Runoff conditions in the Fuping Basin under an ensemble of climate change scenarios*

ROMAINE Ingabire,CAO Bo,CAO Jiansheng,ZHANG Xiaolong,LIU Xia,SHEN Yanjun**

(1.Center for Agricultural Resources Research,Institute of Genetics and Developmental Biology,Chinese Academy of Sciences / Key Laboratory of Agricultural Water Resources,Chinese Academy of Sciences / Hebei Key Laboratory of Water-saving Agriculture,Shijiazhuang 050022,China; 2.University of Chinese Academy of Sciences,Beijing 100049,China)

Abstract: Changes in runoff are of great significance for water resources management,especially under the changing climate.In the Fuping Basin,one of the basins in the upper reaches of the Daqinghe Basin,the water resources are facing changes which show great importance of further studies on runoff conditions in the future in this basin.Hence,in this paper,MIKE11-NAM model was applied to simulate daily runoff(2008-2017) and future runoff conditions under a changing climate in the near future(2025-2054) in the Fuping Basin.After bias correction,an ensemble of four regional climate models(RCMs) was used to develop future climate data under three shared socio-economic pathways(SSP1-2.6,SSP2-4.5,and SSP5-8.5) scenarios.The obtained results showed a good performance of the MIKE11-NAM model in simulating daily runoff.R2 and Nash-Sutcliffe efficiency coefficient(NSE) were 0.82 and 0.81 for calibration,0.87 and 0.87 for validation,respectively.Although uncertainties remain,the correlation between observed and simulated RCM data was improved after bias correction for all models.Precipitation and temperature were projected to increase under all scenarios compared to the baseline period(1985-2014).Annual temperature and precipitation will increase by 2.45 ℃ and 124 mm under the SSP5-8.5 and SSP2-4.5 scenarios,respectively.However,precipitation is expected to mainly increase in summer while temperature will increase in all the seasons.The projected annual runoff will increase under SSP2-4.5 while decreasing under SSP1-2.6 and SSP5-8.5 scenarios.Seasonally,the future runoff will decrease during spring and summer under all the scenarios.Generally,the changes in runoff conditions will be more obvious in the future.Our findings can be important for integrated water resources management and planning in this region.

Keywords: MIKE11-NAM; Daqinghe Basin; Simulation; Runoff; Climate change

Climate change is a serious problem that strongly impacts the environment.The hydrologic cycle is one of the affected parts of the environment,which is expected to face modifications in the upcoming future(Shen et al.,2008; Xia et al.,2017).Precipitation affects runoff directly while temperature affects it indirectly through evapotranspiration,snow,and glacier melt(Molini et al.,2011).Runoff conditions are highly impacted by climate change(Shen et al.,2018),thus,the changes of runoff in present and near future need to be focused on when dealing with water resources management and planning.

MIKE11-NAM model,being a part of the MIKE11 module,has been widely used by many researchers(Amir et al.,2013; Doulgeris et al.,2012; Filianoti et al.,2020; Hafezparast,2015; Makungo et al.,2010; Odiyo et al.,2012; Shrestha et al.,2020).Makungo et al.(2010)coupled MIKE11-NAM and the Australian Water Balance Model(AWBM) to simulate runoff hydrograph of the ungauged Nzhelele River catchment in South Africa,their research confirmed the effectiveness of both MIKE11-NAM and AWBM models to estimate the runoff in ungauged catchments.Odiyo et al.(2012) used MIKE11-NAM to estimate the flows that Latonyanda River contributes to Luvuvhu River downstream of Albasini Dam in South Africa,to make the planning and management of water requirements for downstream users possible; the authors also estimated the ungauged flows.Doulgeris et al.(2012) also used MIKE11-NAM to simulate rainfall-runoff processes of the Strymonas River and Lake Kerkini in Greece.The results showed that the model can simulate with precision both the daily and monthly runoff of the Strymonas River that ends up in Lake Kerkini.Hafezparast(2015) used the model to study the peak and monthly runoff at the Sarisoo River Basin in the North West of Iran and concluded that MIKE11-NAM being a hydrological model that does not require a lot of input data to calculate daily and monthly runoff,which made it a vital tool for water management model on the large scale modeling with middle and longterm simulation period.Amir et al.(2013) used MIKE11-NAM by considering the multi-objective calibration to get the optimum parameters for the Fitzroy Basin.They concluded that these parameters could be used to set up an integrated hydrologic and hydraulic model which could predict flood with the best accuracy.Filianoti et al.(2020) performed a comparison among HEC-HMS,SWMM,MIKE11-NAM and WEC-FLOOD models in the Mesima Torrent watershed(Italy).The comparative study showed that MIKE11-NAM had a sufficient prediction accuracy and could be used to simulate runoff under climate change.Overall,as shown by the above researchers,MIKE11-NAM is a model of conceptual type which has been widely used in different regions around the world with different physical properties and under different climatic conditions.In addition to that,it is a model that gives good results with low input data.Furthermore,it is even suitable for basins with data scarcity problems.

Model approaches with climate models’ outputs from the Global Circulation Models(GCMs) driven by greenhouse gas emission scenarios are usually established to study the future changes in regional hydrologic cycles(Islam et al.,2019).Shrestha et al.(2020) assessed the impact of climate change on hydrology and river morphology in the Chindwin River Basin using MIKE models and GCM output under two Representative Concentration Pathways Scenarios(RCP4.5 and RCP8.5),they found that the discharge over a period from 2018-2040 is projected to increase under RCP4.5 and RCP8.5 respectively.Luo et al.(2019) analyzed the impact of climate change on water resources in Xinjiang using SWAT model with GCM output also under RCP4.5 and RCP8.5 scenarios,the results showed that runoff will increase under RCP4.5 and RCP8.5 scenarios in the future.Bao et al.(2020) established the Variable Infiltration Capacity(VIC) model in the Haihe River Basin and projected the future runoff by using the outputs from 18 general circulation models(GCMs) of the Coupled Model Inter-comparison Project Phase 5(CMIP5),together with three RCP scenarios.Their research showed that the Haihe River Basin will become warmer and wetter and that the runoff might increase under RCP2.6,RCP4.5 and RCP8.5 scenarios in future.Since GCMs outputs provide historical and future climatic data that can be used as input for hydrological models,it is important to couple hydrological models with GCMs outputs to study the future impacts of climate on hydrology,which will help to better understand the future runoff conditions under a projected climate change.

In northern China,a decrease in runoff was detected since 1950(Liu et al.,2017).As shown by past studies,the reason behind this decline was two main contributing factors namely climate change and human activities(Feng et al.,2019; Li et al.,2013; Li et al.,2018).The Daqinghe Basin is interesting from a flood simulation point of view as suggested by previous researchers(Li et al.,2014),however,recent study also indicated that precipitation and temperature,which generally reflect the changes of climate,showed a remarkable increase trend,the runoff condition is likely to become more unstable(Xu et al.,2014; Zhang et al.,2021a).Up to the present,the studies carried out in different study periods on the climate change and runoff trend showed that: precipitation decrease was not significant whereas temperature increased significantly(Wang et al.,2018).Moreover,a decreasing trend of runoff was noticed in this basin(Chen et al.,2011; Gao et al.,2017; Zhou et al.,2011;Liu,2020).Runoff in this basin is mainly affected by climate change other than human activities(Zhou et al.,2011; Liu et al.,2021; Wang et al.,2012; Yang et al.,2018).This reflects that although researches have been done on the effect of climate change on runoff in the Fuping Basin,climate change is still a driving factor affecting runoff which still needs to be studied now and in the near future for effective water resources management and planning in this basin.Therefore,in light of the importance of runoff conditions under climate change,the main objectives of this study are to evaluate the MIKE11-NAM model performance and applicability in simulating runoff,and to estimate the runoff change behaviors in the Fuping Basin under climate change scenarios.The results can be useful for better understanding the variability of water resources in the Fuping Basin,especially under the changing climate.

1 Study area,datasets and methods

1.1 Study area

The Fuping Basin is one of the basins in the upper reaches of the Daqinghe Basin which is located in the middle of the Haihe River Basin in the northern part of China.The basin lies between 113°39′-114°34′E and 38°10′-39°30′N,and covers an area of 2204.4 km.The region is characterized by mountainous relief with an elevation range of 253 to 2781 m above sea level(a.s.l)(Fig.1).The basin is characterized by a warm temperate continental monsoon climate where the average annual rainfall is 650 mm,most of which occurs in the summer,especially in July and August.Its annual average temperature is 7.56 ℃.The Fuping Basin is an important water source for downstream rivers and reservoirs.

Fig.1 Location and elevation of,and spatial distribution of meteorological and hydrological stations in the Fuping Basin

1.2 Data and exploratory analysis

Time series(2008-2017) of daily maximum,mean and minimum temperature(,and),wind speed,solar radiation and relative humidity were obtained from China Meteorological Data Assimilation(CMDAS,http://www.cmads.org/).The CMDAS dataset is developed by using many different scientific approaches and technologies including loop nesting of data,projection of resampling models,bilinear interpolation and LAPS/STMAS.Furthermore,this data was selected for the study because it has been successfully applied in hydrological researches in different regions in China(Wang et al.,2021; Zhao et al.,2018).Long-term time series(1980-2017) of daily,,and precipitation in the Fuping Basin were collected from the China Meteorological Data Service Center(CMDC)(https://data.cma.cn).Daily rainfall(12 stations from 2008 to 2017) and discharge(from 2008 to 2017) data for the Fuping Basin were collected from the hydrological yearbook of China.The calculation of average rainfall was performed in ArcMap using the Thiessen polygon method.Potential evapotranspiration was calculated by using the Penman-Monteith equation(Allen et al.,1988).The China Digital Elevation Model(DEM) of 30 m resolution was generated from the geospatial data cloud(http://www.gscloud.cn/home).It was used in watershed delineation.

To analyze runoff conditions in the Fuping Basin in the future,both the historical and an ensemble of climate change scenarios data from Regional Climate Models(RCM) were also utilized during the simulation processes.These datasets were obtained from Coupled Model Intercomparison Project(CMIP6)(https://esgfdata.dkrz.de/projects/cmip6-dkrz/),in which the selected models for this study are ACCESS_ESM1_5,EC_Earth3,CanESM5 and MPI_ESM1_2_LR(Yang et al.,2021; Tian et al.,2021).Daily,,,and precipitation are included in these datasets.They were downloaded under baseline and future Shared Socialeconomic Pathways(SSP1-2.6,SSP2-4.5,and SSP5-8.5)scenarios.The selected baseline period was 1985-2014 and the future period selected under SSP1-2.6,SSP2-4.5,and SSP5-8.5 was 2025-2054.

To reduce bias between observed and simulated RCM climate data,bias correction was performed.The linear scaling(LS) method of bias correction was chosen,which can be applied for correcting both precipitation and temperature(Teutschbein & Seibert,2012).The linear scaling method being a mean-based approach implements a constant corrected factor that is estimated by the difference between original RCM simulations and observations for each calendar month.Precipitation is adjusted with a multiplier which is the ratio between the mean monthly observed and simulated historical data whereas temperature is adjusted using additive factor which is the difference between mean monthly observed and simulated historical data as shown in equations 1 to 4(Luo et al.,2018; Teutschbein & Seibert,2012).The bias correction factors are assumed to be constant for both historical and future runs in this study.

To reflect the bias correction performance,the correlation coefficient() and root mean square error(RMSE) were calculated between observed and RCM simulated climate data.

1.3 MIKE 11 NAM model setup and sensitivity analysis

MIKE11-NAM is a part of MIKE11 model which was developed by the Danish Institute of hydraulics(DHI)(DHI,2009).Although it is a conceptual rainfallrunoff model,the model was widely used in many parts of the world within catchments which differ in sizes,soil properties,climatic conditions and land use due to a low number of data input,user-friendly working environment,easy setup,high prediction accuracy.The Fuping Basin has fewer human activities and is a relatively natural basin.Since the main objective is to simulate the changes in runoff,thus,all these characteristics lead MIKE11-NAM model to be chosen for simulating hydrological processes in the Fuping Basin.

MIKE11-NAM model takes into account up to its four interrelated storages which are surface storage,root zone storage,groundwater storage and snow storage; and covers nine parameters(Fig.2,Table 1)(DHI,2009).However,snow storage was not considered in this study because the amount of snow present in this region is negligible due to its warm temperate climate.Furthermore,the fundamental modeling components are described below:

Table 1 Parameters used in the MIKE11-NAM model during simulation

Fig.2 MIKE11-NAM structure(DHI,2009)

The overland flow is given by Equation 5,it is proportional to the excess rainfall and varies linearly with the soil moisture content of the lower zone storage.

Where,CQOF is the overland flow runoff coefficient(0 ≤CQOF ≤1); TOF is the threshold value for overland flow(0 ≤TOF ≤1);is excess rainfall; the interflow QIF is given by Equation 6.It is routed through two linear reservoirs in series with time constant CK1,2.

Where CKIF is the time constant for interflow and TIF is the root zone threshold value for interflow(0 ≤TIF ≤1).

The amount of infiltrating waterrecharging the groundwater storage depends on the soil moisture content in the root zone,and is given by Equation 7.

Where TG is the root zone threshold value for groundwater recharge(0 ≤TG ≤1).

Sensitivity analysis was carried out to find out which parameters are sensitive to affect the shape of the hydrograph and peak runoff,and to determine the good fitness of model parameters during calibration between the observed and simulated discharge.Model parameters were adjusted using both manual and auto-calibration(Madsen,2000).We first performed manual calibration to find the relatively sensitive parameters.Furthermore,auto-calibration was adopted.Both the coefficient of correlation() and Nash-Sutcliffe efficiency(NSE)(Equation 8) were used(Knoben et al.,2019; Nash & Sutcliffe,1970).

2 Results and discussion

2.1 Trend analysis of Tmean,precipitation and runoff

Due to data availability,trend analysis of,precipitation and runoff data were performed using longterm meteorological and hydrological data in the Fuping Basin from 1980 to 2017 based on the Mann-Kendall test.Annual temperature showed a significant upward trend(＜0.05),annual runoff showed an insignificant downward trend and there was no significant trend for precipitation as shown in Table 2 and Fig.3.Adding further to the trend analysis,the seasonal analysis showed that temperature increased significantly in all seasons,especially in winter [0.7 ℃·(10a),＜0.001].Annual precipitation showed no significant trend,however,the precipitation increased significantly in autumn [17.5 mm·(10a),＜0.001].Moreover,runoff increasing and decreasing trends were not significant in all seasons(Table 2).

Table 2 Trends of temperature,precipitation and runoff from 1980 to 2017 in the Fuping Basin based on the Mann-Kendall test

Fig.3 Trend analysis of annual precipitation,temperature and runoff in the Fuping Basin from 1980 to 2017

2.2 MIKE11-NAM model performance

The MIKE11-NAM model could capture the hydrological dynamics on a daily scale in the Fuping Basin.Observed and simulated discharge during both calibration(2009-2011) and validation(2012-2017) periods show a reasonably good fit,withand NSE values of 0.82 and 0.81 for the calibration period and 0.87 and 0.87 for the validation period,respectively(Fig.4).It is worth noting that although low flows were well captured during both calibration and validation periods,some of the peak flows were underestimated during the calibration period.The major peak runoff events are relatively consistent with major peak rainfall events,and the peak runoffs were reproduced in the simulation during summer seasons,especially during July and August.

Fig.4 Observed vs.simulated runoff for both calibration(a,2009-2011) and validation(b,2012-2017) periods of the Fuping Basin

Nine parameters were selected for calibration:,,CQOF,CKIF,CK1,2,TOF,TIF,TG,CKBF(Table1).In addition to the parameter calibration,Table 1 also lists the value ranges and the final value of parameters that were obtained in this study,which are further discussed in the discussion part.Furthermore,these parameters represent the catchment hydrological characteristics.Overall,the evaluation of the runoff simulations suggests that the MIKE11-NAM model could be applied for runoff projection under climate change scenarios.

2.3 Bias correction of RCM data with historic observations

In the historical period,all model data from RCMs were underestimated when compared to the observed data before bias correction.The correlation between RCM temperature data and observed data was reasonable,especially for MPI_ESM1_2HR and ACCESS_ESM1_5 models than Can_ESM5 and EC_EARTH3.Although the correlation was reasonable,the data still needed to be improved by conducting bias correction.After temperature bias correction the relationship between RCM models and the observed temperature was improved with increasedand reduced RMSE(Table 3),which explains a relative good performance of the linear scaling approach in correcting temperature.

In contrast to temperature,precipitation was more underestimated before bias correction.The correlation between observed data and simulated RCM data was relatively low.Can_ESM5 correlation with the observed precipitation data was the weakest compared to other RCMs,followed by MPI_ESM1_2HR,ACCESS_ESM1_5,and EC_EARTH3 in ascending order.After bias correction,the correlation between observed and simulated RCM data was improved for all models.Although bias correction was conducted for improving data used in this study,this did’t mean that all uncertainties found were eliminated rather reduced as shown in Table 3.Fur-thermore,the bias factors were also applied to the future climate data.Thereafter,future climate change projection was done.Moreover,for the future runoff projection,the multi-model ensemble(MME) for both precipitation and temperature was used as input for the model because the multi-model ensemble outperforms any individual RCM(Gu et al.,2018).

Table 3 Bias correction performance of four regional climate models for monthly precipitation and temperature estimation compared to the observed data

Precipitation is the main source of runoff and as temperature also affects runoff,the climate change characteristics were studied for both precipitation and mean temperature() in the future(2025-2054) under SSP1-2.6,SSP2-4.5,and SSP5-8.5 scenarios.It was found that the precipitation will increase in the future.The baseline annual precipitation(670 mm) will increase by 97 mm,124 mm,and 99 mm in the future under SSP1-2.6,SSP2-4.5,and SSP5-8.5 scenarios,respectively.The precipitation change is more significant under the SSP2-4.5 scenario than SSP5-8.5 and SSP1-2.6,and will increase by 18.5%(124 mm) compared to the baseline(Table 4).On monthly basis,an increasing trend of precipitation was observed in all seasons and it was more significant in the summer season,especially in July and August.In the wettest season(June,July,August and September) of the study period,the rainfall would increase by 24 mm(20%) in the future compared to the baseline period(Fig.5).

Table 4 Precipitation and temperature change comparison between historical observed data(1985-2014) and future data(2025-2054) based on four regional climate models(RCM) and multimodel ensemble under 3 scenarios

Fig.5 Future(2025-2054) monthly precipitation projection based on multi-model ensemble of four regional climate models(MME) compared to the baseline(1985-2014) precipitation

Temperature,like precipitation,may also increase under all three scenarios in the future.The annual mean temperature in the baseline was 7.8 ℃,which would increase by 1.87 ℃,1.92 ℃,and 2.45 ℃ under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios in the future,respectively.However,unlike precipitation,temperature under the highest emission scenario(SSP5-8.5) would have a remarkable increase +2.45 ℃ compared to the baseline temperature(Table 4).On the monthly basis,the temperature was projected to increase in all seasons and months.In addition to that,the warming was projected to be higher in colder months than in warmer months.During winter,the temperature would increase by 1.92 ℃,2.10 ℃ and 2.56 ℃ and in summer would increase by 1.78 ℃,1.70 ℃ and 2.19 ℃,respectively,compared to the baseline under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios.The increase in temperature was projected to mainly occur in winter than in summer(Fig.6).

Fig.6 Future(2025-2054) monthly temperature projection based on multi-model ensemble of four regional climate models(MME) compared to the baseline(1985-2014) temperature

2.4 Runoff condition under climate scenarios

After future climate projection,the calibrated MIKE11-NAM was used to predict the future runoff(2025-2054) in the Fuping Basin under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios.Compared to the baseline(1985-2014),the future annual runoff was projected to increase by 3.5 mm under SSP2-4.5,and decrease by 12.0 mm and 11.0 mm under SSP1-2.6 and SSP5-8.5 scenarios,respectively.Its monthly variability was projected to increase under all scenarios in October and November,increase only under SSP2-4.5 and SSP5-8.5 in January and September,and it increased only under SSP2-4.5 in December.However,it will decrease under all scenarios in February,March,April,May,June,July and August(Fig.7).Adding further to the analysis of the runoff condition in the Fuping Basin,a seasonal analysis showed that runoff would increase by 0.3 mm,10.6 mm and 3.3 mm during autumn,respectively,under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios.During winter,the runoff would increase by 1.1 mm under SSP2-4.5 while decrease by 1.1 mm and 0.5 mm under SSP1-2.6 and SSP5-8.5 scenarios,respectively.However,the runoff would decrease by 2.3 mm,1.2 mm and 1.9 mm,respectively,during spring; and by 9.0 mm,7.1 mm and 12.9 mm,respectively,during summer,under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios,respectively.

Fig.7 Projected future monthly runoff(2025-2054) in the Fuping Basin under SSP1-2.6,SSP2-4.5,and SSP5-8.5 scenarios compare to baseline(1985-2014)

3 Discussion

MIKE11-NAM model showed some levels of uncertainties which were possibly due to the input data and model parameters used when running the model.Odiyo et al.(2012) found that the number of rainfall gauging stations in the study area might affect the estimation of peak and low flows.Therefore,as the estimation of peak and low flows depends on the mean rainfall in the catchment,more rainfall gauging stations data are recommended in catchments where they are available.Although this study developed a multi-objective calibration of parameters by following the sensitivity analysis,parameters still played an important role in facilitating the calibration process.Therefore,each parameter sensitivity was explained below:known as the soil moisture content in the root zone available for vegetation transpiration.During calibration,it was found important and very sensitive in our study because while it increased the peak runoff decreases significantly,and vice versa,which affects the shape of the hydrograph and strongly the total volume.Moreover,as mentioned by DHI(2009),it was an important parameter to be adjusted to satisfy the water balance condition during simulation because evaporation is strongly dependent on it,especially during dry seasons; and so does the distribution of rainfall into evaporation,runoff and groundwater(Joynes,2009).CQOF describing the fractions of the excess rainfall that generates overland flow,and infiltration in the catchment,was found sensitive in this study.In addition to that,the CQOF value obtained after the model calibration was higher in the Fuping Basin,espacilly in catchments with steep slopes and big sizes.Therefore,our study’s findings are consistent with(DHI,2009)’s statements because although the Fuping Basin is a small catchment,it is dominated by highlands.

The time constant for routing overland flow(CK1,2) during calibration was very sensitive to the shape of the hydrograph and was calibrated up to 40.5 hours,which means that it took a long time lag for the peak discharge to occur and the reason behind this was that the amount of water existing in the surface and root zone(and) storages may be lower.In addition to that,the latter does not have more effect on the total runoff volume which is consistent with other findings(Keskin et al.,2008).The CKBF parameter strongly affects the timing of runoff through hydrograph recession determined during dry periods and was calibrated to a moderate value in this study.TOF is the root zone threshold value for overland flow.It means that for the flow to occur the moisture content of the lower root zone storage must be above the threshold value which makes this parameter important for model calibration.Adding further to this,our study’s findings showed that this parameter was very sensitive.Moreover,,CQOF,CK1,2,TIF,and TOF were recommended as the most sensitive parameters that need to be calibrated during model simulation(Trinh and Molkenthin,2021).Some other studies found that,,CQOF,and CKIF were important parameters(Makungo et al.,2010; Odiyo et al.,2012).Our study findings were slightly different from theirs as discussed previously.This is possibly due to the difference in climatic conditions,soil physical properties,topography and land cover of the study areas.

For the trend analysis of temperature,precipitation and runoff in the historical periods(1980-2017),the annual temperature showed a significant upward trend whereas runoff showed an insignificant downward trend and precipitation showed no significant trend.Our findings were consistent with other studies that temperature increased significantly,runoff decreased and precipitation showed no significant trend(Jiao et al.,2021; Li et al.,2018; Xu et al.,2020; Gao et al.,2017).Moreover,the annual precipitation and temperature were projected to increase under SSP1-2.6,SSP2-4.5,and SSP5-8.5 scenarios in the future(2025-2054) in this study.

From an extensive literature review,there was a distinct lack of research information on the future climate projection and runoff condition under the change condition in the study area,particularly in this catchment.However,on a regional scale,particularly in northern China,previous studies’ results were follows: mean temperature would increase between 1 ℃ and 5 ℃ in the future in North and Northeast China under all scenarios by 2100 compared to the baseline based on CMIP6 data(Zhang et al.,2021b).It was expected to increase under higher latitudes and altitudes of China(Yang et al.,2021).Rao et al.(2019) by using CMIP5 data found that an extreme increase in annual precipitation was projected to occur in northern China in the future(2046-2065 and 2080-2099).Furthermore,Gao et al.(2022) also by using CMIP5 data(2031-2065 and 2066-2100) reported that future precipitation would increase in the Daqinghe Basin.Since Fuping is located in the Daqinghe Basin which is in the northern part of China,thus,all of these studies were generally consistent with the present study’s findings.However,the existing little difference in the rate of increase compared to this study was possibly due to different types of data or the study periods used.

Moreover,an increase in future runoff was projected to occur in northern China under all scenarios in previous studies.Guan et al.(2021) using CMIP6 data found that future runoff would increase in northern China due to the increase precipitation in that region.Xing et al.(2018) using CMIP5 data(2040-2069) found that runoff would increase in the northern China basins.Bao et al.(2020) using CMIP5 data(2010-2099) found that runoff was projected to increase in the Haihe Basin.Since Fuping is located in the Daqinghe Basin belong to the Haihe Basin and northern China,these studies’ findings are slightly different from the present study’s findings because runoff was projected to increase only under the SSP2-4.5 scenario.Moreover,the future runoff rise may be due to the projected extreme increase in precipitation in northern China(Guan et al.,2021).

Furthermore,in this study,the projected runoff increased under SSP2-4.5 may be due to the highest increase of precipitation and moderate temperature increase observed under the scenario whereas the highest temperature increase observed under SSP5-8.5 may induce the high evaporation under this scenario,therefore,as seen previously that temperature affects runoff through evaporation,which may lead to the decrease in future runoff observed under the latter scenario.

Therefore,the projected increase in runoff is expected to contribute to relieving water shortages in the Fuping Basin as well as in the rest of northern China as shown by past studies(Jiang,2009).Although the increase in available river runoff is expected to reduce water shortage problems as shown by past studies(Guan et al.,2021),it may be expected that serious extreme hydrological events like floods may be inevitable hazards in the future due to the projected future extreme climatic and hydrological modifications which need further studies.

4 Conclusions

Our study provided an assessment of the applicability and reliability of MIKE11-NAM in runoff simulation and estimation,and the runoff condition under future climate change scenarios in the Fuping Basin.Annual temperature showed a significant upward trend,especially in the winter season.There was no significant trend for annual precipitation.However,it increased significantly in autumn.Annual runoff showed an insignificant downward trend in the Fuping Basin.Both temperature and precipitation would increase significantly under SSP1-2.6,SSP2-4.5 and SSP5-8.5 scenarios in the Fuping Basin in the future.The highest temperature and precipitation rise were expected to occur under the SSP5-8.5 and SSP2-4.5 scenarios,respectively.MIKE11-NAM model can simulate and estimate runoff conditions in the Fuping Basin both in the calibration and validation periods.Furthermore,the hydrological characteristics of the Fuping Basin were described in terms of the final parameters obtained after the model calibration and validation.Driving with RCM data,the annual runoff would increase under SSP2-4.5 and decrease under SSP1-2.6 and SSP5-8.5 scenarios compared to the baseline.Moreover,the future runoff would increase in autumn under all scenarios,while it would increase only during winter under SSP2-4.5 scenarios and decrease during spring and summer under all scenarios.Although uncertainties remain,this study showed a future picture of the hydrological conditions of the Fuping Basin,which could help the government and planners to make future projects.The variability of extreme hydrological events(e.g.river floods) in this mountainous region still required further research under the changing climate.

We thank the Belt and Road fellowship and the President’s fellowship for their financial support to International students.