of the maximum river stage for the 2080s, not the instantaneous river stage.

The results of our study have several practical implications. First, the changes in the amount of extreme precipitation in the future under all three scenarios will have direct implications in terms of the availability of water resources in the BRB. If managed properly at the basin-scale through integrated water resources management involving all co-riparian countries, such an increase in water resources can be utilized to augment flow during lean season and can be harnessed for multipurpose uses. Second, such an increase in water availability will result in untimely flooding that will have to be managed to reduce the damage to property and ecosystems. Third, increased flooding propensity will cause damage to crops, life, and livelihood in the BRB. The existing flood control and mitigation efforts will have to be retrofitted to tackle additional flood depth, duration, and frequency. Fifth, collaboration among co-riparian countries in the BRB will be essential for optimal utilization of water resources and for minimizing future damage caused by increased severe flooding in the BRB.

5 Conclusions

  1. 1.

    The ensemble mean of the percentage changes in basin-scale mean annual precipitation for the 2080s compared to the climate normal period of 1981–2010 for the SSP126, SSP245, and SSP585 are 7.48%, 8.01%, and 19.37%. All future scenarios project a moderate increase in basin-scale precipitation in the early to mid-twenty-first century (2015–2050). From 2050 onwards, SSP585 projects a significant increase in mean annual precipitation, and SSP126 and SSP245 show a moderate increase. The uncertainties associated with the projections grow as we move to the distant future.

  2. 2.

    The tail index determined for precipitation extremes shows that by the 2080s the BRB will experience more intense precipitation under various SSPs as compared to the historic period. The extreme precipitation will increase under all SSPs.

  3. 3.

    The ensemble mean of percentage changes in peak discharge at Bahadurabad is higher for higher warming scenarios. Based on SSP585 and 1.5-degree warming, the peak discharge will potentially increase by 39% which will increase the flood risk in Bangladesh.

Considering the complex topography of the study region, further downscaling is recommended before using them to quantify the climate change impact assessments on natural and built systems. Fine-scale processes are known to influence the regional distribution of precipitation changes. Therefore, the use of high-resolution climate scenarios as generated by Regional Climate Models is also recommended.