The runoff from the Upper Indus River Basin (UIB) is considered vital for sustainable water supply and agriculture. Climate change may alter the runoff and affect the availability of water. This study employs a hydrologic model and the climate projections of fine-resolution Meteorological Research Institute-Atmospheric General Circulation Model (MRI-AGCM) (0.1875°) to assess the impacts of climate change in the future (2075–2099) on the flows in the UIB and its sub-basins. The simulations have shown satisfactory results compared to the observed ones. Furthermore, simulated snow-cover was compared with estimated snow-cover by Moderate Resolution Imaging Spectroradiometer (MODIS). According to the results, meltwater from glaciers contributes to about two-thirds of the annual runoff from the UIB. According to the projections, the future temperature will increase by 5.3 °C and annual precipitation will increase by 17% across the UIB. Despite this increase, the annual river flows were projected to decrease by 14% in the future because of depleted glacier melt. Furthermore, the detailed hydrologic responses in selected sub-basins showed different patterns in the future depending on the climate and elevation. In an eastern sub-basin, flows will increase by 58% because of enhanced glacier melt and higher precipitation. Contrarily, two western sub-basins will experience the reduction of the flow by 30–70% because of the glacier retreat.

  • Runoff in the Upper Indus River Basin (UIB) is melt-dependent.

  • The precipitation regime in the east is monsoon dominated while in the west it is snowfall dependent.

  • Hydrologic response across the UIB is contrasting because of climate and geography.

  • Owing to the fine resolution of the MRI-AGCM, distinct climatic features across the UIB have been projected.

  • Climate change and glacier retreat will alter the quantity and timings of the hydrographs in the future.

APHRODITE

Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation

GCM

General Circulation Model

ICIMOD

International Centre for Integrated Mountain Development

MODIS

Moderate Resolution Imaging Spectroradiometer

MRI-AGCM

Meteorological Research Institute-Atmospheric General Circulation Model

RCM

Regional Climate Model

RCP

Representative concentrative pathway

RRI

Rainfall–runoff–inundation

UIB

Upper Indus River Basin

The Indus River Basin covers an area of 1,140,000 km2 in Afghanistan, China, India, and Pakistan. The length of the river is approximately 2,800 km, and its origin lies in Tibet (Figure 1) (Ali 2013). The Upper Indus River Basin (UIB) supplies river flows that are essential for one of the world's largest contiguous canal irrigation systems (Lutz et al. 2016).
Figure 1

Map of the UIB with its outlet at Bisham Qila.

Figure 1

Map of the UIB with its outlet at Bisham Qila.

Close modal

Climate in the UIB is quite variable, in winter (October to March) temperatures often remain sub-freezing where average temperature remains around 9 °C, while the average temperature in summer (May–September) is approximately 13 °C (Mcsweeney et al. 2012). Reggiani & Rientjes (2015) used the mass-balance technique and estimated the total annual precipitation in the basin to be 681 ± 100 mm. Similar estimates of total annual precipitation have been reported by Dahri et al. (2018) and Immerzeel et al. (2015). In winter, precipitation falls as snow which accumulates in the season and melts in spring; this meltwater from snow and glaciers results in higher flows from July to September (Archer 2003; Mukhopadhyay & Khan 2014). At the downstream of Bisham Qila, the Tarbela Reservoir supplies water essential for irrigation, stores flood water, and generates hydropower (Figure 1) (Lutz et al. 2016).

The hydrological cycle of the UIB is under stress because of climate change. Rising temperatures will cause more water to evaporate, which will increase the moisture content in the atmosphere and render the spatiotemporal trends in precipitation more variable (Wester et al. 2019). In addition to precipitation variability, the increased temperature will alter the hydrological cycle in the UIB (Immerzeel et al. 2010). The topographic features and climate vary across the UIB which controls the hydrologic cycle. There is a need to explore the hydrology at sub-basins in the UIB to understand the shift in the quantity and timings of the runoff under the future climate. Future climate change in the Himalayan–Karakoram region, where the UIB is located, has been projected by Sanjay et al. (2017) using the dynamic downscaling of Regional Climate Models (RCMs). According to these projections, temperatures will increase between 2.4 and 3.6 °C by the mid-century scenario and between 3.0 and 5.0 °C by the end of this century. The annual precipitation will increase in the range of 3.5–12.8% in the mid- and late centuries in the region. The two studies that employed a fine-resolution General Circulation Model (GCM) (∼25 km) and the latest CMIP6 have projected wetter summers in the UIB in the future (Ali et al. 2021; Abbas et al. 2022). Glaciers in the region will not be able to retain their mass despite the increase in precipitation because the increase in temperatures will make the current mass balance negative (Lutz et al. 2016). This situation will result in an increase in glacier melt rates in the coming decades until the last quarter of this century (Rounce et al. 2020).

The hydrologic analysis in UIB has used various techniques and datasets. Often global datasets are employed owing to the unavailability of the in situ data. Lutz et al. (2016) and Su et al. (2016) included glacier mass-balance modeling in the hydrologic analysis to evaluate the impacts of climate change in the UIB and its sub-basins. The simulations demonstrated the contribution of glaciers along with snow and rainfall which helped in understanding the seasonal changes in the stream flow discharge. Ougahi et al. (2022) and Shah et al. (2020) used GCMs to project the river flows in the UIB using the SWAT hydrologic model. They have projected that river flows will increase in the mid- and late century, although the rate of increase will be milder in the later century. Higher increases are projected in the spring and summer because of earlier snow melt and precipitation. Kiani et al. (2021) focused on the temperature increase of 1.5–2 °C as the critical threshold in the UIB. Under both the threshold scenarios, the annual river flows are projected to increase in the range of 11–26% because of the earlier onset of snow and glacier melt. Similar temperature thresholds were used by Hasson et al. (2019) to evaluate the hydrologic response in the UIB. However, their modeling approach uses a variety of glacier coverage scenarios ranging from 0 to 100%. Their simulations show that annual river flows will increase in the case of the intact glacier while decreasing in all other glacier coverage scenarios. Ali et al. (2021) used two finer resolution climate models in the UIB, the results suggested that river flows will continue to increase in the mid-century (50–60%) and late century (83–87%). The finer resolution enables better understanding of the climatic variability and accurate hydrologic simulations across the UIB.

Most previous studies have used coarse-resolution GCMs focusing on larger river basins. Although such studies have shown reliable runoff projections over larger basins, regional conditions are often neglected. The coarse resolution climate models may be insufficient to simulate the complex climatic processes like monsoon in high mountainous regions, where higher resolution RCMs are essential (Sanjay et al. 2017). The climate simulated with RCMs in the Himalayas has shown well-represented physical processes like moisture transport (Dimri et al. 2013). The Atmospheric General Circulation Model (AGCM) provides climate projections at fine resolutions and shows the influence of monsoon and winter precipitation across the UIB.

Our study focuses on the entire UIB as well as three representative sub-basins to reveal their responses to climate change. The three selected sub-basins have different topography and climate characteristics, which are ideal for understanding inter-basin hydro-climatic variability. This study employs the Meteorological Research Institute-Atmospheric General Circulation Model (MRI-AGCM) with a spatial resolution of 0.1875°. The model simulates four scenarios based on different spatial patterns in sea surface temperatures (Mizuta et al. 2012). The fine-spatial resolution has enabled us to simulate the distinct spatial patterns of precipitation across the UIB (Supplementary material, Figure S1). The objectives of this study are as follows:

  • Simulate the hydrological cycle of the UIB and its sub-basins with a distributed hydrological model forced by climate datasets. Snow and glacier melt models are coupled with the hydrological model based on geographic datasets including land use and glacier cover datasets. By conducting various scenario simulations, we quantify the runoff components of river flows from different sources including rainfall, snowmelt, and glacier melt.

  • Assess the impacts of climate change on the stream flow of the UIB and its sub-basins. The runoff components are projected under future climate conditions to better understand the future stream flow changes in the entire UIB and its sub-basins.

The next section explains the study area including the topography and climate followed by the methodology with detailed descriptions of the models used in this study. The results and conclusions are presented in the last two parts.

The UIB has an area of 192,861 km2, and the mountainous region has an elevation range of ∼6,800 m (Figure 2). The total glacier area in the basin is 14,087 km2, and most of the mass is concentrated in the northwestern region (ICIMOD 2011). The UIB receives runoff from the Himalayas, Karakoram, and Hindukush mountains, which mostly consist of melt from snow and glaciers (Lutz et al. 2014). Supplementary material (Figures S2 and S3) shows the spatial distribution of glaciers in the basin. To assess the climatic and runoff conditions at the regional level, three representative sub-basins across the UIB, namely, Gilgit, Shigar, and Kharmong were selected. These sub-basins cover the entire geographic range and differ in terms of glacier coverage and climate (Figure 2). The Gilgit sub-basin is located at the western edge of the UIB and has an elevation of 1,472–6,392 m. Glacier coverage in the basin is 1,684 km2 (13.2%) with almost 1,181 km2 (70%) of the glaciers located in the elevation band of 4,000–5,000 m. The Kharmong sub-basin is located in the eastern region of the UIB, it is 70,000 km2 and 3.6% of which is glacier covered. The Shigar sub-basin lies in the northern region and has an area of 6,900 km2. It has the largest glacier coverage (28.8%) among the sub-basins (Table 1).
Table 1

Area, elevation, and glaciers of the UIB and sub-basins

BasinArea (km2)Elevation (min, mean, and max) (m)Glacier (%)
UIB 192,861 733, 4,600, 7,528 7.67 
Gilgit 13,100 1,472, 4,030, 6,392 13.2 
Shigar 6,900 2,189, 4,520, 7,360 28.8 
Kharmong 70,000 2,409, 4,790, 6,470 3.6 
BasinArea (km2)Elevation (min, mean, and max) (m)Glacier (%)
UIB 192,861 733, 4,600, 7,528 7.67 
Gilgit 13,100 1,472, 4,030, 6,392 13.2 
Shigar 6,900 2,189, 4,520, 7,360 28.8 
Kharmong 70,000 2,409, 4,790, 6,470 3.6 
Figure 2

The UIB and its sub-basins with their elevations and outlets.

Figure 2

The UIB and its sub-basins with their elevations and outlets.

Close modal

The observed monthly discharge (1980–2005) at the Bisham Qila is shown in Supplementary material, Figure S4. The hydrograph shows a peak in summer (May–September) which is due to the contribution of the meltwater (Mukhopadhyay & Khan 2014). In winter (October–April), the average monthly flow is sustained at approximately 450 m3/s. This base flow is mainly due to the remnant melt and rainfall runoff, as explained by Mukhopadhyay & Khan (2015). For the sub-basin analysis, the river flow data of three sub-basins have also been acquired (Figure 2).

For hydrological modeling of river flows, the 0.25° Asian Precipitation Highly Resolved Observational Data Integration Towards Evaluation (APHRODITE) (Yatagai et al. 2012) climate dataset was used along with glacier cover (ICIMOD 2011). Figure 3 and Supplementary material (Figure S5) show the average precipitation and temperature in the basin. The original APHRODITE dataset of precipitation was corrected using the precipitation extrapolation technique proposed by Immerzeel et al. (2015). They used glacier mass balance to estimate the precipitation gradient in the basin. Furthermore, these gradients are used to correct the precipitation bias in the APHRODITE dataset. The difference between our study and Immerzeel et al. (2015) is that they have done spatial downscaling of the APHRODITE dataset and performed bias-correction on the sub-basin level within the UIB. According to the dataset, winter (November–April) receives more precipitation than summer (May–October). Higher elevation zones receive heavy precipitation in winter in the form of snowfall. The eastern edge of the UIB is shown to have a high rate of precipitation in summer because of the monsoon influence (Figure 3). The snow cover for the calibration of the model is extracted from the MODIS product (MOD10A2) with a 500-m resolution (Hall & Riggs 2001).
Figure 3

Spatial distribution of average temperature and precipitation in the UIB (1980–2005).

Figure 3

Spatial distribution of average temperature and precipitation in the UIB (1980–2005).

Close modal
The rainfall–runoff–inundation (RRI) model introduced below was run with snow and glacier melt models. The rainfall, snowmelt, and glacier melt were calculated for each grid cell of the APHRODITE climate data using daily time series. For snow and glacier melt, separate degree-day models were used which require daily precipitation, daily temperature, and topographic information (Figure 4).
Figure 4

Flow chart explaining the data and models used in this study.

Figure 4

Flow chart explaining the data and models used in this study.

Close modal

The simulation period consisted of 26 years from 1980 to 2005. The simulation was divided into two periods, one for calibration and the other for validation. The calibration covered 13 years from 1980 to 1992, whereas the validation covered the remaining years. The model was calibrated and validated with observed river flows at Bisham Qila, and the performance was checked using several statistical parameters. The river flows at sub-basins were validated with the observed river flows and simulated snow covers were compared with the MODIS snow cover.

Cema-Neige snow melt model

The Cema-Neige snow melt model was adopted from Valéry et al. (2014). This is a degree-day melt model that uses mean air temperature (Tmean) to distinguish between rainfall and snowfall. The main outputs of the model are rainfall and snowmelt for each grid cell of the APHRODITE data. In this application, if the temperature is below 0 °C, all precipitation is considered to be snow, otherwise it is treated as rainfall when the temperature is above 0 °C.

The snowpack temperature (Snowpacktemp) defines the internal thermal state of the snowpack, which is used to quantify the melt. If the internal temperature rises to 0 °C, the melt takes place. Equation (1) estimates the snowpack temperature.
(1)
where Snowpacktemp,t is the snowpack temperature at time step t in °C and X is the snowpack inertia factor, which is set by calibration. The potential melt Meltpot (mm/day) is computed when the snowpack temperature reaches 0 °C and the daily mean air temperature is greater than 0 °C.
(2)
where DDFs (mm°C−1 day−1) is the degree-day factor. The melt cannot exceed snow storage, i.e. Meltpot is limited to the snow storage. The accumulation of snow is updated daily based on the previously stored snow and the sum of solid precipitation as Psnow as indicated by the following equation.
(3)
where SSt (mm) is the snow storage at time step t updated after the accumulation or melt of snow (mm), and Meltact is the actual melt (mm/day). Then actual melt Meltact (mm/day) is estimated by the following empirical expression (Equation (4)). The snow cover area is also employed in this function as follows:
(4)
The snow cover area is a unique feature of the model. The model uses Psnow and annual average snowfall to estimate the percentage of the river basin covered with snow (Equation (5)).
(5)
where Z is the average annual snow precipitation (mm). More detailed information on Cema-Neige can be found in Valéry et al. (2014).

Glacier melt model

For the present climate glacier melt calculation, we used the glacier coverage estimated by ICIMOD (2011). For each grid of the APHRODITE, the glacier cover (%) is calculated and the glacier melt was quantified using a degree-day model explained by Terink et al. (2015). Equation (6) was used to calculate the daily melt from clean ice and debris-covered glaciers.
(6)
where MeltCI/DC refers to the daily glacier melt from clean ice and debris-covered glacier, DDFCI/DC (mm°C−1 day−1) are the degree-day factors and FCI/DC are the proportion of clean ice and debris-covered glaciers in the grid. The degree-day factors are set by the model calibration. The total glacier melt is the sum of both MeltCI and MeltDC.

Glacier area

In this study, the future glacier area was extracted from the estimates given by Huss & Hock (2015). Their study simulated the average glacier area in the future using 14 GCMs for the western parts of South Asia, where the UIB is located. The projected average glacier area for the entire western South Asia by 2075–2099 is 9,144 km2. This is roughly 28% of the present area, which is 32,814 km2 (Supplementary material, Figure S6). In this study, we assume the same ratio of glacier loss in the UIB region as estimated by Huss & Hock (2015) for the entire region of western South Asia. Thus, we assume that the current glacier area of 14,068 km2 in the UIB will reduce to 3,521 km2 by 2075–2099. In terms of the spatial variation of the glacier cover under future climate conditions, we also theorize that the glacier area will mainly be lost from the lowest elevation areas due to the higher temperatures; this pattern of glacier evolution is confirmed by Huss et al. (2010).

It should be noted that there are several glacier mass balance models, e.g. volume–area scaling, finite-element ice-flow models, etc. In addition to this, Huss et al. (2010) introduced empirical formulae to quantify the change in the thickness of glaciers. Their formulae are the result of field studies and topography analysis of 34 glaciers of varying sizes. Furthermore, they compared the results with a three-dimensional (3D) ice-flow model and found the empirical model performs satisfactorily. Compared to those physically based glacier models, our approach here based on temperature-based melt models is fairly simple. However, the advantage of the approach in this study can be based on the robust glacier area projected by the previous study, in this case by Huss & Hock (2015), and the use of the consistent degree-day factors for both glacier and snow melt, which enables us to interpret the results more directly.

GCM downscaling

The delta-change method was selected to remove the biases in temperature present in the MRI-AGCM. This method is a widely used technique that considers the mean bias as the difference between the observed and GCM data (Miao et al. 2016). The future projection can be adjusted as (Equation (7)):
(7)
where x is the meteorological variable of either o (observed) or m (modeled) for a historic training period or current climate (c) or future projection . Linear correction is another bias-correction technique that utilizes a scaling factor between observed and GCM simulations to reduce the bias in future projection. This is used for the bias correction of precipitation data as suggested by many previous studies, for example Miao et al. (2016) (Equation (8)).
(8)

Future climate scenarios in the UIB and sub-basins are explained with the help of figures in the results section.

RRI model

The outputs of the Cema-Neige model (rainfall and snow melt) and glacier melt model were used for the simulation of the river flows using the RRI model. The original RRI model uses only rainfall and potential evapotranspiration as the input together with DEM and land use. In this study, we used the Cema-Neige model to first distinguish precipitation between rainfall and snowfall and update the snow coverage and snow melt as well as evapotranspiration. In the case of no-snow coverage with rainfall, we give the rainfall and glacier melt as the input of the RRI model. In the case of snow coverage existence, we give either rainfall or snowmelt depending on the form of precipitation along with the glacier melt as the input of the RRI model. The RRI model is a two-dimensional (2D) model that is capable of simultaneously representing rainfall runoff and flood inundation (Sayama et al. 2012). The flow on the slope grid cells was calculated using the 2D diffusive wave model, while the channel flow was calculated using the 1D diffusive wave model.

For better representations of RRI processes, the model simulates lateral subsurface flow, vertical infiltration flow, and surface flow. The vertical infiltration flow was estimated using the Green–Ampt model (Sayama 2015).

Model calibrations and validations

The simulated discharge at the Bisham Qila and sub-basins are shown in Figure 5, Supplementary material (Figures S7, S8, and S9), and the model performance is summarized in Table 2. Both the simulated and observed hydrographs show good agreement. For checking the performance of the simulations Nash–Sutcliffe (NS) coefficient, root-mean-square error (RMSE), and percent bias (PBIAS) have been calculated. The NS of 0.54–0.6 at the Bisham Qila is higher than the sub-basins indicating room for improvement in the simulation process.
Table 2

Nash–Sutcliffe (NS) coefficient, PBIAS, and root-mean-square error (RMSE) between the simulations and observed river flows in the UIB and sub-basins

StationCalibration
Validation
NSPBIASRMSE (m3/s)NSPBIASRMSE (m3/s)
Bisham 0.60 0.35 0.58 0.54 0.32 0.61 
Gilgit 0.53 0.035 0.75 0.57 0.04 0.7 
Shigar 0.50 0.3 0.7 0.48 0.27 0.68 
Kharmong 0.45 0.28 0.72 0.41 0.32 0.64 
StationCalibration
Validation
NSPBIASRMSE (m3/s)NSPBIASRMSE (m3/s)
Bisham 0.60 0.35 0.58 0.54 0.32 0.61 
Gilgit 0.53 0.035 0.75 0.57 0.04 0.7 
Shigar 0.50 0.3 0.7 0.48 0.27 0.68 
Kharmong 0.45 0.28 0.72 0.41 0.32 0.64 
Figure 5

Calibration and validation results of the hydrological model at Bisham Qila.

Figure 5

Calibration and validation results of the hydrological model at Bisham Qila.

Close modal
A comparison of snow-covered areas between the Cema-Neige simulations and MODIS imagery is shown in Figure 6, which shows snow cover from 2000 to 2004.
Figure 6

Comparison of MODIS snow cover with simulated snow cover.

Figure 6

Comparison of MODIS snow cover with simulated snow cover.

Close modal

The Cema-Neige model does not consider the slope and aspect of the grid cell in the simulation, which is not the exact representation of the physical process. In addition, an important parameter of the Cema-Neige model is the division of precipitation into rainfall and snowfall, which is based on using a 0 °C threshold. These factors are the sources of disagreement between the model results and the actual snow cover. Overall, simulations show satisfactory results, but improvements are required in the calibration of summer flows.

Estimation of snow and glacier contributions under present climate

Figure 7 shows the monthly contribution of rainfall, snowmelt, and glacier melt to the discharge. Under the current climate conditions, glaciers contribute 68%, rainfall 20%, and snowfall 12% of river flows. Rainfall runoff is generated mainly from May to October, and the maximum rainfall discharge occurs in September. With the increasing temperature in April, snow starts to melt from the lower elevation zones, and its discharge reaches approximately 1,000 m3/s in April and approximately 2,400 m3/s in May. The river flows continue to increase in the following months because of the inclusion of glacier melt. In summer from June to September, glacier melt dominates the hydrograph and contributes 85% of the total flow.
Figure 7

The monthly contribution of glacier, snow, and rain in the UIB (1980–2005).

Figure 7

The monthly contribution of glacier, snow, and rain in the UIB (1980–2005).

Close modal

Lutz et al. (2014) reported the annual glacier contribution in the UIB at the Bisham Qila to be 67%, which is similar to that of our simulations (Table 3). Upon the comparison of runoff components with other studies, differences among estimates were found because of different scales of the study area, estimation approaches, datasets, etc.

Table 3

Summary of precipitation (P) (mm), temperature (T) (°C), runoff (Q), and glacier coverage (%) in (a) the present (1980–2005) and (b) the future (2075–2099)

StationsP (mm)T (°C)Q (mm)Glacier (Qg)Snow (Qs)Rain (Qr)Glacier (%)
(a) Present (1980–2005) 
Bisham 440 1.77 656 367 (68) 62 (12) 105 (20) 7.67 
Gilgit 491 1.34 962 614 (75) 77 (10) 123 (15) 13.2 
Shigar 450 −1.67 2,400 1,805 (85) 145 (7) 151 (8) 28.8 
Kharmong 405 2.18 226 57 (35) 42 (26) 63 (39) 3.6 
(b) Future (2075–2099) 
Bisham 516 7.4 566 280 (60) 33 (7) 150 (33) 1.83 
Gilgit 553 7.9 272 36 (17) 41 (19) 136 (64) 
Shigar 642 6.7 1,640 1,128 (81) 87 (6) 176 (13) 12.5 
Kharmong 453 6.6 357 137 (50) 20 (7) 116 (43) 3.5 
StationsP (mm)T (°C)Q (mm)Glacier (Qg)Snow (Qs)Rain (Qr)Glacier (%)
(a) Present (1980–2005) 
Bisham 440 1.77 656 367 (68) 62 (12) 105 (20) 7.67 
Gilgit 491 1.34 962 614 (75) 77 (10) 123 (15) 13.2 
Shigar 450 −1.67 2,400 1,805 (85) 145 (7) 151 (8) 28.8 
Kharmong 405 2.18 226 57 (35) 42 (26) 63 (39) 3.6 
(b) Future (2075–2099) 
Bisham 516 7.4 566 280 (60) 33 (7) 150 (33) 1.83 
Gilgit 553 7.9 272 36 (17) 41 (19) 136 (64) 
Shigar 642 6.7 1,640 1,128 (81) 87 (6) 176 (13) 12.5 
Kharmong 453 6.6 357 137 (50) 20 (7) 116 (43) 3.5 

Note: Runoff contributions by glacier (Qg), snow (Qs), and rain (Qr) are expressed in mm. Values in the parenthesis show the percentage of the given components in the runoff.

Sub-basin river simulations

In total, runoff from these three sub-basins contributes to 40% of river flow at the Bisham Qila station, and in the glacier contribution, the sub-basins’ share is 34%. Figures S7, S8, and S9 show simulations of discharge for Gilgit, Shigar, and Kharmong, respectively. They are highly glacierized, and the runoff is meltwater-dependent. The common characteristics of runoff in the three sub-basins in the UIB are the significant proportion of melt-runoff, which explains the significance of winter precipitation in the basin (Table 3).

In Gilgit and Shigar, over three-fourths of the annual river flows are contributed by meltwater alone because of heavy winter precipitation and larger glacier covers. In Kharmong, rainfall runoff is higher than snow and glacier melt because of the stronger influence of monsoon precipitation (Table 3).

Future climate in the UIB

Figure 8 shows the change in basin temperature and precipitation during 2075–2099 based on the four MRI-AGCM scenarios with different SST patterns. Across the UIB, the temperature will increase by 5.3 °C, and the largest increase of 6.6 °C is projected in June, July, and August (JJA). In December, January, and February (DJF), the temperature will increase by 4.5 °C. The comparison with other studies in Table 4 suggests that the increase in annual temperature is similar to the other studies based on different GCMs or RCMs. Note that the annual temperature rise reported by Su et al. (2016) is relatively smaller (+3.8 °C) but this is mainly due to their target period that is between 2041 and 2070.
Table 4

A comparison of climate change impact studies in the UIB with this study

This study(a) Lutz et al. (2016)) (b) Su et al. (2016) (c) Hasson et al. (2019) (d) Ali et al. (2015) (e) (Babur et al. 2016)(f) Khan et al. (2020) 
Future period to be studied 2075–2099 2071–2100 2041–2070 2070–2099 2071–2100 2071–2100 2071–2100 
The definition of UIB Upstream of Bisham Qila Upstream of Bisham Qila and other sub-basins including the Kabul, Jhelem, Chenab, Satluj.
At Bisham Qila for the discharge 
The region in UIB which is above 2,000 m Upstream of Bisham Qila and the Kabul river basin Upstream of Bisham Qila Mangla River Basin in the south of UIB Upstream of Bisham Qila 
Used GCMs and scenarios MRI-AGCM, RCP8.5, four SSP patterns Four selected GCMs (MPI-ESM-LR, IPSL-CM5A-LR, CSIRO-Mk3-6 − 0, MIROC5), RCP8.5 (in this table) 20 GCMs from CMIP5, RCP8.5 (in this table) RegCM4.3, RCP8.5 (in this table) Two GCMs (CCAM and RegCM), RCP8.5 (in this table) 7 GCMs (RCP4.5 and 8.5) 5 CORDEX RCMs 
Glacier meltmodels Future glacier cover is estimated based on Huss & Hock (2015) and temperature-index for the melting Glacier mass-balance model included based upon temperature-index Glacier mass-balance model included based upon temperature-index No direct glacier meltmodel but a temperature-index snow meltmodel No direct glacier melt model but an energy-based snow meltmodel SWAT model which has no direct glacier model SWAT model which has no direct glacier model 
Method to estimate the glacier coverage for present climate ICIMOD database Randolph glacier Inventory Randolph glacier Inventory – – – – 
Projected annual temperature change (°C) +5.3(+5.14 ∼ +5.64) +6.5(+5.5 ∼ +8.0) +3.8 +5.1 +5.8 +5.4 +6.2 
Projected annual precipitation change (%) +17.0(10–21) +10.6 (−25 to +125) +4.5 30 to +40 +20.7 +26 +38 
Projected annual discharge change (%) −14(−16 to −10) 30(−15 to +60) 6–22 17.4 87 −27 to +74 +8(−27 to +27) 
Monthly temperature change (°C)        
DJF 4.6 6.3 3.9 3.6 6.5 5.58 - 
MAM 4.6 6.0 3.7 3.7 6.1  
JJA 6.6 6.15 3.6 5.4 5.9 4.72  
SON 5.3 5.8 3.7 3.9 5.05 5.28  
Monthly precipitation change (%)        
DJF 12 16 −20 3.5 0.58  
MAM 14 −18 −2 18 32.9 20  
JJA 35 15 13 33 13 53  
SON −1.7 40 2.5 25 33.2 58  
Monthly discharge change (%)        
DJF 35 100 20 15 147 104 50 
MAM −12 110 47 264 300 20 200 
JJA −21 22 57 14 −46 
SON −5 60 29 124 52 43 
This study(a) Lutz et al. (2016)) (b) Su et al. (2016) (c) Hasson et al. (2019) (d) Ali et al. (2015) (e) (Babur et al. 2016)(f) Khan et al. (2020) 
Future period to be studied 2075–2099 2071–2100 2041–2070 2070–2099 2071–2100 2071–2100 2071–2100 
The definition of UIB Upstream of Bisham Qila Upstream of Bisham Qila and other sub-basins including the Kabul, Jhelem, Chenab, Satluj.
At Bisham Qila for the discharge 
The region in UIB which is above 2,000 m Upstream of Bisham Qila and the Kabul river basin Upstream of Bisham Qila Mangla River Basin in the south of UIB Upstream of Bisham Qila 
Used GCMs and scenarios MRI-AGCM, RCP8.5, four SSP patterns Four selected GCMs (MPI-ESM-LR, IPSL-CM5A-LR, CSIRO-Mk3-6 − 0, MIROC5), RCP8.5 (in this table) 20 GCMs from CMIP5, RCP8.5 (in this table) RegCM4.3, RCP8.5 (in this table) Two GCMs (CCAM and RegCM), RCP8.5 (in this table) 7 GCMs (RCP4.5 and 8.5) 5 CORDEX RCMs 
Glacier meltmodels Future glacier cover is estimated based on Huss & Hock (2015) and temperature-index for the melting Glacier mass-balance model included based upon temperature-index Glacier mass-balance model included based upon temperature-index No direct glacier meltmodel but a temperature-index snow meltmodel No direct glacier melt model but an energy-based snow meltmodel SWAT model which has no direct glacier model SWAT model which has no direct glacier model 
Method to estimate the glacier coverage for present climate ICIMOD database Randolph glacier Inventory Randolph glacier Inventory – – – – 
Projected annual temperature change (°C) +5.3(+5.14 ∼ +5.64) +6.5(+5.5 ∼ +8.0) +3.8 +5.1 +5.8 +5.4 +6.2 
Projected annual precipitation change (%) +17.0(10–21) +10.6 (−25 to +125) +4.5 30 to +40 +20.7 +26 +38 
Projected annual discharge change (%) −14(−16 to −10) 30(−15 to +60) 6–22 17.4 87 −27 to +74 +8(−27 to +27) 
Monthly temperature change (°C)        
DJF 4.6 6.3 3.9 3.6 6.5 5.58 - 
MAM 4.6 6.0 3.7 3.7 6.1  
JJA 6.6 6.15 3.6 5.4 5.9 4.72  
SON 5.3 5.8 3.7 3.9 5.05 5.28  
Monthly precipitation change (%)        
DJF 12 16 −20 3.5 0.58  
MAM 14 −18 −2 18 32.9 20  
JJA 35 15 13 33 13 53  
SON −1.7 40 2.5 25 33.2 58  
Monthly discharge change (%)        
DJF 35 100 20 15 147 104 50 
MAM −12 110 47 264 300 20 200 
JJA −21 22 57 14 −46 
SON −5 60 29 124 52 43 

SON, September, October, and November.

Figure 8

Change in (a) temperature and (b) precipitation across the Upper Indus River Basin according to MRI-AGCM. Change in (c) temperature and (d) precipitation on a monthly basis in the future (2075–2099).

Figure 8

Change in (a) temperature and (b) precipitation across the Upper Indus River Basin according to MRI-AGCM. Change in (c) temperature and (d) precipitation on a monthly basis in the future (2075–2099).

Close modal

In terms of annual precipitation change, our study projects its increase by 17%. In particular, the precipitation in the JJA season shows a higher increase by 35%, indicating a wet summer season. Furthermore, the winter precipitation is also projected to increase which could increase the snow depth and spring runoff. The projections of annual precipitation summarized in Table 4 show higher variations among different GCMs and RCMs by different studies. For example, four selected GCMs by Lutz et al. (2016) indicated the range between −25 and 125% depending on the GCMs.

The spatial pattern in future temperature change is fairly homogeneous across sub-basins (Figure 9). The difference among the SST scenarios was also not very evident. On the other hand, the precipitation changes according to the MRI-AGCM are spatially more variable. In winter (October–March), precipitation will increase in most of the basins with some differences in the magnitudes. For the season between April and September, the eastern region of the UIB is projected to increase even up to 100% and the western part of the UIB showed some decreasing trend in the seasonal precipitation (Figure 10). Among the sub-basins we selected, the Kharmong sub-basin located in the eastern part of the UIB may experience a higher increase in the precipitation from July to September (Figure 10). On the contrary, the Shigar sub-basin located in the northern part of the UIB may show a decrease in precipitation in some seasons.
Figure 9

Increase in temperature across four MRI-AGCM scenarios in the future based on seasons in the UIB.

Figure 9

Increase in temperature across four MRI-AGCM scenarios in the future based on seasons in the UIB.

Close modal
Figure 10

Change in precipitation across four MRI-AGCM scenarios in the future based on seasons in the UIB.

Figure 10

Change in precipitation across four MRI-AGCM scenarios in the future based on seasons in the UIB.

Close modal
The trends in the change in the spatial features of the future precipitation can be confirmed by Figure 11 which shows the projected monthly changes in temperature and precipitation in the UIB and its three sub-basins. There is a consensus among the scenarios in terms of temperature change in the three sub-basins. An insightful variability exists in the precipitation regime; Shigar will have an increase in precipitation during October–April in the future, and in Gilgit during July–September, but a decrease in the other months. On the other hand, Kharmong lies adjacent to the Himalayan mountain range and it has a different precipitation regime, which is summer-dominated as opposed to that of Shigar and Gilgit. The precipitation during July–August is projected to increase by 60–80% in the sub-basin.
Figure 11

Change in climate in the sub-basins of the UIB according to four MRI-AGCM scenarios (left: monthly temperature; right: monthly precipitation).

Figure 11

Change in climate in the sub-basins of the UIB according to four MRI-AGCM scenarios (left: monthly temperature; right: monthly precipitation).

Close modal

Future flow regime

Figure 12 shows the future (2075–2099) and present (1980–2005) monthly discharge at the Bisham Qila and its sub-basins. At the Bisham Qila and its sub-basins, significant shifts in the spring and summer hydrographs are evident. As summarized in Tables 3 and 4, at the Bisham Qila, annual river flows will decrease by 14% while the the monthly discharge will decrease by 12% in March–May (MAM) and 21% in JJA and increase by 35% in DJF. The patterns and magnitudes are different in the sub-basins because of the glaciers’ elevations and prevalent climate regimes. The Gilgit and Shigar will be the most affected and face reduction of up to 70 and 32% in average annual flows (Table 3). In Gilgit, the decrease in pre-monsoon precipitation will negatively impact the river flows as well (Figure 11). The monthly discharge at Gilgit will decrease by 85% in JAS. The runoff in both of these basins will be impacted by the retreat of glaciers and higher evaporation because of warmer temperatures. Kharmong will have higher river flows (+58%) in the future because of the combined action of precipitation and excess glacier melt (Table 3).
Figure 12

Present (1980–2005) and future (2075–2099) river flows in the UIB and sub-basins. The first row shows Bisham, the second shows Gilgit, the third shows Shigar, and the last shows Kharmong. The present river flows are shown in black.

Figure 12

Present (1980–2005) and future (2075–2099) river flows in the UIB and sub-basins. The first row shows Bisham, the second shows Gilgit, the third shows Shigar, and the last shows Kharmong. The present river flows are shown in black.

Close modal

Other studies have also projected changes in future monthly discharge in the UIB. For example, Lutz et al. (2016) showed that under the representative concentrative pathway (RCP)8.5 scenario, the annual discharge at the Bisham Qila will increase by 30% with a more significant increase by 100 and 110% in DJF and MAM, respectively (Table 4), mainly because of precipitation and temperature increases. Notably, the projection by Lutz et al. (2016) presented that under the RCP4.5 scenario, the runoff will decrease by approximately 20% at the Bisham Qila because the rate of increase in temperature is not as significant as in the RCP8.5 case. In that scenario, the reduction of the glacier coverage plays a more dominant role and consequently, the annual discharge is projected to be decreased. In that regard, our projection at the Bisham Qila is similar to the case of RCP4.5 by Lutz et al. (2016) as both of them project the reduction of annual discharge.

Figure 13 explains the shift in the runoff components by different sources including rainfall, snow, and glacier melt. At the outlets of the Bisham Qila and the sub-basins, rainfall–runoff is projected to increase because of higher precipitation in the summer and change in the phase of precipitation in the winter from snow to rainfall because of higher temperatures.
Figure 13

Comparison between the present (1980–2005) and the future (2075–2099) runoff components of the UIB and sub-basins. Row wise, the simulations represent Bisham Qila, Gilgit, Shigar, and Kharmong, respectively. The black line represents the present and other colors MRI-AGCM.

Figure 13

Comparison between the present (1980–2005) and the future (2075–2099) runoff components of the UIB and sub-basins. Row wise, the simulations represent Bisham Qila, Gilgit, Shigar, and Kharmong, respectively. The black line represents the present and other colors MRI-AGCM.

Close modal
This phase change will negatively impact the snowmelt contribution, resulting in a decrease of its runoff by half of the present quantities. The runoff from glacier melt will also decrease, but the magnitudes are different depending on the elevation distribution of glaciers in the region. In Gilgit, the glaciers will retreat to the extent that it will diminish their contribution to the hydrograph because most of its glaciers are located below 5,000 m. Shigar will also experience a decrease in summer flows because of the glacier retreat and high evaporation. On the other hand, Kharmong will have an excess contribution of glaciers to the hydrograph because its glaciers are located at higher elevations compared to the other two sub-basins (Supplementary material, Figure S3) and (Table 3). Overall, the spring and summer season will be drier in the western region because of the deficit of meltwater in summer (Figure 14). On the other hand, the eastern region will see excess river flows because of wetter monsoon and largely intact glaciers which will melt with higher rates because of higher temperatures.
Figure 14

Percentage changes in the monthly river flows in future (2075–2099) in the UIB and its sub-basins with respect to the present (1980–2005).

Figure 14

Percentage changes in the monthly river flows in future (2075–2099) in the UIB and its sub-basins with respect to the present (1980–2005).

Close modal

The methodology of this research has a few shortcomings which have resulted in biases in the river flow simulations. The snow melt and glacier melt models depend upon the empirical degree-days approach which is reliable but approximates the physical processes. Another source of bias is the spatial resolution of the climate datasets and estimation of precipitation in the mountain regions (Immerzeel et al. 2015). For future projection of river flows, the extent of glaciers is estimated from Huss & Hock (2015) which could be improved by integrating a robust glacier mass-balance model to improve the accuracy of the simulations.

Atmosphere/Ocean General Circulation Models (AOGCMs) are the foundation for the simulation of the climatic response to anthropogenic activities and results from such simulations have provided the information for the majority of the climatic impact studies. However, the coarse resolution of AOGCMs raises constraints on their usefulness, especially in complex physiographic features. Therefore, high-resolution and time-slice AGCM experiments have been developed to improve the regional information of AOGCMs (Mearns et al. 2003). An AGCM is forced by observed SST and sea-ice distribution. Then, the future climate is simulated with the boundary conditions (SST and sea-ice), which are already projected by AOGCM (Mizuta et al. 2012). A high-resolution AGCM is vital for realistic precipitation simulations in mountainous regions and for obtaining meaningful future projections as well. A study used four-member 20-km MRI-AGCM and reported that the greatest extreme precipitation increases are found in the South Asian region (Kitoh et al. 2016). Furthermore, Krishnan et al. (2015) used the high-resolution Laboratoire de Météorologie Dynamique general circulation model (LMDZ) and found a robust increase in the frequency of heavy precipitation (>100 mm/day) over central India.

The fine-resolution MRI-AGCM also contains some biases and uncertainties. For example, Rahman et al. (2013) and Rajendran et al. (2012) employed them in India and Bangladesh to study the monsoon precipitation and found biases in the monsoon precipitation because of the dependence of the simulations on a specific scenario of greenhouse gas emission and boundary conditions. Another study reported that the Asian monsoon precipitation is sensitive to the cumulus parameterization schemes rather than SST (Endo et al. 2012). Therefore, the inclusion of a coupled air-ocean model with AGCM while projecting present and future climate will yield more reliable projections.

The biases of the MRI-AGCM could affect the simulations of the summer river flows in the UIB, especially in the eastern region where monsoon is more prevalent. Nevertheless, MRI-AGCM is a step forward in the climate modeling because of its resolution and improved physical parameterizations which have improved the simulations of heavy monthly-mean precipitation around the tropical Western Pacific, the global distribution of tropical cyclones, and the seasonal march of East Asian summer monsoon (Kitoh et al. 2016).

The UIB is the major source of water for agricultural and domestic water demands in Pakistan. The UIB receives heavy winter precipitation and glaciers also contribute significant quantities of runoff annually. The river flows in the UIB are vulnerable to changing climate, which will cause an imbalance in the water supply and demand. Therefore, for the assessment of the impact of future climate on river flows, this study used the MRI-AGCM with the spatial resolution of 0.1875° and the RRI model coupled with snow and glacier melt models. The MRI-AGCM projected an annual increase of temperature by 5.3 °C and annual precipitation by 17% in the future (2075–2099). Although both temperature and annual precipitation are projected to increase, the annual river flow from the UIB was projected to decrease by 14% mainly due to the deficit in glacier runoff.

Furthermore, the timing of runoff will be altered by the climate change, i.e. increase by 35% in DJF while decrease by 12 and 21% in MAM and JJA, respectively. Our study demonstrated that such patterns are different across the UIB mainly because of the spatial variation in glacier melt and precipitation regime. Such regional analysis and information would be useful for better water management in the UIB and have been enabled because of the fine-spatial resolution of the AGCM.

The simulation can be improved by using finer and accurate climate datasets. Furthermore, the simulations require a robust glacier mass-balance model which will enhance the quality of the analysis. Hence, future studies should be done with dynamic modeling of glacier extent and coverage under future climate conditions.

Special thanks to the Water and Power Development Authority (WAPDA) and the Pakistan Meteorological Authority (PMD) for the provision of hydro-meteorological data.

S.B. was funded by the Japanese government through the MEXT scholarship for PhD studies at Kyoto University. In addition to this, the authors are thankful to Kyoto University for providing additional funding for data collection.

S.B. and T.S. conceptualized the study. S.B. prepared the dataset and performed the modeling tasks. M.Y. contributed to the glacier melt analysis. S.B. wrote the manuscript and all the authors contributed to the review.

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

The authors declare there is no conflict.

Abbas
A.
,
Ullah
S.
,
Ullah
W.
,
Waseem
M.
,
Dou
X.
,
Zhao
C.
,
Karim
A.
,
Zhu
J.
,
Hagan
D. F. T.
,
Bhatti
A. S.
&
Ali
G.
2022
Evaluation and projection of precipitation in Pakistan using the coupled model intercomparison project phase 6 model simulations
.
Int. J. Climatol.
https://doi.org/10.1002/JOC.7602
.
Ali
A.
2013
Indus Basin floods: Mechanisms, Impacts, and Management
.
Mandaluyong City
.
Ali, S., Li, D., Congbin, F. & Khan, F. 2015 Twenty first century climatic and hydrological changes over Upper Indus Basin of Himalayan region of Pakistan. Environmental Research Letters 10, 014007
.
Ali
S.
,
Reboita
M. S.
&
Kiani
R. S.
2021
21st century precipitation and monsoonal shift over Pakistan and Upper Indus Basin (UIB) using high-resolution projections
.
Sci. Total Environ.
797
,
149139
.
https://doi.org/10.1016/j.scitotenv.2021.149139
.
Archer, D. 2003 Contrasting hydrological regimes in the upper Indus Basin. Journal of Hydrology 274, 198–210
.
Babur
M.
,
Babel
M. S.
,
Shrestha
S.
,
Kawasaki
A.
&
Tripathi
N. K.
2016
Assessment of climate change impact on reservoir inflows using multi climate-models under RCPs – the case of Mangla Dam in Pakistan
.
Water
8
,
389
.
https://doi.org/10.3390/W8090389
.
Dahri
Z. H.
,
Moors
E.
,
Ludwig
F.
,
Ahmad
S.
,
Khan
A.
,
Ali
I.
&
Kabat
P.
2018
Adjustment of measurement errors to reconcile precipitation distribution in the high-altitude Indus basin
.
Int. J. Climatol.
38
,
3842
3860
.
https://doi.org/10.1002/joc.5539
.
Dimri
A. P.
,
Yasunari
T.
,
Wiltshire
A.
,
Kumar
P.
,
Mathison
C.
,
Ridley
J.
&
Jacob
D.
2013
Application of regional climate models to the Indian winter monsoon over the western Himalayas
.
Sci. Total Environ.
468–469
,
S36
S47
.
https://doi.org/10.1016/J.SCITOTENV.2013.01.040
.
Hall
D. K.
&
Riggs
G. A.
2001
MODIS/Terra Snow Cover 8-Day L3 Global 500 m SIN Grid, Version 6 [WWW Document]
.
Hasson
S. U.
,
Saeed
F.
,
Böhner
J.
&
Schleussner
C. F.
2019
Water availability in Pakistan from Hindukush–Karakoram–Himalayan watersheds at 1.5 °C and 2 °C Paris agreement targets
.
Adv. Water Resour.
https://doi.org/10.1016/j.advwatres.2019.06.010
.
Huss
M.
&
Hock
R.
2015
A new model for global glacier change and sea-level rise
.
Front. Earth Sci
.
https://doi.org/10.3389/feart.2015.00054
.
Huss
M.
,
Jouvet
G.
,
Farinotti
D.
&
Bauder
A.
2010
Future high-mountain hydrology: a new parameterization of glacier retreat
.
Hydrol. Earth Syst. Sci.
14
,
815
829
.
https://doi.org/10.5194/hess-14-815-2010
.
ICIMOD
2011
Clean Ice and Debris Covered Glaciers of HKH Region [WWW Document]
.
Immerzeel
W. W.
,
van Beek
L. P. H.
&
Bierkens
M. F. P.
2010
Climate change will affect the Asian water towers
.
Science (80-.)
328
,
1382
1385
.
https://doi.org/10.1126/science.1183188
.
Immerzeel
W. W.
,
Wanders
N.
,
Lutz
A. F.
,
Shea
J. M.
&
Bierkens
M. F. P.
2015
Reconciling high-altitude precipitation in the Upper Indus basin with glacier mass balances and runoff
.
Hydrol. Earth Syst. Sci.
19
,
4673
4687
.
https://doi.org/10.5194/hess-19-4673-2015
.
Khan
A. J.
,
Koch
M.
&
Tahir
A. A.
2020
Impacts of climate change on the water availability, seasonality and extremes in the Upper Indus Basin (UIB)
.
Sustainability
.
https://doi.org/10.3390/su12041283
.
Kiani
R. S.
,
Ali
S.
,
Ashfaq
M.
,
Khan
F.
,
Muhammad
S.
,
Reboita
M. S.
&
Farooqi
A.
2021
Hydrological projections over the Upper Indus basin at 1.5 °C and 2.0 °C temperature increase
.
Sci. Total Environ.
788
,
147759
.
https://doi.org/10.1016/j.scitotenv.2021.147759
.
Kitoh
A.
,
Ose
T.
&
Takayabu
I.
2016
Dynamical downscaling for climate projection with high-resolution MRI AGCM-RCM
.
J. Meteorol. Soc. Japan. Ser. II
94A
,
1
16
.
https://doi.org/10.2151/jmsj.2015-022
.
Krishnan
R.
,
Sabin
T. P.
,
Vellore
R.
,
Mujumdar
M.
,
Sanjay
J.
,
Goswami
B. N.
,
Hourdin
F.
,
Dufresne
J. L.
&
Terray
P.
2015
Deciphering the desiccation trend of the South Asian monsoon hydroclimate in a warming world
.
Clim. Dyn.
47
(
3
),
1007
1027
.
https://doi.org/10.1007/S00382-015-2886-5
.
Lutz
A. F.
,
Immerzeel
W. W.
,
Shrestha
A. B.
&
Bierkens
M. F. P.
2014
Consistent increase in High Asia's runoff due to increasing glacier melt and precipitation
.
Nat. Clim. Change
4
,
587
591
.
https://doi.org/10.1038/nclimate2237
.
Lutz
A. F.
,
Immerzeel
W. W.
,
Kraaijenbrink
P. D. A.
,
Shrestha
A. B.
&
Bierkens
M. F. P.
2016
Climate change impacts on the Upper Indus hydrology: sources, shifts and extremes
.
PLoS One
.
https://doi.org/10.1371/journal.pone.0165630
.
Mcsweeney
C.
,
New
M.
&
Lizcano
G.
2012
UNDP Climate Change Country Profiles Pakistan
.
Mearns
L. O.
,
Giorgi
F.
,
Whetton
P.
,
Pabon
D.
&
Lal
M.
2003
Guidelines for Use of Climate Scenarios Developed from Regional Climate Model Experiments
.
Miao
C.
,
Su
L.
,
Sun
Q.
&
Duan
Q.
2016
A nonstationary bias-correction technique to remove bias in GCM simulations
.
J. Geophys. Res. Atmos.
121
,
5718
5735
.
https://doi.org/10.1002/2015JD024159
.
Mizuta
R.
,
Yoshimura
H.
,
Murakami
H.
,
Matsueda
M.
,
Endo
H.
,
Ose
T.
,
Kamiguchi
K.
,
Hosaka
M.
,
Sugi
M.
,
Yukimoto
S.
,
Kusunoki
S.
&
Kitoh
A.
2012
Climate simulations using MRI-AGCM3.2 with 20-km grid
.
J. Meteorolog. Soc. Japan
90A
,
233
258
.
https://doi.org/10.2151/jmsj.2012-A12
.
Ougahi
J. H.
,
Cutler
M. E. J.
&
Cook
S. J.
2022
Modelling climate change impact on water resources of the Upper Indus Basin
.
J. Water Clim. Change
13
,
482
504
.
https://doi.org/10.2166/wcc.2021.233
.
Rahman
M. M.
,
Rafiuddin
M.
,
Alam
M. M.
,
Kusunoki
S.
,
Kitoh
A.
&
Giorgi
F.
2013
Summer monsoon rainfall scenario over Bangladesh using a high-resolution AGCM
.
Nat. Hazards
69
,
793
807
.
https://doi.org/10.1007/s11069-013-0734-7
.
Rajendran
K.
,
Kitoh
A.
,
Srinivasan
J.
,
Mizuta
R.
&
Krishnan
R.
2012
Monsoon circulation interaction with Western Ghats orography under changing climate
.
Theor. Appl. Climatol.
110
,
555
571
.
https://doi.org/10.1007/s00704-012-0690-2
.
Reggiani
P.
&
Rientjes
T. H. M.
2015
A reflection on the long-term water balance of the Upper Indus Basin
.
Hydrol. Res.
46
,
446
462
.
https://doi.org/10.2166/nh.2014.060
.
Rounce
D. R.
,
Hock
R.
&
Shean
D. E.
2020
Glacier mass change in high mountain Asia through 2100 using the open-source Python Glacier Evolution Model (PyGEM)
.
Front. Earth Sci.
7
.
https://doi.org/10.3389/feart.2019.00331
.
Sanjay
J.
,
Krishnan
R.
,
Shrestha
A. B.
,
Rajbhandari
R.
&
Ren
G.-Y.
2017
Downscaled climate change projections for the Hindu Kush Himalayan region using CORDEX South Asia regional climate models
.
Adv. Clim. Change Res.
8
,
185
198
.
https://doi.org/10.1016/j.accre.2017.08.003
.
Sayama
T.
2015
RRI User manual
.
Sayama
T.
,
Ozawa
G.
,
Kawakami
T.
,
Nabesaka
S.
&
Fukami
K.
2012
Rainfall–runoff–inundation analysis of the 2010 Pakistan flood in the Kabul River basin
.
Hydrol. Sci. J.
57
,
298
312
.
https://doi.org/10.1080/02626667.2011.644245
.
Shah
M. I.
,
Khan
A.
,
Akbar
T. A.
,
Hassan
Q. K.
,
Khan
A. J.
&
Dewan
A.
2020
Predicting hydrologic responses to climate changes in highly glacierized and mountainous region Upper Indus basin
.
R. Soc. Open Sci.
7
,
191957
.
https://doi.org/10.1098/rsos.191957
.
Su
F.
,
Zhang
L.
,
Ou
T.
,
Chen
D.
,
Yao
T.
,
Tong
K.
&
Qi
Y.
2016
Hydrological response to future climate changes for the major upstream river basins in the Tibetan Plateau
.
Global Planet. Change
136
,
82
95
.
https://doi.org/10.1016/j.gloplacha.2015.10.012
.
Terink
W.
,
Lutz
A. F.
,
Simons
G. W. H.
,
Immerzeel
W. W.
&
Droogers
P.
2015
SPHY v2.0: spatial processes in Hydrology
.
Geosci. Model Dev.
8
,
2009
2034
.
https://doi.org/10.5194/gmd-8-2009-2015
.
Wester
P.
,
Mishra
A.
,
Mukherji
A.
&
Shrestha
A. B.
2019
The Hindu Kush Himalaya Assessment
.
Springer International Publishing
,
Cham
.
https://doi.org/10.1007/978-3-319-92288-1.
Yatagai
A.
,
Kamiguchi
K.
,
Arakawa
O.
,
Hamada
A.
,
Yasutomi
N.
&
Kitoh
A.
2012
APHRODITE: constructing a long-term daily gridded precipitation dataset for Asia based on a dense network of rain gauges
.
Bull. Am. Meteorol. Soc.
93
,
1401
1415
.
https://doi.org/10.1175/BAMS-D-11-00122.1
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Supplementary data